Kinvermark's response is spot on.

The theoretical advantage of grading in 32 bit is insignificant when impressed on 8 bit source, even when rendered >8 bit.

Just asked the same question on the Logarist thread. Yes, it is supposed to provide more precision when color correcting. I don't know all the math, but apparently even when converting 8 bit codec, rounding errors are possible and something can be lost. However, you can always try some tests to see if you notice anything.

Yes, I've heard that too; never seen QUANTIFIED tests, however...

The other thing I've noticed is that promoters of those theories universally forget to factor in the effect of quantization and dither noise on the video shadow detail. Even when I show them they disregard it.

I think in most cases your pretty fine using 8 bit projects.

You need to use a float point project in Vegas Pro though if you want to process a 10 bit pipeline from input to output (assumed you use 10 footage which Vegas Pro is capable to process as 10 bit, unfortunately Vegas Pro doesn't support that many 10 bit file types).

To me the biggest advantage of using a float point project over using an 8 bit project is for certain compositing and color correction/grading tasks where clipping may occur, then transfers and might add and also kind of ping-pong effect of clipping and compressing may happen.
In float point projects you will also see such clipping in the preview but internally (because of the technical math behind float point processing) there is no such thing as clipping and thus enables to withdraw clipping and to avoid such ping-pong effects.

Just asked the same question on the Logarist thread. Yes, it is supposed to provide more precision when color correcting. I don't know all the math, but apparently even when converting 8 bit codec, rounding errors are possible and something can be lost. However, you can always try some tests to see if you notice anything.

The Logarist is - similar to ACES - a Color Management System. Means, that it will do much more then one could do with the typical color correction Tools in Vegas - for example Transfer footage to the correct color space as decided. And here it is not about simple "rounding Errors" - but it is about to get the Colors right.

It is quite clear that with 10bit or higher footage such a System has some clear Advantages since it offers that. But there are now a lot of cameras that shoot slog with 8bit. Here you can ask if a simple LUT would not be enough - can be applied with external Tools also in Vegas. What would you do here.

You need to use a float point project in Vegas Pro though if you want to process a 10 bit pipeline from input to output.

Right - - but with 8 bit source its a little like putting mascara on a pig.

(because of the technical math behind float point processing) there is no such thing as clipping

Well, I suppose there is a theoretical bit saturation level but I don't know how one would go about filling it.

I am not a professional, my video work is for my own personal use (friends and family). My cameras provide 8 bit video only (Sony FDR-AX100, GoPro Hero 3, DJI Phantom 4 Pro, DJI OSMO, etc.). Is there any reason to use anything other than 8 bit pixel format in VP Project Properties?

Thanks in advance,

Rich in San Diego

Rich, are you shooting and delivering in 1080p or 4K? You could convert the AX100 4K files from 8bit to 1080p 10bit, Catalyst Browse does a good job, by transcoding it to XAVC-I 1080p. Not sure how well the DJI Phantom 4, OSMO and other 4K footage would hold up during transcoding to 1080p 10bit but the AX100 files do very well.

Now, is that all worth the trouble? From personal experience I tell you no, stick with 8bit. Not only is it much easier to edit you also loose the ability to "reframe" once you are down to 1080p. As they say... "been there done that".

One more thing I learned from this forum. Working with 10 bit in and out of Vegas and previewing requires 10bit capable hardware too.

OldSmoke,

I deliver in 1080P exclusively. I also shoot in 1080P with all cameras except the DJI Phantom 4 Pro where I shoot 4K so that I can crop.

I've played with transcoding to other formats based on discussions on this forum, but at the end of the day it seems a lot of extra work and I don't see a difference. This is not to say there is no difference, only that I don't see it. I wouldn't be surprised that if someone showed me the difference it would be obvious. Reminds me of "artifacts", once I knew what to look for, they are easily spotted and wonder how I ever missed them.

Rich

Rich, you've answered your own question.

Shoot an open blue sky scene in 8 bit. You will see banding.

Color correct and grade in 32 bit float.

You will see the same banding.

That's all anyone should need to know.

If you are just making straight cuts with no crossfades or any other fade fades, and no colour correction then working in 8 bit should produce a similar (but not necessarily the same) output as 32 bit. However, if you want to use crossfades and colour correction, no matter what the input or output format you will get possibly different (arguably better) results working in 32 bit. I say this as someone that only ever works with Vegas in 8 bit mode as I have found everything apart from that mode to be problematic, unpredictable, glitchy and woefully slow. But that does not detract from the simple fact that modes that work in higher than 8 bit should provide superior output through higher precision and reducing clipping. If I remember correctly, when the 32 bit mode was introduced in Vegas 8, it was the only way to do optically correct crossfades in the application, and they did look vastly superior to the crossfades in 8 bit mode.

Rich, you've answered your own question.

Shoot an open blue sky scene in 8 bit. You will see banding.

Color correct and grade in 32 bit float.

You will see the same banding.

That's all anyone should need to know.

Shot your sky in 10bit. Grade in 10bit. Render to 10bit.

You will see no banding.

That is the first point anyone should know.

The second one is that all the now available HDR displays come near to DCI P3 - typically 98% or so. And if you shoot with log you are ready for that, but 10bit or more are a must. The 32bit floating point method together with Color management approches like ACES or Logarist are the best path to be fit for the future. And the future will not be 8bit rec709 420 any more.

And the future will not be 8bit rec709 420 any more.

Need to qualify that a bit. Streaming ability for universal delivery is going to remain just there (720p) for some time. Remember that China only very recently reached 5Mbps connection capabilities, still less than needed for reliable 1080p Internet streams...

I really feel like it's one of those questions where if you have to ask, don't change the setting. You either need it or you don't, and if you need it, you will know you need it.

Most people that buy new drones and low end camera gear do not know what codec they are shooting in, much less the color depth.

Remember that China only very recently reached 5Mbps connection capabilities,

Completely wrong. You are assuming things you don't know.

My source is Akamai.

Your weak attempt at a personal attack has been reported as inappropriate.

My source is Akamai.

Your attack has been reported as abuse.

Thank you Musicvid. My source is I'm living in China for several years.

Your weak attempt at a personal personal attack has been reported as inappropriate.

Seriously????

And I bet you have really great connectivity!

if you live in a metropolitan area almost anywhere in the (developed) world you're more than likely going to have a reasonable to good, if not excellent internet connection.

for the vast majority of people living OUTSIDE of metro areas, it's another story entirely. on a 'good' day i get 2.5mbs, i have friends around the world who, like myself, live 'rurally', who share the same lackluster connections.

as for china, my wife often visits there, going to more remote areas where there's NO internet at all.

it all depends on where YOU are.

Data through Q1 2017. Source:Akamai.net

Rich, you've answered your own question.

Shoot an open blue sky scene in 8 bit. You will see banding.

Color correct and grade in 32 bit float.

You will see the same banding.

That's all anyone should need to know.

I live in Belgium, when will there ever be an open blue sky?

You need to use a float point project in Vegas Pro though if you want to process a 10 bit pipeline from input to output.

Right - - but with 8 bit source its a little like putting mascara on a pig.

(because of the technical math behind float point processing) there is no such thing as clipping

Well, I suppose there is a theoretical bit saturation level but I don't know how one would go about filling it.

Some artists like the famous Wim Delvoye (Belgian neo-conceptual artist) go further than mascara on a pig:

https://wimdelvoye.be/work/tattoo-works/tattooed-pigs/