My DSLR captures at 12 bit and I would love being able to treat my stills at that color depth in Vegas...

For that level of production, you might as well use more dedicated tools for that kind of thing? You would have the money to take your cut into dedicated programs that handles effects work and color correction.

I think Vegas should be geared towards doing what it does best and geared towards its current audience. Higher bit depth doesn't really help towards either.

2- I think there's still a lot of improvement that can be done for basic things. For example, the color corrector filter doesn't work too well... it doesn't entirely get rid of color casts. The eyedropper should be changed, so that it fixes things by clicking on a neutral color. Currently it takes the use of two eyedroppers, which is unnecessary.

In terms of visual output quality, Vegas could improve the way it handles 16-235 color space... all the filters are designed for 0-255.

Color correction wise, motion tracking, better interface, and better performance would do a lot more than moving to higher than 8 bit.

3- If you look at Final Touch HD, it works in 32-bit but its output quality is hindered by some simple things. The falloff/feathering on its version of masking causes the visual appearance of bands/halos.

I'd like to see at least 2 more bits of depth so that the Gen media can get a little smoother gradient - or a floating point depth for it. Would help out a good bit.

Dave

Adobe has great demos of the benefit of higher bit depths on their current whirlwind tour of the U.S.

The difference is, shall we say, night and day. Really.

But please don't just take my word for it. Go see for yourselves, Adobe is showing this in quite a few locations across the U.S.

As more and more video cameras drop the use of the anachronism that is videotape, the need for 8-bit junior quality because of video tape capacity and bandwidth concerns just disappears.

It's not just about banding, which can be enough of an issue at times. It's also about how many effects look, and how composites appear. And more.

Glenn,
being able to simply cut 10 bits in Vegas would help no end. At the moment the best you can do is edit 8 bit and then match back in another system, pretty crude when PPro which is far from expensive can handle this task. Why would you bother if a competitors system can do the whole job.
Also a simple task such as working with BetaSP benefits from 10bit, without it the generational loss if you're doing multiple passes back to tape is pretty bad.
Thing is it's now possible to do a full 35mm 2K DI in post on affordable desktop PCs. With the performance of the second generation prosummer HD cameras pushing the envelope we're getting closer to avoiding the costs of 35mm acquisition, digital projection is finally starting to take off but even without that having a DI that's appropriate for a film out is also desirable.

I agree about the rest of it, specialist tools and specialist operators for task such as color grading. Problem I'm talking about is at the moment Vegas doesn't fit anywhere in the workflow.

As for handling the Studio RGB space better, well yes but why limit your dynamic range? If you have to conform to Studio RGB wouldn't that be better done as part of a 'mastering' process? Certainly that's the way the audio gods handle it for obvious reasons.

If motion tracking linked to masks for CC can be accomodated in Vegas then I'm all for it but there's plenty of tools that handle that nicely at the moment and many of them have a 10 bit and beyond pipeline. I'd be interested to know if the problems mentioned with FinalTouch HD are still there with 10 bit sources.

Bob.

i'm betting 16-bit color will be in V7. HDR monitors & video cards are on the imediate horizon for consumers (not just high end stuff) and with the new generation of multi-core & 64-bit it seems like the perfect time.

Of course we'll be back to 2003 render times but the picture will look better. :)

As for handling the Studio RGB space better, well yes but why limit your dynamic range? If you have to conform to Studio RGB wouldn't that be better done as part of a 'mastering' process? Certainly that's the way the audio gods handle it for obvious reasons.
This is more of an interface thing... for example:

Your DV video comes in as 16-235. This is desireable because DV will record superwhites... they give you 0.3+ stops of exposure latitude. If the DV codec decoded to 0-255 instead, you will get clipped highlights. It'll also hinder fixing white balance.

You apply Magic Bullet Editor's and increase contrast. Suddently, blacks/shadows will go below 16. If you encode to DVD here, your blacks will get clipped.

Vegas' filters have similar issues. If you increase exposure in Levels (via Input end), then that will raise black level. You can hack around this, but that involves "juggling" between two settings.

The filters can EASILY be more user-friendly so you only have to touch one setting. All that's required is some simple algebra in the interface.



Exporting to that program would be annoying. It's like what you're saying here.. you would like Vegas to do the whole job.

I'd be interested to know if the problems mentioned with FinalTouch HD are still there with 10 bit sources.
Yes. It actually has nothing to do with 10 or 8 bit and is a different sort of banding. see http://tig.colorist.org/wiki2/index.php/Human_Perception_Quirks

FT HD changes often though, so maybe they've fixed that issue now. Or not.

Coursedesign:
The difference is, shall we say, night and day. Really.
That may be because the filters are doing something else. They are probably undo-ing video's gamma correction (i.e. by taking the values to the exponent of 2.2), appling the filter, and then putting the gamma correction back in.

Vegas can do this to a degree, as filters can do 32-bit internally (i.e. the Color Corrector, MBE). If you want many of your filters working in 32-bit, then having the engine be 32-bit could be advantageous (less rounding error for starters).

2- To take your position, working in 10-bit or greater will be more for film-like capture/results. Cameras like RED, Dalsa Viper, and the Andromeda-modded DVX100 really need 10-bit (or better) for the increased exposure latitude they record.

On the other hand, how many people would be shooting with those cameras?
RED's not out yet.
The Dalsa Viper is quite expensive.
The Andromeda is somewhat expensive (about $15k for the entire system) and isn't very portable or popular.

The benefits of 10-bit and higher is not primarily in working with higher than 8-bit acquisition formats, so the availability of 10-bit+ cameras are not a primary issue (even if you will see more affordable 10-bit cameras this year).

It makes one heck of a difference in processing 8-bit footage, even if the end result is squeezed back into 8-bit again.

I think Adobe's tour shows this as well as any other of the many examples I have seen before.

I'm thinking that their tour show is even available online somewhere, and I guarantee that you could see a huge difference even after web compressions.

This is not about polishing the 14th layer of wax.

It's about Earl Scheib vs. a Mercedes factory-authorized paint job, and the difference is not dependent on if it's Columbo's old Peugeot or a new S550.

Does anyone know if the directX filter architecture supports 10-bit? From the little I know, I don't think it does. Changing the filter architecture to support 10-bit might mean that 3rd party filters won't work, possibly lower performance, and would move Vegas away from its original "premise" (no complicated hardware acceleration, open filter architecture).

There is an open source OpenFX architecture that might work.

2- I think Vegas just needs better marketing. Bob, I hear what you're saying about having a "leader" feature where Vegas does high-end work. But I think I'd rather have a leader feature which is useful to most of the user base.

The audio features in Vegas are really strong... isn't it like one or two steps away from Pro Tools? I'm not an audio person, but it seems like Vegas and Pro Tools would more or less give you the same quality output (the sound being dependent on talent, and the filters used). Vegas may have the workflow advantage in that you can change the picture edit and not have to re-conform the audio.

3- Coursedesign:

I think everyone is talking about some different yet related things here.

A- Working in the linear domain (i.e. HDR calculations are done this way):
I am guessing this is what you saw demo'ed. To see Vegas do the same thing:
Add any filter to a clip... i.e. Color Corrector, saturation = 1.6.
Add the Levels filter before the gaussian blur, and set gamma to 0.45
Add the Levels filter after the gaussian blur, and set gamma to 2.2.
The difference should be noticeable.

Another example would be film-style dissolves versus video dissolves.

Filters for Vegas can be designed so that all these calculations take place in 32-bit float within a single filter (i.e. like the color corrector, or MBE). Not as good as having rendering from filter to filter take place in more than 8-bit.

B- Rounding error (i.e. rendering in 32-bit float would eliminate this):
Start with a clip with real-world footage.
Add the color corrector, gain = 0.5
Add the color correction, gain = 2.0
Toggle the FX on/off (i.e. in the video preview window). I find it difficult to see a difference.
If you use the gradient generator instead, then the difference is easier to spot.

C- Being able to import 10-bit footage.

Glenn,
NO. Double blind tests show that Protools summing algorithms produce better sound than Vegas / Sound Forge. This as said isn't 14th layer of wax, this is lay people asked to sit and listen to the same piece mixed through various systems. Summing involves some pretty serious maths that goes much over my head but I can see why better algoritms that might be slower produce better results. It's not uncommon amongst the audio geeks to feed each track out of digital system into an analogue mixer to do the summing, it's that critical. It's not a matter of just adding two binary values, binary values between samples are derived by interpolation, then the summing occurs and then that is resampled back using decimation, at least that's my quick take on it.

When you dive into the world of digital audio and then go back to video you start to wonder, a lot. It's a good thing the eye is far more forgiving than the ear.

But back to video.

Some 10 bit acquisition systems are very expensive however the XL H1 and Wafian recorder will set you back $25K. You might add a Mini 35 and a set of primes to that but there are people here already using cameras more expensive than that. This system is recording 10 bit.

Can't speak for the USA but there's plenty of locals shooting 16mm and having that scanned, again not cheap but affordable, enough for many entries in local events such as Tropfest to be shot that way. Some will work with 10bit transfers, some Cineon.

BetaSP is a pretty common (if not dead) format. Try handling that through a 8 bit pipeline and things get ugly real quick. We've lost a few from here simply because of this. The issue is going from analogue to digital to analogue to digital and back to analogue. With only 8 bit resolution the errors add up quickly and the image degrades. Under $1000 buys you a BMD card that does it's best to minimise this by converting to 10 bit but Vegas kills it with only 8 bit output. When that's printed back to SP it goes through a D->A conversion and you can bet the exact same analogue values are not written back to tape. Take that tape and repeat the process and again another set of errors are introduced by the A->D converters.

This as i said at the outset isn't something I need, heaven forbid that I've ever got to edit BetaSP, fortunately down here it's pretty much a legacy format although there's still a few using it. But the emerging HD acquistion systems are 10bit, they're not overly expensive either. Add to that film scans and the possibility of any of us being able to make use of 10 bit is increasing. But even that's not the point, the point is that most of the other players have bitten the bullet and are able to handle 10 bit, it is being used and it's creating a marketing advantage. We could argue the merits of 10bit versus 8bit acquisition but no one could suggest that 10bit is worse than 8bit. Just look at the audio world, 16bit audio is more than good enough for most things, heck there's even audio that's only 12 bit and it's usable, but just try selling a product that's only 16 bit, sure there's a lot of hype mixed in there, who really has kit good enough to justify 32/192K but given two products at around the same price what you gonna buy, the one with the bigger numbers, yes but also if its the one used by someone who does take advantage of the bigger numbers that counts for a lot.
Bob.

Bob,
To me (and I'm just cynically) that stuff about audio summing smacks of the Monster cable effect... but you point out that there's been a double blind test done, which should set things straight. Would you happen to have the link? (If it's online.)

2- Speaking of Monster cable, the cables sound better precisely because of the marketing. I do think there is merit to 10-bit in the appropriate situations, but at the same time you gotta wonder if 10-bit is overhyped.

3- You know, even 6-bit is not that bad. A lot of computer LCDs are 6-bit (with dithering to make it look better). 6-bit is definitely noticeable, although personally it takes me some time to notice it.

On the other hand, in contrived lab situations, some observers can discern up to 11 to 12 bits. see matt cowan's post at
http://www.cinematography.net/Pages%20DW/ColorBitDepth.htm
In practice, it's unlikely that better than 8-bit output would be achieved. It still is better to capture with high exposure latitude + bit depth.

I suppose the issues are:
a- In real world situations, do audiences see better than 8-bit?
b- If so, is it really worth it to get 10-bit equipment? (i.e. is HDV good enough compared to better formats? are 6-bit LCDs good enough?)
c- Is 10bit something that editors look for, that it really is a marketing advantage?

For me, Vegas 7 will be a make-it-or-break-it release. I could be an AP user by the end of the year. Users have complained about the marketing of Vegas since I came aboard at v.3. The message is clear, the marketing isn't going to change. Sony's management appears uninterested in taking the necessary steps to keep up with FCP, Adobe, Avid or...? Meanwhile they flog the dead horse known as Xpri.

Make no mistake, I like Vegas a lot and would love to be wrong about this but Sony has only one more chance to keep me in the fold.

Just sayin'...

Glenn,
8bit 4:2:2 is probably adequate for content delivery just as 16/48K or even 16/44.1 is good enough for general audio delivery. That's a different world though to acquisition. Try bringing the lowlights up 2 stops on 8 bit video, you just don't have enough data. Same goes with audio, record 24 bits and you can happily add 20dB gain in post, gets pretty ugly though if you've only recorded 16 bits.

It's not about what we see or hear, it's about performing manipulation of the digital data, more data to work with is always very desirable.

Bob.

Yep, totally agree.

Glenn, most of your points are valid, but only really apply if the image remains as it comes out of the camera. In the case of a dramatic production where the camera may be set up for maximum range with a grading process in mind, the more bits available, the more the picture can be manipulated without degredation. This is an indisputable fact. It is precisely the reason the BBC insisted on Digitbeta for so many years before HD came along.

By all means deliver on 8-bit formats. But 10-bit gives a lot more scope for post adjustment. Even if filters can work internally at 32-bits, that makes no difference to me. I want the option of seeing 10-bits, and delivering a master in 10-bit colour. If Vegas is hardware agnostic this shouldn't be a weird request.

The most dramatic example I've seen of this is scanning 35mm slides. I can scan at 8 or 16 bit. They look pretty much the same on my 6bit LCDs but start manipulating the image in PS and THEN you see the difference.

Bob.