AF100 and judder

100 replies [Last post]
Piotr
Offline
Joined: Jun 9 2006

I notice a worrying amont of judder with AF100 when panning.
It is as usual pronounced on contrasty verticals, but seems inordinately high even at slow pan speeds. Can it be connected to a particular sampling scheme used on the sensor? Or some setup configuration? The shutter is half and camera in 1080/25p, detail -2, film mode, but judder seems too high for me on slow pans. Any gives?

infocus
Offline
Joined: Jul 18 2003

Is it worse than any other 25p camera? (I don't see why it should be.)

Shutter speed setting? Is it set to 1/50 sec?

Piotr
Offline
Joined: Jun 9 2006

yes, half shutter. it is as bad as on gh1, present even at low panning speeds, most pronounced on contrasty verticals.

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

It's exactly the same as in any other camera shooting 25p. It isn't the frame rate that matters, it's the fact that captured images (frames) are duplicated on presentation. Classic film motion.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

MAGLINK
Offline
Joined: Mar 8 2007

Hmmm it may be that we have another HPX370/371 type situation, they claimed a new sensor on that camera but it was the filtering that was causing the smearing which was a lot more than motion blurr, a firmware release basically gave you an HPX300/301 again, there is a thread about the AF100/101 here: http://www.dvxuser.com/V6/showthread.php?233634-Ghosting-and-Trails

Piotr
Offline
Joined: Jun 9 2006

I subjectively see it as worse than on other 25p/24p acquisition systems. Many users on dvxuser.com confirm it's worse, some think it's not any worse. Is there any objective way of measuring and comparing this phenomenon?

Dave R Smith
Offline
Joined: May 10 2005
infocus wrote:
Is it worse than any other 25p camera? (I don't see why it should be.)

Shutter speed setting? Is it set to 1/50 sec?

As implied, I think, slower shutter speed will reduce judder(filmic effect), but if shutter speed is slower than frame rate, ie 1/26 -1/50 or slower, then 1 captured frame would result in 2 recorded frames.

Can you extract frame by frame stills, overlay them (say in photoshop) and check that there aren't consecutive frames the same.

For comparitive test, perhaps use moving vehicle at known speed (say 30 mph) so you know 'object to camera' relative movement is smooth, not distorted by any lacking fluidity in tripod head (or freehand) movement.

Piotr
Offline
Joined: Jun 9 2006

The shutter used was "half" (180deg or 1/50). Unfortunately I won't be able to conduct any tests in the next two weeks as I'll be away. However I have a hunch the character of the AF101 judder has something to do with the CMOS rolling shutter or the ways Panasonic tried to minimize it's impact. I don't know if it's the right track but clues are:
-it's different from our CCD based cameras (sdx900 and hpx3100).
-it's most pronounced on pans with many verticals, especially contrasty ones
-it's not so bad on oblong shapes moving across the frame (cars)
Of course there is a possibility we are seeing things, that's why I was curious for a scientific method to quantify the phenomenon.

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

CCDs don't have rolling shutters, and so are irrelevant here.

Conventional CMOS camera shave a rolling shutter, in that the image scanning process is done directly on the sensor itself, just like on a tubed camera. The AF101 does the scanning a lot faster, reading the image out into a transfer store, just like in a CCD, but the transfer time isn't instantaneous as it is in FIT CCDs. I don't know what the transfer time is, but it's a loot shorter than the scan rate. Motion capture looked just like any other progressive 25Hz camera to me, and the rolling shutter effect is a lot less than in other CMOS cameras.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

drc_online
Offline
Joined: Apr 19 2010

We did some testing with the AF101 a few weeks back and noticed a motion artefact that was more like ghosting on some shots.

It's not judder, as you would normally describe it - rather a feint double-image on movement that was much more pronounced than the normal motion blur effect.

Unfortunately I didn't shoot the shots concerned so I don't have a definitive note of ISO, NR settings, detail etc. (the AF101 was also set to record offspeed, which may or may not be significant)

I'm going to try to re-create this effect at BVE next week and show it to the Panasonic folks, as it is quite disconcerting and, if repeatable, would certainly raise some concerns.

Dave

MAGLINK
Offline
Joined: Mar 8 2007
drc_online wrote:
We did some testing with the AF101 a few weeks back and noticed a motion artefact that was more like ghosting on some shots.

It's not judder, as you would normally describe it - rather a feint double-image on movement that was much more pronounced than the normal motion blur effect.

Unfortunately I didn't shoot the shots concerned so I don't have a definitive note of ISO, NR settings, detail etc. (the AF101 was also set to record offspeed, which may or may not be significant)

I'm going to try to re-create this effect at BVE next week and show it to the Panasonic folks, as it is quite disconcerting and, if repeatable, would certainly raise some concerns.

Dave

Yup certainly sounds like the same ghosting that was reported on the HPX370/371

Piotr
Offline
Joined: Jun 9 2006

This whole issue of judder/strobing/ghosting (people and me mix it up and no wonder as the symptoms are similar) is stirring more emotions than I thought.
I found this article with links to some of Allan's work:
http://broadcastengineering.com/hdtv/solutions-judder-problem/
Also this tempestous thread :
http://www.dvinfo.net/forum/panasonic-avccam-ag-af100-camcorder/489109-does-af-100-line-skip.html
Seemingly it's on the sampling scheme of the AF100 sensor, but really it boils down to "how good is the AF100". Now it's at the stage when some posters showed resolution charts with 650-700 TVL (same as my experience which I thought wrong), Panasonic nervously and vehemently denies, mods announce the thread ran its course, challengers ask for Panny charts, Panny finally yields and promises to deliver...and now it's silence. Fascinating like a good thriller.
I also found that BBC setups changed from hpx301 to 371--now the coring is up. Is it connected to ghosting/judder symptoms?
Question to Alan Roberts. When can we expect the AF100 tests to be published?

MAGLINK
Offline
Joined: Mar 8 2007

Yes I also noticed that the coring on the 370 had gone up to +5 when it was 0 on the 301.

It does sound like panny have put certain filtering on the AF100/101 to keep noise down like they did on the 370/371(what new chip???), problem is that it is causing the same problems so they may have to do the same firmware release to offer two settings as they did on the 370/371, what that will do to the noise level may be too much with the AVCHD codec and on the 370/371 it took it back to 300/301 levels.

Personally I have never had a problem with noise on the 301 and with the AF101 I would rather have it slightly noisy against funny ghosting or smearing!

I had to sigh when I got the Preston's BVE release which says that the AF101 can shoot in the dark with +18db gain!

Gyr
Offline
Joined: Jan 17 2005
Piotr wrote:
Also this tempestous thread :
http://www.dvinfo.net/forum/panasonic-avccam-ag-af100-camcorder/489109-does-af-100-line-skip.html
Seemingly it's on the sampling scheme of the AF100 sensor, but really it boils down to "how good is the AF100". Now it's at the stage when some posters showed resolution charts with 650-700 TVL (same as my experience which I thought wrong), Panasonic nervously and vehemently denies, mods announce the thread ran its course, challengers ask for Panny charts, Panny finally yields and promises to deliver...and now it's silence. Fascinating like a good thriller.

You're right Piotr that is a fascinating thread. It was wierd how the Moderator said the comments of the Panasonic person were the definitive answer, so the thread was locked. Then the Moderators post was removed without explanation and the thread was open again.

It was pretty obvious to anyone that the Panasonic answer wasn't "definitive"; defensive might be a better description.

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

Bear in mind that my settings for any camera are based on what I saw and measured during the tests of that camera. Generally, I deliberately ignore previous tests on similar models, unless someone has specifically asked me to match them. So, things like coring, which don't grossly affect the performance, and wee bit lax. The higher the gain setting you intend to use, the higher the noise level will be, and so coring gets more important. It could just be that I wasn't as worried by noise in the 301 than the 371, it could be that I was more picky with the 371. I can't be sure, because I don't keep notes of my emotions.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

infocus
Offline
Joined: Jul 18 2003
Piotr wrote:
I also found that BBC setups changed from hpx301 to 371--now the coring is up. Is it connected to ghosting/judder symptoms?

It is conceivable that 0 on one camera and +5 on another represent the same amount - but the scale reference is different. Just a thought.....

I've given my general feelings about coring before. It may improve the noise performance - but it comes at a nasty price, removing the low level "subtle" detail as well as noise. Avoid it if at all possible.

I've certainly seen the "ghost" issue myself on an HPX371, but would doubt it's related to any "judder" concerns. My memories are that the "ghosts" are quite subtle on normal pictures most of the time - but can really come out to bite on any post processing. (A problem with chroma keying was one I saw demonstrated.)

I understood they were a function of inter-frame noise reduction techniques. That's quite understandable when in a camera with 1920x1080 chips of only 1/3", but I wouldn't expect such to be needed in a camera with such a large sensor, surely?

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

Knowing how Panasonic cameras develop, I'd be amazed if the DSP for the 371 is different from the 301. So, In theory, you can use the settings from one for the other and get the same (or very similar) results. The same's true for Sony, settings in the PMW500 work in the 350 and 320, and in the PDW700, and for Canon (the XFs).

I have no detailed notes for my choosing of different coring levels, but I can surmise that I'm getting more picky about what signal processing does, and wasn't happy about the subtle static detail loss that noise reduction does.

Don't read too much into this, there's nothing suspicious or secret. The settings I derive are done one the day, and usually in a rather short test. When I started doing this, I was taking 3 days to rummage one camera. Now, I'm often asked to do 3 cameras in a day. The only way I can do that is to make assumptions about the controls and check, rather than measure everything, which is what I used to do and still vastly prefer to do. So, the emphasis has changed; in the early days, I was only trying to get best pictures, these days the tests are much more to do with whether the camera is acceptable or not (e.g. Canon 5D), the settings come as a side benefit. So, regard the settings as a starting point, advisory, but read the notes in the documents which usually explain how I came to end up with those settings.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

drc_online
Offline
Joined: Apr 19 2010
Example image of ghosting

Hi all,

So I didn't make it to BVE due to illness, but here's a sample frame of what I was seeing.
This is with a 50th shutter, camera set to 1080p25, but recorded externally on a NanoFlash as 1080i50.

(I've looked at the in-camera clips recorded at 50fps and they are fine - i.e. they don't have this artefact)

Notice the striping that indicates it's some kind of interlace artefact, but remember the camera thinks it's in 25p mode.
So it looks very much like this is to do with how the camera creates an interlaced output from a 25p sensor output.

..and in case anyone asks, no I don't have permission to post video I'm afraid.

I would be very happy to know this is normal 25p recorded as 50i artefacts, as it would put my mind at rest.

Dave

Dave R Smith
Offline
Joined: May 10 2005
drc_online wrote:
Hi all,

... here's a sample frame of what I was seeing.

Dave

forgotten something??

drc_online
Offline
Joined: Apr 19 2010
Dave R Smith wrote:
forgotten something??

Eh, no Dave - sample image showing up fine in my browser - anyone else not see it?

Here's a direct link
http://dl.dropbox.com/u/1107736/doubleimage3.jpg

Anyway, here's an update - I looked again at the in-camera recording and there is indeed a little bit of a double image on movement - it's pretty subtle, and nothing like what i've shown above (which I'm now convinced is a setup issue between the AF101 and NF), but it is there.

It should be visible in this image around the nose and mouth.

http://dl.dropbox.com/u/1107736/50fpsdoubleimage2.jpg

Dave

Dave R Smith
Offline
Joined: May 10 2005

I see no link or evidence of mis-typed link in earlier post, and later post I can see links but the web page is white screen with url typed within it.
Possibly my anti-virus controls are preventing background activation scripts from working or its a permission issue?

mooblie
mooblie's picture
Offline
Joined: Apr 27 2001
drc_online wrote:
Eh, no Dave - sample image showing up fine in my browser - anyone else not see it?

Sorry: I can't see anything, Dave, either. Neither embedded in your post, nor by clicking on the external link.

Both of your "http://dl.dropbox.com/u/1107736/doubleimage3.jpg" and "http://dl.dropbox.com/u/1107736/50fpsdoubleimage2.jpg" URLs seem to be broken.

Martin - DVdoctor in moderation. Everyone is entitled to my opinion.

drc_online
Offline
Joined: Apr 19 2010
mooblie wrote:
Sorry: I can't see anything, Dave, either. Neither embedded in your post, nor by clicking on the external link.

Both of your "http://dl.dropbox.com/u/1107736/doubleimage3.jpg" and "http://dl.dropbox.com/u/1107736/50fpsdoubleimage2.jpg" URLs seem to be broken.

OK, wierd.
Let's try this one (for the in camera recording)

http://db.tt/X28QRuj

and this one for the nanoflash recording

http://db.tt/X28QRuj

Dave

Dave R Smith
Offline
Joined: May 10 2005
Dave R Smith wrote:
I can see links but the web page is white screen with url typed within it.

..still the same.
Normally view/page source lets me see the web page code - e.g. this website - but for the link, it's greyed out disabled.
As its a drop box, I'm guessing it's accessed via a logon/cookie you have, but we don't.

Access your end via a different browser may reveal problem to yourself.

drc_online
Offline
Joined: Apr 19 2010

These SHOULD be links to my public folder which doesn't need a login.

If anyone wants to see them, please PM me and I'll email you the files directly.

Sorry about that.

Dave

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

Still can't see any images at those locations.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

foxvideo
foxvideo's picture
Offline
Joined: Sep 9 1999

Alan, on the dropbox link - right click and save as.

Dave Farrants Fox Video Editing

Duncan Craig
Offline
Joined: Nov 19 2008

I can see the problems for sure, OSX & SAFARI.

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

OK, got it. yes, there's definitely some inter-frame processing going on there.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

infocus
Offline
Joined: Jul 18 2003

I also had problems opening the images – Photoshop reported a “bad file format”. Another image programme did manage it, and the results you’re seeing seem a bit weird.

It doesn’t seem like the problem that exists with the HPX371, since in your image the effect seems to be more one of a double exposure, whereas with the HPX371 there is one main image with a “noise ghost” where there is movement. What your images show is also far more “in your face” than the 371 fault – the latter can be quite hard to spot in normal material, but can come out to bite when post processing is tried.

It looks as if the camera is generating 50 pictures every second but the 25p mode is recording two – a double exposure – every 1/25 sec? It would be useful if you could post 2 or 3 successive frames?

This obviously raises all sorts of questions. It seems far from correct – but the question must be whether it’s a mis-set menu item in your camera, or an inherent fault which affects all units. Hopefully the former – but if the latter it would explain Piotrs initial observations about the “judder” being far worse looking with the AF101 than any other 25p camera. My initial reaction was 25p is 25p - but if every frame is actually a double exposure.........

Dave Jervis
Offline
Joined: Feb 21 2006

I know I am stepping into territory where there are far better informed people than me thinking about this.... but I would like to know what the source of the light was when that shot was taken.... (I also agree that 2 or 3 successive frames might help....).
dave

Piotr
Offline
Joined: Jun 9 2006

I am following this with deep interest from far away without access to the camera.
Meanwhile two sites with sections on Af100 closed/deleted whole threads raising serious questions about the AF100. One was on the real measured resolution where the Panasonic rep first promised to deliver evidence substantiating 800TVL (the actual topic was on sampling scheme), then rescinded, then the mod said "case closed". Another thread on a different site signalled a bizzarre flare according to the author caused not by a lens but rather an internal reflections in the camera body. Here the thread just disappeared without any explanation. The original poster presented his claims first in a linked blog though:
http://af100info.wordpress.com/2011/02/20/is-panasonic-af100-haunted/

Plot thickens.

Dave R Smith
Offline
Joined: May 10 2005

Haven't seen photo's but..

infocus wrote:
– but the question must be whether it’s a mis-set menu item in your camera, or ......

In view of extensive/hidden menus, a 'reset' to default may be in order..unless of course the default settings are where the fly comes in.

Piotr wrote:
... The original poster presented his claims first in a linked blog though:
http://af100info.wordpress.com/2011/02/20/is-panasonic-af100-haunted/

Plot thickens.

In above link, the delay/ghosting is >1/25 and perhaps 1 sec or more, so if ghosting on Daves camera is for a fraction of a second it's a different issue - unless berighter the source, the longer the ghost affect as candle is a fairly extreme test.

Suggestion:
Suspend torch say 3m from camera from 2 bits of string and oscilate torch from side to side. Ghosting extent should help show how many frames worth of 'problem' there is, be it sensor overload, frame superimposition, internal flare.

Maybe do 2 tests with lens zoomed at each extreme and 2 more tests with iris fully open/closed to see if it influences internal flare/ghosting.

I reccomend you report this fault without delay to m/f - assuming you're within g'tee period. If you decide to keep/tolerate later on, fair enough, but delayed reporting could impede case for repair/refund.

MAGLINK
Offline
Joined: Mar 8 2007
Piotr wrote:
I am following this with deep interest from far away without access to the camera.
Meanwhile two sites with sections on Af100 closed/deleted whole threads raising serious questions about the AF100. One was on the real measured resolution where the Panasonic rep first promised to deliver evidence substantiating 800TVL (the actual topic was on sampling scheme), then rescinded, then the mod said "case closed". Another thread on a different site signalled a bizzarre flare according to the author caused not by a lens but rather an internal reflections in the camera body. Here the thread just disappeared without any explanation. The original poster presented his claims first in a linked blog though:
http://af100info.wordpress.com/2011/02/20/is-panasonic-af100-haunted/

Plot thickens.

Yes NAB is coming so the USA censoring police are on the case, it all went very quiet about the so called new sensor on the HPX370/371 when it was found that all they had done was slap a filter on which made the camera worse, a magic firmware release later that gave you a HPX300/301 again and all talk of new sensors was off the agenda!

Panny have to be really careful here as trying to hype new camera's will cause them to get their fingers seriously burnt!

SimonMW
Offline
Joined: Nov 16 2004

I find myself constantly unimpressed with Panasonic as a company as a direct result of episodes like the one Gary mentions. They are always coy about the specifics of their cameras. With regard to the measured resolution (and more importantly the aliasing issue that presented itself) they still haven't released their tests, and I doubt that they ever will.

As a company they need to be far more open, and if they are asked technical questions or presented with evidence that is contrary to their claims they should deal with it. Instead they throw a tantrum and treat people like an enemy of the state.

With regard to the DVinfo thread Jan Crittendon completely ignored the specifics of what the chart was showing her. Nobody said it was a crap camera. They just wanted to know why it wasn't meeting the specifications of their claims. Those being that they had completely eliminated the aliasing (clearly they haven't) and that it is a 1080p camera (recorded format yes, but certainly not on the actual resolution front!)

She wouldn't comment on how the picture was being obtained from the sensor, if as she claimed it wasn't performing any line skipping or pixel binning of any sort. So why the poor performance if it is using such a supposedly efficient OLPF and isn't doing any pixel binning etc? It should be an extremely simple question to answer if they want to be honest with their customers. In contrast Sony are usually very open about this sort of thing, which is one reason I much prefer them as a company.

Quote:
Panny have to be really careful here as trying to hype new camera's will cause them to get their fingers seriously burnt!

I think they are well past that. They need to start to realise that we live in 2011 and they can't just sweep things under the carpet and behave in the closed manner that they are used to.

infocus
Offline
Joined: Jul 18 2003
Piotr wrote:
The original poster presented his claims first in a linked blog though:
http://af100info.wordpress.com/2011/02/20/is-panasonic-af100-haunted/

Plot thickens.

I'd not seen that before, and when I followed your link tended to first just think "so, a bit of lens flare, so what? Silly to blame a camera for a lens problem!"

But then you read on and - well, indeed the plot thickens. The authors explanation seems to stand up very well, and the explanation of why it goes away with internal ND is very convincing - I went from a sceptic to thinking "what a well reasoned argument" in about two minutes.

I'm not sure it explains what drc_online posted though - the reflection described above seems to give an inverted second image. I believe drc_onlines problem didn't show that, rather what looked like a basic double exposure? (I can only open the images on another computer.)

MAGLINK
Offline
Joined: Mar 8 2007

I would agree with you totally Simon but when I had a Z7 and an S270 I had even more problems with sony plug-ins not working with FCP and long gop HDV causing serious problems with our on-line streaming, sony never sorted either out so I switched to panasonic.

I think a lot of the problem is the prosumer aspect of these camera's but I was seen as the anti christ when I questioned the new sensor aspect of the 370/371, just glad I stuck with my 301 as I could have wasted a lot of money upgrading.

drc_online
Offline
Joined: Apr 19 2010
Ignore really bad image - wrong settings!

OK since I started this I thought I better chip in again.

I can confirm that the really bad double image with interlace banding I was seeing was a direct result of having the wrong setting for PSF handling in the NanoFlash and nothing to do with the AF101.

There may be other issues but I don't want to be contributing to any scaremongering for something that isn't the camera.

We are getting a camera soon and will be able to do more detailed testing.

I think the key here is to have realistic expectations. For a £3500 camera with a large sensor the AF101 does pretty well and the usability improvements beat shooting on a HDSLR every time.

However I now believe it is a mistake to pimp it up with PL lenses, rails, etc. Etc. and expect it to perform like an Arri Alexa.

We are going to keep our rig super simple and light and see how that works out.

Dave

Richard Payne
Offline
Joined: Sep 15 2000

I now have my demo production AF-101 and when I get the chance and a certain camera expert 'who shall remain nameless' has a little less work on, I'll try to get to the bottom of this.

All I can say is that in normal use I have not seen what has been described here.

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

It makes sense to query the Nanoflash settings, because that could cause what I saw in the example. the report and setting doc is with Canon for comment. I completed it, sent it to them and said that unless I heard to the contrary, I'd publish it (last Friday). the quick reply was 'we want to discuss it in-house first' which is perfectly reasonable. I've not been nasty about anything, just picked up one thing which may affect some specialist users, but not many. In the meantime, I can tell you all that you can use the settings from the 305 and they'll do pretty well.

If the camera's set to interlace and the Nano to progressive, that's the effect I'd expect. It hadn't occurred to me that such a mismatch could happen. But it reinforces my belief that I'm going to have to get a Nanoflash, I've already set wheels in motion, and hope to have one before the end of this financial year, for lots of good reasons :)

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

infocus
Offline
Joined: Jul 18 2003
Alan Roberts wrote:
It makes sense to query the Nanoflash settings, because that could cause what I saw in the example.

One of drc_onlines examples showed interlace - but we've already agreed that that was a red herring - a mis-set. It's the others - both in-camera and via nanoFlash - that we're talking about, the ones that show what I can best describe as a "double exposure" on movement. That's being seen on non-nanoFlash footage.

Any chance of posting about three successive frames around the problem, drc_online?

Quote:
....the report and setting doc is with Canon for comment. I completed it, sent it to them and said that unless I heard to the contrary, I'd publish it (last Friday). the quick reply was 'we want to discuss it in-house first' which is perfectly reasonable.

It's the (Panasonic) AF101 that this issue has been seen in, and is being discussed here - I believe you may be getting confused with the (Canon) XF100/105?

infocus
Offline
Joined: Jul 18 2003
drc_online wrote:
I think the key here is to have realistic expectations. For a £3500 camera with a large sensor the AF101 does pretty well ..........

I agree wholeheartedly with the first sentence, and indeed much of the trouble may be down to some people having hopes of an Alexa on the cheap. Then being in a state of denial when it turns out not to be so.

But by the time you factor in the cost of lenses etc, the AF101 cost is comparable to an EX1, so is it not therefore realistic to expect performance to that sort of level? (Prestons are selling an EX1 - not EX1R - for about £400 more than a AF101 body only.)

The EX1 has the (practical) advantages of servo zoom, true manual operation, better codec etc - the AF101 has dof of a large format sensor. Given the comparable cost, wouldn't you at least expect comparable technical performance to an EX?

Jhoge
Offline
Joined: Feb 24 2011
Piotr wrote:
I notice a worrying amont of judder with AF100 when panning.
It is as usual pronounced on contrasty verticals, but seems inordinately high even at slow pan speeds. Can it be connected to a particular sampling scheme used on the sensor? Or some setup configuration? The shutter is half and camera in 1080/25p, detail -2, film mode, but judder seems too high for me on slow pans. Any gives?

I thought I posted here last night, but it must have gone into the deep void of cyberspace.
We have the exact problem you describe Piotr. In 25P shutter off, any movement is choppy/juddery/stroby hard to put into words.

It looks like there is no motion blur at all, like shooting on 90 degree shutter.

It´s actually so bad that the production I´m on decided to either change cameras or change to interlaced, because we can´t broadcast 25p the way it looks now.

It´s not so easy to see the problem on a LCD monitor, I was first made aware of the problem form our editors when they watched the footage on Panasonic reference monitors. I then took the footage home to watch on my 42" plasma and I could clearly see the judder. It looks really bad and so far we have not found away around it.

If it was only me and my camera on this production i would have thought it was me, settings or my camera that caused it, but we are 2 cameramen with each our own 101, and we both get the same problems

Jhoge
Offline
Joined: Feb 24 2011
infocus wrote:
I agree wholeheartedly with the first sentence, and indeed much of the trouble may be down to some people having hopes of an Alexa on the cheap. Then being in a state of denial when it turns out not to be so.

But by the time you factor in the cost of lenses etc, the AF101 cost is comparable to an EX1, so is it not therefore realistic to expect performance to that sort of level? (Prestons are selling an EX1 - not EX1R - for about £400 more than a AF101 body only.)

The EX1 has the (practical) advantages of servo zoom, true manual operation, better codec etc - the AF101 has dof of a large format sensor. Given the comparable cost, wouldn't you at least expect comparable technical performance to an EX?

I´ve shot 25P on my EX-3 since the day it came out now and have had no issues with 25P, even for "reality" like documentaries.
The 101 25P just looks really bad, at least at the moment. I really hope some one could give me a setting to make it work, but I doubt it´s down to settings at this point

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

Sorry, yes, the 101. I'm tired after 4 very long work days involving a hell of a lot of driving, and there's a lot more to come,. I'll shut up. But the 101 report is also with the BBC and Panasonic, waiting for them to agree on it. So I can't comment properly on the camera yet, not until the end of next week at the earliest.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

MAGLINK
Offline
Joined: Mar 8 2007

I think a lot of the problems are as said that people are thinking that the AF101 is going to be an F3 or an Alexa, OK the codec is not there but people now think that anything with a large sensor can do full shallow DOF and with the addition of a nanoflash or full recorder then all will be OK.

Sadly it isn't going to be so and as we are seeing the AF101 may have some of the attributes of a DSLR to remove all of the associated problems means that compromises have to be made in other areas.

I suppose you get what you pay for at the end of the day but with the hype of DSLR's and their use to do HD has blurred the marketplace regarding camera's and with most of the wanabe DOP's wanting RED or Alexa for less than £5k it is no wonder that this camera may have a very short lived shelf life.

I found it very interesting that Preston's at BVE did not seem to be demo-ing it or pushing it in the same mode as their pre show hype as a camera that can "shoot in the dark with +18db's of gain" ??????????

Jhoge
Offline
Joined: Feb 24 2011

Well, I never had any hope that the 101 would be in the league of an F3 and certainly not an Alexa. I did however hope that it would be on par with the Sony EX cameras and that I would be able to shoot 25P without any problems. Now it looks like that´s not possible., and that´s dissapointing

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

I had a chance, today, to look at a longish AF101 shoot, with lots of motion, on a 50" plasma from very closeup. We had it on a HDCAMSR and could hold and examine frames. No sign of any inter-frame problems at all, everything clean as a whistle (subject to the usual caveats about spatial aliases, using too short a DoF so that the tips of the ears were in focus but not the eyes, the presenter wobbling about ads well as the camera so that nothing stayed in focus and the Bokeh effect on the railings behind was quite disturbing, and the varying angle of lean of verticals as the motion tracked past detail, and so on).

I suspect the BBC's directive is going to be that it will be allowed, but only on a tripod.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

infocus
Offline
Joined: Jul 18 2003
Alan Roberts wrote:
I suspect the BBC's directive is going to be that it will be allowed, but only on a tripod.

I thought one of the planks of the EBU/BBC requirements was full 1080p resolution? (With exceptions made when no viable alternative, such as for a Varicam.)

Since this camera doesn't even make full raster 720p, let alone 1080, why may the BBC allow it?

Can I also ask what setting was the material you saw - 1080i/25, 1080p/25?

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

Consider this.

Not all HD cameras need to make full spec HD. If the programme is not ever going to be made available to another broadcaster for sale, e.g. News, then full HD is much less important.

Even top-end drama coproductions have the 20% getout clause, where non-full-spec HD can be used, called as upconversion.

Productions where it is impossible to fit or use a full-spec camera for physical or safety reasons can use anything that will make a picture, e,.g. Top Gear in-car cameras.

There are lots of cases where a camera is perfectly acceptable without being full spec.

Mind you, I don't make the rules, I only measure and report.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

infocus
Offline
Joined: Jul 18 2003

All points fully noted. It does therefore sound as if the implication is that the allowance for use will be "subject to the upconversion rules", that's quite a different kettle of fish to unrestricted approval. (Which is what I think interests a lot of people.)

Can I check what the parameters of the material you saw today were? I'm assuming 1080, but 1080i/25, 1080p/25, or 1080psf/25? I think as regards the comments about judder in this thread, that the system is an important variable. The reports indicate it being a problem in some formats, not in others.

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

It was 1080psf/25, recorded on Nanoflash. I suspect the problems being reported are down to finger trouble, getting mismatched settings between the Nano and the camera.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

Piotr
Offline
Joined: Jun 9 2006
AF101 clip with judder (?)

Back from holidays I recorded today this short clip showing some of the behaviour I wrote about. Please download the 60MB file for 1080p viewing. All parameters described therein.
Any comments appreciated.

http://vimeo.com/20429696

Piotr
Offline
Joined: Jun 9 2006

nobody? nothing? everything ok? judder within the norm?

Duncan Craig
Offline
Joined: Nov 19 2008

I downloaded the file, my Mac wasn't playing it very well. It seemed to skip frames. So I can't comment on judder, but a cleaner and shorter shot would be appreciated, it took ages to download I only got 40kbps down.

Jhoge
Offline
Joined: Feb 24 2011

I saw the judder, and I commented on it, but over at dvxuser.
I´m sure it looks much worse on a TV though

infocus
Offline
Joined: Jul 18 2003
Piotr wrote:
nobody? nothing? everything ok? judder within the norm?

Similar response to Duncan - can't really be sure if any juddering is down to what I currently have available for playback.

Can I ask what you and Jhoge have to say to Alans last thought:

Alan Roberts wrote:
I suspect the problems being reported are down to finger trouble, getting mismatched settings between the Nano and the camera.

Were you both using nanoFlash when you experienced the problem?

Piotr
Offline
Joined: Jun 9 2006
Duncan Craig wrote:
I downloaded the file, my Mac wasn't playing it very well. It seemed to skip frames. So I can't comment on judder, but a cleaner and shorter shot would be appreciated, it took ages to download I only got 40kbps down.

I can't make it really any shorter, because one has to be able to compare low-contrast sequence with the high-contrast one when windows enter the frame. You need a fairly strong machine to replay 1080/25p mpeg4 codec. Also I suggest that the clip is viewed as 1920x1080 mapped 1:1 onto your display, as I noticed that any rescaling affects the strength of the phenomenon.
I wonder if the SD card file ingest into the Apple software does anything to the sequence/codec. It used to be that eg. HDV was transoded into some intermediary codec 4x longer to be able to access each frame of the long GOP individually. Not sure how it's done now. Also how Vimeo treats the incoming 1080p file. I can only confirm that the original file and the final file downloaded from Vimeo are the same volume in megabytes.

infocus
Offline
Joined: Jul 18 2003

It occurs to me that one way I could easily and reliably watch such footage is if I could download the original AVC-HD file folder from the SD card (with just the one shot on it), then burn it to an SD card, and playback directly using the SD card slot in the TV. That would also distinguish whether any problem being seen is inherent to the AF101, or a camera-NLE mismatch.

Piotr
Offline
Joined: Jun 9 2006
infocus wrote:
Similar response to Duncan - can't really be sure if any juddering is down to what I currently have available for playback.

Can I ask what you and Jhoge have to say to Alans last thought:

Were you both using nanoFlash when you experienced the problem?

Just as described in the Vimeo post, I was using the SD card only. Also all settings are listed there.

Piotr
Offline
Joined: Jun 9 2006
New developments

perhaps we are getting somewhere with this phenomenon at dvxuser....

http://www.dvxuser.com/V6/showthread.php?238369-First-Impression-serious-drawbacks/page7

let me quote here my last post there:

QUOTE
LPowell, this is the first plausible hypothesis to what might be really happening. I hope it can be verified through further experiments. Thanks.

Meanwile here are new developments.

Firstly thanks to the generous offer of Fohdeesha, my "AF100 judder" clips* are now available on his FTP server.
Go to :
http://fohdeesha.com/data/other/public/
You'll see four clips/links for download. The top two "AF101PANLONG" is the longer 65 second/170MB series of panaromas of varying speed. The MTS container is the original camera file. The MOV container is the same file re-wrapped into Quicktime by ClipWrap (divergentmedia.com) . This does not change the sample structure of the original video file (no transcoding), it's the same codec with a different container that allows it to watch in QT.
You will need a strong machine to do it without problems, best a recent (2010-2011) multicore Intel Mac with Snow Leopard/ QT X and/or hardware acceleration. H264 is that demanding. Go to Divergentmedia site for an excellent tutorial .**
The bottom two clips at the FTP site is the shorter (25sec/70MB) single slow panaroma, again in MTS and MOV versions.

First I watched the clips on my home system (iMac i7 /8GB + EIZO cg243W coloredge in REC709 preset) and played from AF101 (original card/file) via HDMI onto my HD plasma. When watching on a computer display it's important to have the clip mapped 1:1 onto 1920x1080 as I noticed rescaling changes the strength of the phenomenon.
Today I projected the clip again from AF101 via HDMi and Panasonic PT3000 HD projector in the Cinema1 preset. Screen diagonal 120 inches (3 meters)

I found out the seeming "judder" (maybe it's the "farkle"?) depends on display type and viewing distance. It's strongest on the computer display (although the cg243 is also used as reference monitor for rec709 video), medium on the plasma, and weakest via projection. That also corresponds to the typical viewing distance: less than screen diagonal with the computer, more than a diagonal with plasma and projection. But the distance seems to be secondary factor to the replay/display method. What really made me wonder is the virtual absence of "judder/farkle" from my clips via projection. But in its place I noticed on the screen a strong chromatic aberration (greenish on one side of the window framing, purplish on the other). This is very unusual, as this lens (Zuiko SHG 14-35/2) is known for very low CA, virtually non-existent at f7. Any other explanation? Those greenish/purplish tints also seem to be "ringing" (changing intensity) with the movement. Is it in line with the explanation LPowell gave? (below quoted). Strangely enough the "CA"/"non-CA" is less visible on the plasma and I missed it first on the computer. There the phenomenon looks more like a typical judder, not a "ringing of CA artefacts".

That's where I am today with all this. LPowell's hypothesis is the most promising explanation , and my projection observation of disappearance of judder and appearance in its place of "CA ringing" the most intriguing--for me. I still plan to conduct comparative tests with AF101 and HPX3100 some time this month. Hopefully we are getting closer to an explanation.

* I placed the clips first at Vimeo with all setup parameters
http://vimeo.com/20429696

** http://www.divergentmedia.com/blog/f...ing_on_the_mac

http://www.divergentmedia.com/blog/f...oduction_codec

Quote Originally Posted by lpowell View Post
I downloaded the MP4 file from Vimeo and imported it into a Premiere Pro CS5 25p sequence. At 100% magnification, I see a prominent flicker on the edges of the potted plant as it scrolls across the frame, both at 25p and single-stepping through the frames. I've seen this distracting effect numerous times on high-contrast scrolling shots I've taken in daylight, particularly when shooting at high shutter speeds.

As an experiment, I displayed the video at 200% magnification in Premiere Pro (with Full Playback Resolution). This dramatically reduced the flickering effect, though I could still detect small traces of flickering on the edges of the plant stalks. It's obvious to my eyes that Premiere uses a smoothing filter at 200% magnification that reduces the contrast on sharp edges.

I believe this flicker is a side-effect of the sharpening filter used in the demosaicing algorithm the camera uses to convert the RAW Bayer Filter output into RGB pixels. When an object with a high-contrast edge moves across the frame, its alignment with the image sensor RGBG grid varies slightly with each frame. As a result, the calculated RGB values of the pixels on the edges can vary noticeably from frame to frame. In the early days of low-resolution 3D computer animation, we used to call these edge effects "farkle" to distinguish them from the flicker due to low frame rates.

If you single-step through the video, you'll see that this edge flickering is only apparent in the instant that one frame is replaced by the next - it does not persist for the entire duration of each frame. In addition to the sharpening filter in the camera, the response time of the video monitor will have an effect on the flicker. Unfortunately, hi-res displays with faster response times (such as LCD's designed for gaming) will make the flicker appear more prominent. Slower displays will tend to smooth the high-contrast edges and minimize the flickering.

There are two approaches that may help suppress this flickering effect. When shooting high-contrast moving subjects, using slow shutter speeds and minimum in-camera sharpening should help smooth out the sharp edges. In addition, post-production smoothing filters can be used to selectively blend problematic edges into the background. I'm sure the gross improvement I saw using Premiere's 200% magnification could be done with far more skill and subtlety.

UNQUOTE

What do you think of the LPowell "Farkle/DeBayer artefact" hypothesis ?
How about the CA ringing visible in projection?

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

Aha, some light appears. I've looked at lots of stuff from this camera and never seen the problem. But I looked at it on proper TV displays. Looking on a computer editor display will inevitably involve the display doing a standards conversion. TV displays show the images at the image rate of the signal (e.g. 50Hz), but computer displays run at a fixed image rate, usually 60Hz or higher, so viewing video this way will always result in a frame-rate conversion, and they're the hardest thing to do right in this business.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

delphiplasma
Offline
Joined: Dec 5 2008

Alan got there before me (I'll try and confuse the issue). Quite right, with the refresh rate of 50hz or 60hz. Judder is very discernable on 24P footage. Displays need to be compatible. Most Blu Ray titles have a raw 24P on them.
I've only just been reading up more into 'judder' as I've noticed the odd Blu-ray film with 'Judder' effect.
Either I'm very sensitive too it (As no one else seems to see it), or the display isn't 24P compatible?
From what I hear, it is very difficult to exclude from the filming process.

Piotr
Offline
Joined: Jun 9 2006

Let's just keep in mind that my footage is 25p (25 fps).
What would be display frequency of:
-computer screen
-plasma
-HD projector
in our 50Hz part of world?

Aren't plasma displays 60Hz by default, whereever they sell?
Computers most likely are 60HZ (or 59,8) all over the world.
I understand we talk about vertical frequency here.
Not sure what projectors (mine is panasonic pt3000) do...

How all that affects the perception of movement?

StevenBagley
Offline
Joined: Aug 14 2000
Piotr wrote:
Let's just keep in mind that my footage is 25p (25 fps).
What would be display frequency of:
-computer screen

Varies. But I suspect most LCDs will default to 60Hz

Quote:
-plasma

Varies, mine certainly does 50Hz, 60Hz and some multiple of 24p depending on the input.

Quote:
-HD projector

Same as a plasma.

For 25p footage, a proper TV (fed the footage as video, and not by a computer) will probably be showing it at 50Hz (each frame displayed twice). On a computer screen, it will be mapped to the refresh of the display (usually by dropping/repeating frames). If the display is set to 60Hz, then this will result with frames being repeated (probably in an odd pattern to help avoid it being obvious -- you need to make 12 output frames from every 5 input frames, so something similar to f1 f1 f2 f2 f2 f3 f3 f4 f4 f5 f5 f5 would work -- scarily some UK BBC Bluray disks used a similar technique to generate a 60Hz master that could be released anywhere -- they look awful, although that is in part due to the chroma ending up a field out of phase).

There's an easy to tell whether the judder is coming from the camera or the playback though -- find a sequence where the judder is very visible on playback and then step through it frame by frame in your NLE. If it is visible frame by frame, its burnt in -- if it isn't its a playback issue...

Steve

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

In general, a display sold as a TV set will refresh the screen at the incoming image rate, or a, integer multiple of it (so a PAL-land TV set will refresh at 50, 100Hz etc). Mine certainly do, and if I give them a 59.94Hz signal then they run at 59.94Hz, which is what they should do.

But displays sold for computer use will run at a fixed rate, usually 60Hz or higher, and will tell the graphics card what rate they run at. The cheapest graphics cards just throw the latest image to the display at the video image rate, so it will do a 5:6 frame-rate conversion just by duplicating frames, and will look awful (this was what the first BBC standards converter was doing in 1966/67, before we at BBC R&D took over the project and got it a bit better).

Computers are not a sensible way of looking at video, that's why I've got an HDspark in my i7 editing machine, feeding a TV display for monitoring.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

infocus
Offline
Joined: Jul 18 2003
delphiplasma wrote:
Quite right, with the refresh rate of 50hz or 60hz. Judder is very discernable on 24P footage. Displays need to be compatible. Most Blu Ray titles have a raw 24P on them.

I think we need to wind back quite a few posts. Look back at post #2 and I started off sceptical:

Quote:
Is it worse than any other 25p camera? (I don't see why it should be.)

Move forward and Piotr and others say:

Piotr wrote:
I subjectively see it as worse than on other 25p/24p acquisition systems.

The issue is not whether different displays will show judder differently (they will) but whether, *all else equal*, the AF101 has a worse judder problem than other 24/25p cameras. Piotr gave a definite "yes", then on this forum, Jhoge posted (post 44):

Jhoge wrote:
I´ve shot 25P on my EX-3 since the day it came out now and have had no issues with 25P, even for "reality" like documentaries.

The 101 25P just looks really bad, at least at the moment. I really hope some one could give me a setting to make it work, but I doubt it´s down to settings at this point.

And according to Piotrs last link, other people are saying similar things - people who, having just bought the camera, would not be expected to publicise it's failings. (Helping to do themselves out of work, at least with their new camera.) And people who are comparing it's 25p footage with 25p from other cameras on the same display equipment.

At first, I could see no reason why such could be so, but Piotrs quote gets me thinking. Piotr, I'm intending to download your file to display on my plasma, and I've also brought this subject up with the person whose zone plate tests first tipped me off to how low the resolution is. A couple of things in what lpowell says make me now begin to suspect there may be a connection.

Piotr
Offline
Joined: Jun 9 2006

On my part I will set up a controlled experiment to compare AF101 with hpx3100 both in 25p and panoraming the same scene with contrasty verticals. This is not just for the sake of it. I have a plan to shoot our next daily drama multicamera (all recording, non-switched) and all edited in post. By multicamera I mean 3-4 cams per crew, so 8-12 cameras in total working concurrently on 2-3 separate sets. These days even big productions look for savings, so ability to do it with a 10k cameras (AF101 +accessories) gives significant advantage over the traditional broadcast level 2/3 cams (say hpx3100+lens etc etc). To attempt that I must be sure of the imaging chain 100%.
PS. The DOF issue is a bonus on the AF101, but not the killer app. We learned workarounds on 2/3 imgers too (longer focal length).

infocus
Offline
Joined: Jul 18 2003
Piotr wrote:
These days even big productions look for savings, so ability to do it with a 10k cameras (AF101 +accessories) gives significant advantage over the traditional broadcast level 2/3 cams (say hpx3100+lens etc etc). To attempt that I must be sure of the imaging chain 100%.

PS. The DOF issue is a bonus on the AF101, but not the killer app.

If the DOF issue is not essential, then why not just go for something like an EX or a XF305 instead?

As far as I can see, the only feature of the AF101 that may be seen as a positive bonus over such as an EX1 is the shallow DOF due to the large format sensor. In pretty well every other respect, the EX is superior. That doesn't just apply to image quality (especially resolution), but also that out of the box it has a power servo zoom of good zoom range, good manual control of iris, focus etc, better viewfinding, and a superior native codec.

I fully appreciate what you say about cost (ability to do it with <10k cameras etc) but by the time you've bought lenses etc, there's not a great deal in it between an AF101 kitted out and a comparable EX1 outfit. So if shallow DOF isn't essential.......

In your shoes, I'd also wait for the 23 March before doing anything when we should hear more about the new large sensor NXCAM.

Jhoge
Offline
Joined: Feb 24 2011

I have to agree with infocus here. I paid pretty much the same for my AF101 as I did for my EX-3 a few years back.
I know Barry Green think the AVCHHD codec superior to XDCAM EX (and he has some shots that shows it), but just looking at wide shots with a lot of detail, the EX is much better IMHO.
The Shallow DOF on the other hand is a nice addition

Piotr
Offline
Joined: Jun 9 2006

I have time with final decisions on production cameras for the soap till the summer, so no rush. Shallow DOF is not critical (IMHO-zealots will differ) for television shows, including drama. Even a big plasma is still watched from couple of diagonals, we tell stories visually differently than those for the cinema (closer shots), so some cinematic isolation tricks are not so critical here. However it is nice to be able to use DOF once in a while, just to differentiate drama from other genres. And here 2/3 inch is a minimum to be able to do it sensibly. 1/2 or 1/3 inch imagers are just to small. Another camera I am contemplating is F3, but until it's in our hands and fully tested I cannot say anything about it. I like what I read about dynamic range, S-log etc ( esp. for exteriors) though.

infocus
Offline
Joined: Jul 18 2003
Alan Roberts wrote:
Aha, some light appears. I've looked at lots of stuff from this camera and never seen the problem. But I looked at it on proper TV displays. Looking on a computer editor display will inevitably involve the display doing a standards conversion.

I've now had the chance to critically examine one of the samples Piotr earlier posted (the shorter version, mts file format) which is the raw file straight out of the camera - so no chance of any processing/recompression etc skewing results.

I used a PS3 to replay the file directly from a hard drive, and looked at it directly via HDMI on a 42" Panasonic plasma, 1920x1080 resolution, with overscanning off. We can forget any thought of it not being a proper TV, or any standard conversion happening.

Firstly, I found it immediately possible to see the effect Piotr and others are describing. It's most noticeable (and objectionable) on high contrast edges, and the window frame edges are the easiest place to look. I agree it's more objectionable than "normal" 25p material. This is at normal (2.5m) viewing distance.

Next step was to step frame by frame through the file. The AVC-HD spec ( http://en.wikipedia.org/wiki/AVCHD - towards the end) implies that AVC-HD will not support 1080p/25 , only 1080i/25 or 1080p/24. I therefore believe Piotrs file is not 1080p/25 as such, but more likely 1080psf/25. The response of the PS3 in frame by frame stepping tends to support this, and I believe that in "PAUSE" mode it's showing a single field of the psf signal - hence a complete frame advance every two presses of the "ADVANCE" button. (So one press makes the frame change to the next, another press and the line pattern changes on near horizontal lines, whilst vertical edges remain in the same place. Second field of the same progressive frame. Exactly what would be expected with psf.)

At first it was difficult to see what was happening, but if anybody can repeat the viewing, just look (stepping field by field) at the section 20 seconds in where the curtains move across frame. Move on to a new frame, and the windows and curtains move to the left (as expected). Then move to the next field, and the left hand (brighter) part of the curtains stay still, whilst the right hand section (darker) seems to jump to catch up with them!

Quite bizarre, and frankly I've never seen anything like it.

The more I looked, the more evidence I could see of this effect. There seem to be very strange differences between the two psf fields of a progressive frame that shouldn't exist.

I can't say too strongly that I do not believe this to be just a display issue. I'm watching on a plasma TV, not a computer monitor. Equally, I've never seen anything like it on any other 25p footage on the same viewing arrangement. I was initially very sceptical about what Piotr was saying - "Is it worse than any other 25p camera? (I don't see why it should be.)" - but I'm now firmly convinced it is indeed a real issue.

StevenBagley
Offline
Joined: Aug 14 2000
infocus wrote:
I used a PS3 to replay the file directly from a hard drive, and looked at it directly via HDMI on a 42" Panasonic plasma, 1920x1080 resolution, with overscanning off. We can forget any thought of it not being a proper TV, or any standard conversion happening.

Be aware, unless told otherwise a PS3 will replay 50Hz material by standards converting (exceptionally well I might add) to 60Hz (there's an option in the menus to force 50Hz playback). However, that shouldn't affect single stepping... And I can't think for the life of me what would cause the effect you describe...

Steve

Piotr
Offline
Joined: Jun 9 2006

somebody mused before at dvxuser why the observations come only from the 25p world*. it would be good to see if they repeat at 24p. I believe however that people recording in 1080/25p prores also reported the phenomenon. that would point that there is something in how the cmos captures 25p and then processing changes it to 25psf for avchd and SDI output. SDI is psf by nature, right?
What really bothers me is why HD projection looks almost normal but the weird CA-like fringing on the edges?
(we really must be 100% sure whether replay/display is 50hz or 60hz).

*wrong. the op of "first impressions, serious drawbacks" thread there, maxfish, noted the same phenomenon in 24p. or it's a convolution of different phenomena stemming from different circumstances.

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

I'll have hands on a 101 today for a short time and will try to generate the problem. I'll report back.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

infocus
Offline
Joined: Jul 18 2003
Alan Roberts wrote:
I'll have hands on a 101 today for a short time and will try to generate the problem. I'll report back.

It may be worth downloading Piotrs file first and looking at the "ghostly curtain" effect I've described, to give you an idea of what to look for. It may also be valid to duplicate his setup settings. (Piotr - can you say whether you've spotted the problem at various settings?)

It does strike me that when Alan looked at footage via an HDCAM SR deck before, was it set to give an output frame or field in pause mode? In frame mode, it would tend to mask the effect I saw when stepping through - I now think much of the problem may lie in the differences between the two fields of the same psf frame. (Steven - I'll try to check the PS settings, but I'm pretty confident it couldn't explain the curtain issue when stepping through. Are you able to examine Piotrs files yourself?)

I will also say that Piotrs material does look very edgy, even though I note he has already set the detail level below default. Apart from the judder, I found the pronounced stairstepping on the high contrast "almost horizontals" very objectionable. I'd be inclinced to try reducing the detail level further.

Jhoge
Offline
Joined: Feb 24 2011

Interesting observation infocus. I do think something odd is going on here. I´ve just watched the first episode of a TV series we shot last year. It was broadcasted on NRK HD (BBC equivalent), it was shot on mostly EX cameras, all 25P. It had action sport, nature, run & gun docu stuff and it never looked anything like the footage from my AF-101.
I haven´t seen the 101 footage broadcasted yet, but I´ve watched it from the camera on the same 42" Panasonic Plasma.
Turns out I won´t see any 25P (at least for the moment) broadcasted from my 101. We changed to interlaced a week into production as the 25P was not acceptable when viewed on a (Panasonic) reference monitor.

Normally I would have tested a camera much more than I did with the 101 before going into production, but the produceers wanted the camera and it just came out a few days before we started

MAGLINK
Offline
Joined: Mar 8 2007
Jhoge wrote:
Normally I would have tested a camera much more than I did with the 101 before going into production, but the produceers wanted the camera and it just came out a few days before we started

Don't you just love it when this happens, I bet the same producers will be the first ones to criticise and blame anyone but themselves when it all goes wrong!

infocus
Offline
Joined: Jul 18 2003
Jhoge wrote:
Turns out I won´t see any 25P (at least for the moment) broadcasted from my 101. We changed to interlaced a week into production as the 25P was not acceptable when viewed on a (Panasonic) reference monitor.

Normally I would have tested a camera much more than I did with the 101 before going into production, but the produceers wanted the camera and it just came out a few days before we started

In your shoes, I'd be hammering on the doors of the dealer I bought it from .......

Have you or Piotr had any formal response from a dealer or Panasonic directly?

Jhoge
Offline
Joined: Feb 24 2011

The other cameraman on the production has been in contact with Panasonic, but nothing came of it yet. They ask if we have done this or that, what kind of software we use etc
As we are in the middle of production is hard to find the time.
He planned to hook his 101 up to the reference monitor at the post house where they first noticed it. Luckily that is a Panasonic monitor.
This way we can (hopefully) determine if it´s the codec or from the camera it self. I think it´s from the camera.

I bought the camera through a friend of mine on his last day at the company. They are more like a stills consumer place and just had one model in as a showcase. I will talk to Panasonic directly instead I thinkl

Piotr
Offline
Joined: Jun 9 2006

I only used the the settings I published (with detail -4 and coring +2 etc, gamma cine d, 25p etc).
I did not contact panasonic yet, fortunately my production is in the fall and I don't want to raise alarm before I know what's going on (if).

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

OK, here's what we did today.

We set the camera in 25p mode and recorded in the camera. then extracted the card and played it on my 42" plasma via the card socket, so no problem. Then took the card out an imported the files into Edius and played out via HDspark in the 245p project. The camera was set to my settings (not yet released).

What we saw, running the files at real speed, was exactly what I'd expect from a 25p camera, where images are double-shown. The motion showed the usual judder and nothing else. Playing it via the HDspark into a LG Flatron also showed exactly the same effect, entirely as expected. Stopping the play and single stepping frame showed exactly what it should, no problem.

We shot from inside my house, looking out through a window. the casements are white plastic with black neoprene seals, as contrasty as it gets. We shot at 1/50 and 1/1000 shutter. We panned slowly at first, building up to ridiculous speeds. Nothing unexpected happened.

I shall be incorporating this result into my test document.

There's nothing wrong with the camera. Richard Payne of Holdan has to original files, I have copies. I expect he'll find a way to make them public.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

infocus
Offline
Joined: Jul 18 2003
Alan Roberts wrote:
OK, here's what we did today.

....... The camera was set to my settings (not yet released).

If you didn't try the settings Piotr was using, I'm not sure that really nails it. It's still conceivable there is a problem - but it's dependent on camera settings.

I can only say that I can definitely see the problem Piotr describes on his material, but with only that available, it's difficult to be too conclusive. Have you tried looking at his examples? I'd be quite happy to bring player/material to your house at a convienient time and demonstrate what is being talked about if you wished.

StevenBagley wrote:
Be aware, unless told otherwise a PS3 will replay 50Hz material by standards converting (exceptionally well I might add) to 60Hz (there's an option in the menus to force 50Hz playback). However, that shouldn't affect single stepping... And I can't think for the life of me what would cause the effect you describe...

Thanks, wasn't aware of that. Normal play was set to "AUTO", but changing it to force 50Hz didn't make any difference at all, so AUTO seemed to correctly work out the frame rate yesterday.

As far as the field by field stepping goes, there's definitely something odd going on, and I can now see the "ghostly curtain" effect at other places as well when looking closely. Look at the stair carpet treads at the beginning. Move on to the first field of the next frame and the basic shape of the carpet moves as it should. Move to the second psf field and the carpet pattern seems to jump to catch up!

(You do need to be certain that you're looking at fields, not frames.)

Are you able to look at it via your PS, Steven?

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

It's perfectly possible to get any camera to look horrid if you tweak the settings far enough. I tested the 101 with the intention of getting settings suited to television broadcast use. And under those conditions, there's nothing wrong with the motion. I haven't the time to chase the misuse of cameras, sorry, I've got more important things to get on with.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

Piotr
Offline
Joined: Jun 9 2006

The good news: the camera is ok.
The bad news: I'm doing something wrong or something is wrong with my eyes.
Better find out.
My setups (to avoid I guess):
scene file 6, operation type film cam, rec format ph1080/25p, vfr off, deatil -4, v detail -2, coring +2, master ped -2, gamma cine-like d, matrix norm1, ISO 400, system freq 50Hz, shutter 180 deg. All other setups factory.

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

Wait a day and my settings will go on-line, then try again.

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

infocus
Offline
Joined: Jul 18 2003
infocus wrote:
Piotr, I'm intending to download your file to display on my plasma, and I've also brought this subject up with the person whose zone plate tests first tipped me off to how low the resolution is. A couple of things in what lpowell says make me now begin to suspect there may be a connection.

lpowell thought there may be a connection between in camera sharpening - "I believe this flicker is a side-effect of the sharpening filter used in the demosaicing algorithm the camera uses to convert the RAW Bayer Filter output into RGB pixels." - and the witnessed problem.

I've just had a reply from the friend I bought the subject up with, and it may be worth quoting in full:

All very interesting.

I agree there is something odd going on, but it’s difficult to pin down exactly what.

As you’re aware, it was the idea of true 1080p slo-mo that first attracted me to the camera, and it was after the shock of how poor the resolution was on the zone plates that turned me off! I often find that disconnected technical issues tend to have a common source, and I think that may be the case here. The “judder” that people are experiencing may be another symptom of the low resolution, and the way the camera processing works - not a separate issue as you suppose.

To recap, my tests indicated a significance via aliasing to about 1150x650, to the extent that it behaves similar to a 3 chip design with each sensor of those dimensions. If I could measure exactly, I think the precise vaues would be 1152x648 since those figures are a quarter of what I think the read sensor area is. (4608x2592) It’s likely that it reads the chip by binning together 16 photosites ( 4x4 block) at a time, so 12 megapixels get reduced to about .75 megapisels each of red, green and blue. So each frame starts off as a 1152x648 image. That then has to be upscaled to 1920x1080.

Taking those figures as read, the thing to first notice about the file you sent me is that the detail level appears set to a very high level – I suspect as the user tries to compensate for the low basic resolution. The edginess also seems to be quite coarse and broad, and this may be where the trouble lies. There doesn’t appear anything too untoward looking at individual frames, but as the high contrast bounday with an overenhanced edge moves along it may disturb the physiology in the eye. Hence no sign of a problem on still frames, but not so nice on movement. Have you heard of the Pulfrich Effect? Something similar may be happening here.

All that’s a guess, but it points to a way forward for investigation. Suggest that the test is repeated for lower detail settings. If I’m right, the judder look should diminish as the level of detail is turned down – though whether the picture will then look unacceptably soft remains to be seen. At very least, I see it intercutting poorly with footage from a camera with true 1920x1080 sensors. If right, this also explains why some people are seeing this badly, others less so. It’s quite likely they may have different camera settings, and displays etc will also treat the edges differently.

If that turns out to be true, then in answer to your question I don’t think it’s right to say the camera has a fault as such, rather it’s just another sign of very lackluster performance. I suspect the front end of the AF101 is little different to a stll camera except for the OLPF, same basic resolution, upscaling, processing. It’s what I’d expect if paying DSLR price, it’s not what Panasonic led me to believe for the 101.

Let me know what happens.

So - the way forward seems to be if Piotr could repeat his test, once with current settings, then with the detail turned well down.

I suspect that may reduce the judder - the question being (as my friend says) will the picture then look too soft? It may be that with this camera the optimum detail settings for static and moving images may be very different.

Alan Roberts
Alan Roberts's picture
Offline
Joined: May 3 1999

I've just updated my AF101 doc and it's gone to the BBC and Panasonic for comment. Unless they raise a valid objection, it'll go off to Daniel Browning lte this afternoon (early morning for him in Vancouver) and can be on-line by tonight.

But, I repeat, I can find nothing wrong with the camera. yes, it's easy to find settings that make it look horrid, but why would anybody want to do that?

Get my test cards document, and cards for 625, 525, 720 and 1080. Thanks to Gavin Gration for hosting them.
Camera settings documents are held by Daniel Browning and at the EBU
My book, 'Circles of Confusion' is available here.
Also EBU Tech.3335 tells how to test cameras, and R.118 tells how to use the results.

MAGLINK
Offline
Joined: Mar 8 2007
Alan Roberts wrote:
but why would anybody want to do that?

The ones who want it to shoot in the dark with +18db of gain with full autofocus!:D

infocus
Offline
Joined: Jul 18 2003
Alan Roberts wrote:
But, I repeat, I can find nothing wrong with the camera. yes, it's easy to find settings that make it look horrid, but why would anybody want to do that?

I'm interested to hear what comment you may have about my friends thoughts, in particular:

Quote:
If that turns out to be true, then in answer to your question I don’t think it’s right to say the camera has a fault as such, rather it’s just another sign of very lackluster performance. I suspect the front end of the AF101 is little different to a stll camera except for the OLPF, same basic resolution, upscaling, processing.

What he seems to be saying is that whilst the "judder" may be exacerbated by too high a detail setting, the "cure" may be to make it look too soft for normal usage - certainly if to be intercut with a true 1920x1080 camera?

Or do you think his measurements of about 1150x650 before aliasing are incorrect?

Piotr
Offline
Joined: Jun 9 2006

I am going to conduct comparative 1080/25p judder tests of Af101 with hpx3100, as I said, but please note that my Af101 detail level was already negative -4 and -2 (vertical) on the scale -7 to +7, so pretty much turned down. Not a lot of scale left to go lower.
BTW, one of my first posts here on AF101 said I measured the resolution at 650TVL, which I did not believe myself at that time.

infocus wrote:
lpowell thought there may be a connection between in camera sharpening - "I believe this flicker is a side-effect of the sharpening filter used in the demosaicing algorithm the camera uses to convert the RAW Bayer Filter output into RGB pixels." - and the witnessed problem.

I've just had a reply from the friend I bought the subject up with, and it may be worth quoting in full:

All very interesting.

I agree there is something odd going on, but it’s difficult to pin down exactly what.

As you’re aware, it was the idea of true 1080p slo-mo that first attracted me to the camera, and it was after the shock of how poor the resolution was on the zone plates that turned me off! I often find that disconnected technical issues tend to have a common source, and I think that may be the case here. The “judder” that people are experiencing may be another symptom of the low resolution, and the way the camera processing works - not a separate issue as you suppose.

To recap, my tests indicated a significance via aliasing to about 1150x650, to the extent that it behaves similar to a 3 chip design with each sensor of those dimensions. If I could measure exactly, I think the precise vaues would be 1152x648 since those figures are a quarter of what I think the read sensor area is. (4608x2592) It’s likely that it reads the chip by binning together 16 photosites ( 4x4 block) at a time, so 12 megapixels get reduced to about .75 megapisels each of red, green and blue. So each frame starts off as a 1152x648 image. That then has to be upscaled to 1920x1080.

Taking those figures as read, the thing to first notice about the file you sent me is that the detail level appears set to a very high level – I suspect as the user tries to compensate for the low basic resolution. The edginess also seems to be quite coarse and broad, and this may be where the trouble lies. There doesn’t appear anything too untoward looking at individual frames, but as the high contrast bounday with an overenhanced edge moves along it may disturb the physiology in the eye. Hence no sign of a problem on still frames, but not so nice on movement. Have you heard of the Pulfrich Effect? Something similar may be happening here.

All that’s a guess, but it points to a way forward for investigation. Suggest that the test is repeated for lower detail settings. If I’m right, the judder look should diminish as the level of detail is turned down – though whether the picture will then look unacceptably soft remains to be seen. At very least, I see it intercutting poorly with footage from a camera with true 1920x1080 sensors. If right, this also explains why some people are seeing this badly, others less so. It’s quite likely they may have different camera settings, and displays etc will also treat the edges differently.

If that turns out to be true, then in answer to your question I don’t think it’s right to say the camera has a fault as such, rather it’s just another sign of very lackluster performance. I suspect the front end of the AF101 is little different to a stll camera except for the OLPF, same basic resolution, upscaling, processing. It’s what I’d expect if paying DSLR price, it’s not what Panasonic led me to believe for the 101.

Let me know what happens.

So - the way forward seems to be if Piotr could repeat his test, once with current settings, then with the detail turned well down.

I suspect that may reduce the judder - the question being (as my friend says) will the picture then look too soft? It may be that with this camera the optimum detail settings for static and moving images may be very different.

infocus
Offline
Joined: Jul 18 2003
Piotr wrote:
......please note that my Af101 detail level was already negative -4 and -2 (vertical) on the scale -7 to +7, so pretty much turned down. Not a lot of scale left to go lower.

Looking back, I now see that yesterday morning I wrote:

"I will also say that Piotrs material does look very edgy, even though I note he has already set the detail level below default. Apart from the judder, I found the pronounced stairstepping on the high contrast "almost horizontals" very objectionable. I'd be inclinced to try reducing the detail level further.

I look forward very much to seeing footage similar to what you've already posted, but with current detail settings and much lower ones.

Another very quick test would be to shoot the same scene in 1080 and 720 modes with the same camera. If my friend is correct, I don't think the 1080 version will show any more real detail......

Richard Payne
Offline
Joined: Sep 15 2000

The footage Alan and I took yesterday, and a couple of stills from the timeline can be downloaded from here. http://www.resource.holdan.eu/AG-AF101_test.zip I made a little 15 second edit with shutter off, then at 1000. It's exported as a .m2ts file.

Piotr
Offline
Joined: Jun 9 2006
Richard Payne wrote:
The footage Alan and I took yesterday, and a couple of stills from the timeline can be downloaded from here. http://www.resource.holdan.eu/AG-AF101_test.zip I made a little 15 second edit with shutter off, then at 1000. It's exported as a .m2ts file.

I wanted to play the file but I have a problem with the m2ts container (why you chose to export in this container?). Any chance you can post the file in the original MTS container or at least rewrapped into something QT friendly (mov)?
TIA

Duncan Craig
Offline
Joined: Nov 19 2008

Footage plays very well for me on a PS3. m2ts is a great format for the PS3 by the way.
But you must have 'worked' on the footage to add your captions so making another file type should be easy enough.

It's worth saying VLC handles .m2ts files OK.

Piotr
Offline
Joined: Jun 9 2006
Duncan Craig wrote:
Footage plays very well for me on a PS3. m2ts is a great format for the PS3 by the way.
But you must have 'worked' on the footage to add your captions so making another file type should be easy enough.

It's worth saying VLC handles .m2ts files OK.

The idea should be to see the original file .mts played out of Af101 to a plasma without re-wrapping and conversion that PS3 might be doing. Otherwise we do not know what others are really watching. VLC is not a very good player BTW.

Duncan Craig
Offline
Joined: Nov 19 2008

The PS3 isn't rewrapping or converting. The file I downloaded played directly from a USB stick in the PS3. It is downscaling to my 720p plasma, but that's all.

VLC is a great player IMO. It's always the first thing I recommend when clients on windows can't handle a file I send their way. Yes, it can be a resource hog, but at least it works and it's free.

Trying to get a client to install players which constantly nag them to pay for upgrades, look like viruses to their windows machines and offer confusing error messages is a real pain. VLC is often the only simple solution.

infocus
Offline
Joined: Jul 18 2003
Duncan Craig wrote:
The PS3 isn't rewrapping or converting. The file I downloaded played directly from a USB stick in the PS3. It is downscaling to my 720p plasma, but that's all.

I haven't had the chance to look at those files yet (the household PS3 got taken back upstairs by it's owner for playing games on....) but Piotrs point is that it's necessary to view the material with no chance of anything compromising the result. As long as the 50Hz setting of the PS is correctly set (as Steven pointed out) I suspect a PS3 is as good as it gets.

Downscaling to 720 is a different matter, and if the issue boils down to detail settings, a 720 downscale could compromise any evaluation - it needs to be watched 1:1 pixel mapped on a 1920x1080 display to be at all scientific. Same in principle with adding titles if that's been done, as it would inevitably mean a decompression/recompression.

Duncan Craig
Offline
Joined: Nov 19 2008

I was simply letting people know that a PS3 can handle the files.

This issue would have to be visible throughout a post production and transmission chain to mean it was a 'real' problem. If the only people who can see the problem are AF101 owners with 1080p plasmas then Panasonic have no problem.

As someone who doesn't own the camera (I have played with it though), I'm going to offer my view on what I see, take it or leave it. It's a public forum and some people on here are starting to worry if their new camera is a lemon. There's no need to get worked up though.

I watch 1080p/i sources regularly through the PS3 downscaled to 720 so I'm used to what scaling looks like on my setup. There does appear to be, on playback and paused, a wide 'edge' to the high contrast areas like the window frame, a darker fringing. It's not CA, but something else. Perhaps it some kind of dynamic range processing going on in the camera?

I'd like to see a side by side shot with another make of camera trying to replicate the problem with these fast pans.

Duncan.

infocus
Offline
Joined: Jul 18 2003
Duncan Craig wrote:
As someone who doesn't own the camera ..... I'm going to offer my view on what I see, take it or leave it. It's a public forum and .......

Duncan, your input is valued, and for my point of view it's useful to know that when I next get access to a PS3 (ie. when it's not being used for it's intended purpose of game playing.... :)) connected HDMI to the main TV, it will work with Richards files.

The point is that for any analysis of a reported problem to be scientific, it's necessary to try to reduce the variables initially, then vary them one by one to see the result. So try to initially replicate Piotrs circumstances, with respect to camera settings, scene content, pan speed, recording method, etc - then alter variables one by one (and the first one I'd look at would be detail level).

All we are saying is that looking at unaltered camera files, 1:1 mapped to 1920x1080 takes a lot of possible variables away.

Quote:
I'd like to see a side by side shot with another make of camera trying to replicate the problem with these fast pans.

You'll always get a judder problem with fast pans and 25p recording. That's inherent to such a frame rate, and looking at fast pans won't prove anything. (Well, other than you shouldn't pan at all quickly in 25p working... ;) )

The whole point of Piotrs complaint is that he is seeing judder on quite SLOW pans, slow enough that he would not expect to see the same effect with a different 25p camera. To go back to what he said in the very first post in this thread:

Piotr wrote:
{It}..seems inordinately high even at slow pan speeds. Can it be connected to a particular sampling scheme used on the sensor? Or some setup configuration? The shutter is half and camera in 1080/25p, detail -2, film mode, but judder seems too high for me on slow pans. Any gives?
Piotr
Offline
Joined: Jun 9 2006

I looked at the file in VLC. It judders like mad, and no wonder as it takes only 3-4 seconds for the window vertical elements to travel across the frame. These are swish pans! So what it tests is indeed the old rule of minimum 7-8 second cross-frame movement in cinema. That has been tested for the past 100 years, why do it again? That 3 seconds is way too short has been proven again.
Also playing from a re-wrapped container (original MTS into m2ts on a computer player of mixed reputation--VLC...). Also the burnt in titles mean the material was recoded indeed.
I am sorry, but this just does not fulfill a minimum of any scientific comparison method.

Duncan Craig wrote:
I was simply letting people know that a PS3 can handle the files.

This issue would have to be visible throughout a post production and transmission chain to mean it was a 'real' problem. If the only people who can see the problem are AF101 owners with 1080p plasmas then Panasonic have no problem.

As someone who doesn't own the camera (I have played with it though), I'm going to offer my view on what I see, take it or leave it. It's a public forum and some people on here are starting to worry if their new camera is a lemon. There's no need to get worked up though.

I watch 1080p/i sources regularly through the PS3 downscaled to 720 so I'm used to what scaling looks like on my setup. There does appear to be, on playback and paused, a wide 'edge' to the high contrast areas like the window frame, a darker fringing. It's not CA, but something else. Perhaps it some kind of dynamic range processing going on in the camera?

I'd like to see a side by side shot with another make of camera trying to replicate the problem with these fast pans.

Duncan.