One for Alan, Sony Hypergamma

[SimonMW]

I've just been reading a new document on the Sony website about how to use their Hypergamma modes on cameras like the F900, PDW-700, F350 etc.

Confusingly (for me) they have stated that Hypergammas 1-2 should be used if you want to output without modification to systems that have a 100% white clip, and to use Hypergammas 3-4 if you are grading.

Now, I can understand why they are saying this, because Hypergammas 3-4 go right up to 109ire. However I am still confused.

Hypergammas 1-2 not only stay within 100ire, but they peak well below it! This creates a very flat looking picture that I would not consider using without some form of grading.

1-2 appear to compress more contrast range into the signal, even though 3-4 are using more tonal range. Though I could be wrong (and often am).

Alan, have you tested all four Hypergammas? And if so, what did you find that the best uses for them would be?

[Alan Roberts]

Yes, all tested. And the instructions are correct, 1 and 2 are two different curves, both go not quite up to 100% video level but cover the full exposure range (about 600% depending on the camera), they deal with the mid-tones differently from each other, and prouce skin tones at lower video levels than would the normal 709 curve with or without knee. Therefore, they compress contrast quite a bit. 3 and 4 curves are each identical to 1 and 2, except that they are stretched vertically to fit into 109% of video range in stead of just under 100%, still with the same exposure range. Therefore they produced skin tones abit brighter.

So, curves 1 and 3 are identical except that 1 peaks near 100%, 3 is stretched to peak at 109%. 2 and 4 are also identical, with the same peak levels as 1 and 2.

There's a very good reason why 1 and 2 don't go exactly to 100%, there's a small guard space left available for the overshooting that excessive detail enhancement produces. That's absolutely normal. 3 and 4 both go right up to 109%, because they assume you know what you're doing and will set other parameters to make sense. That's one reason why I still define normal curves and knees, to give you the options.

You'll find all this covered in the BBC settings docs for the cameras.

[SimonMW]

Thanks Alan.

Its amazing there is such a difference in look between 1-2 and 3-4. I've always found that I have to correct 1-2 quite a bit in post because the exposure for skin tones is so low, and hence the picture appears very murky to my eye (hence my confusion).

Do you think it would be better, and less damaging, to shoot 3-4 and then bring the highlights down with a basic levels adjustment in post than to shoot 1-2 and stretch a rather compressed tonal range?

Also, what is the difference between these curves and the Panasonic ones? This is also why I suppose I am confused, as the Panasonic Vid Rec film curve is more punchy looking than the flatter Film Rec one, factors that appear at first to be the opposite way around for the Sony curves.

[Alan Roberts]

If you shoot 3/4 and reduce the gain in post, you'll get exactly the same as shooting 1/2 in the first place. This is the safer way to go, because you're lowering the noise although you stand a chance of getting colour contours on "important"! colours (skin tones). Shoot with 1/2 and pull up and you increase the noise and generate "missing" signal levels unless the process you use is spatial and clever (which you tend not to get in anything below Avid, evenj FCP doesn't do it right as far as I can tell).

But, if you use any form of gamma curve on it instead, then you're going to win in all directions, and I tend to do it almost shot by shot. I use the lovely Bezier curves in Edius, allows me to do exactly what I want every time.

Panasonic Video Rec uses ITU709 gamma, just like normal Sony, so they should look the same. The Panasonic Film Rec and Sony Hypergamma both aim to mimic real film stocks (over a limited contrast range). Panasonic's curve is a close match to a Fuji neg stock, can't remember which one. I'm not sure what Sony's atre based on, but the logic's the same, like in the F23, F35 and D20.

Beware of making direct comparisons between cameras though, because Sony cameras always make most highly saturated pictures than Panasonics do, which is a real pain when you try matching them together. If you really want to make comparisons, then set both cameras to the same gamma curve (ideally ITU709) and point them at a Macbeth chart. Then, go through either the Sony multimatrix or the Panasonic colour corrector (two names for the same function) and twiddle away for a while until they look the same. tedious, but the only way, a Chrom DuMonde chart won't get you there, the colours are too desaturated to explore the colour space fully.

All this is actually quite hard to do in practice, but it's one of the many things I do for money :D

[SimonMW]

It is interesting that you have mentioned the limitations of the Chroma DuMonde chart. You are the second video engineer that has told me not rely on it.

I had always been led to believe that the Video Rec gamma on the Panny's was also a cine-gamma, with a Standard gamma being for normal use? Looks like I'll have to rethink a lot of stuff.

I use the lovely Bezier curves in Edius

Curves are one reason I love Sony Vegas so much. FCP has virtually no decent beziers at all. Have to use Color for that, and unfortunately at the moment I am suffering from that particular program hanging on me all the time!

[HallmarkProductions]

Simon,

Can you tell me where these modes are in the menu structure of the XDCAM HD F350, please? I take it these are different to the cine gamma settings? I tend to use Cine Gamma 4, and -3db of gain for standard, news type shots (HQ 50i). I also find that my camera, which is fitted with a Fujinon XS17 Lens tends to overexpose a bit, and burn out the highlights when using the default gamma settings, and so have set the autoexposure to -1/2 stop to compensate. (I generally use manual exposure anyway)

Could you also please post a link to the Sony article.

I doubt I will understand it, but I will give it a read.

In summary, I find that the 350 does not give decent pictures "out of the box", unlike my trusty DSR570. I have found that we spend a lot more time playing with it to make the pictures look better - and by better, I mean "not overexposed".

Is it possible to get hold of a copy of the much-guarded setting for the F350?

Thanks,

Chris

[SimonMW]

Can you tell me where these modes are in the menu structure of the XDCAM HD F350, please?

The cinegammas are in the Paint menu.

I take it these are different to the cine gamma settings?

Nope, they are one and the same. :)

I tend to use Cine Gamma 4, and -3db of gain for standard, news type shots (HQ 50i).

Ahaaa, yes. Now according to the Sony article here (http://www.sony.co.uk/biz/view/ShowContent.action?site=biz_en_GB&contentId=1219237429204&parentFlexibleHub=1166605189820) the Cinegamma you are using (4) is for grading, while for straight video output you should be using 1 or 2.

I also find that my camera, which is fitted with a Fujinon XS17 Lens tends to overexpose a bit, and burn out the highlights when using the default gamma settings

The default STD gamma on the 350 is very linear. It can be tuned however with the knee point and knee slope adjustments to stop things burning out as abruptly.

Judging from the information about the Cinegammas it may well be better to use a modified STD gamma rather than the Cines. Also bear in mind that the 350/355/330/335 cameras are not as capable as the 2/3" cameras when it comes to handling highlight information (400% overload compared to 600%). Alan can explain it far better than my bodged understanding :)

[Alan Roberts]

Simon, sorry, yes, you're right, the Video Rec is a film type curve but is made from ITU709 with a built in knee with a smooth transition, something you can't do any other way. I've tested 31 cameras over the years (another half dozen to come in the next few weeks) and I sometimes get hazy on the fine points unless I go back to my notes.

[CJWG]

Would all of this apply with the EX3 too?

[Alan Roberts]

Not sure until I test them. Dates are fixed for November 3/4.

[CJWG]

Not sure until I test them. Dates are fixed for November 3/4.
Thanks. It would be great to hear about your findings. Looking forward to it.

[steve phillipps]

What are you testing Alan, the EX1 and EX3 at last?
Steve

[Alan Roberts]

So far, last week I saw Canon HF10, Canon 5D, Red One (build 17), Panasonic HPX171 and HVX201, Sony Z7. Next week it should be Sony EX1 and EX3, finish off the Z7, and a short retest on Panasonic HPX3000. I'm really looking forward to getting hands on a HPX2700, but there don't seem to be immediate plans for that one. Reports are written and ready on HF10, 5D, RED, HPX171, HVX201, and have gone to the manufacturers for comment before publication. Z7 report's mostly done, awaiting one more test.

The BBC's interest is in finding a replacement for Z1, getting a decent minicam for those difficult POV shots (hence HF10), and new technology top-end cameras (hence RED and 5D). The 5D supplied was an early prototype, and so is probably not representative of the final product, so we'll not be publishing this time.

Don't ask me for opinions, you won't get them until after publication. I can keep secrets :)

P.S. I was at Wood Norton yesterday, training trainers. It seems that all their camera courses are going to be HD by default from now on. And about time :D

[CJWG]

Very interesting tests coming.

[infocus]

The BBC's interest is in finding a replacement for Z1, ...........
For SD, HD or both?
Don't ask me for opinions, you won't get them until after publication. I can keep secrets :)
Without giving away any secrets, is your remit purely technical, image quality etc, or does it include such things as quality of viewfinder, ergonomics, etc?

[Alan Roberts]

1: both, as far as I'm aware.

2: no secrets to give away, apart from the written report which goes out under the BBC's authority, and therefore they are the publisher, I'm only the author. I look at the technical quality of pictures. The EBU looks at codecs. DV Solutions (Chris, Chris, Tim, Andy etc) look at sound and ergonomics. Although, of course, we all peer over each others' shoulders.