GH2 GH2 Vegas Workflow - RGB and DNXHD

S

SuperSet

Well-known member
Hey fellas,

Can some of you share your Vegas Project Properties, Workflow and Render Settings for YouTube? I've been following the settings outlined below but I don't think I'm using the correct settings for color correction or rendering.

Project Properties:
32-bit full range, Compositing Gamma (2.222)

Effects:
Studio RGB to Computer RGB plugin (just started doing this)
Levels plugin for color correction

Render:
Remove Studio RGB to Computer RGB plugin (just started doing this)
Render to DNXHD 8-bit 709
Render using Handbrake
Upload to YouTube

Does it make sense to use 8-bit instead of 32-bit to save rendering time and having to choose Compositing Gamma? The camera records in 8-bit 4:2:0 IIRC.
 
As far as using 32-bit and the color corrector to switch from Studio to Computer RGB and vice versa- this is where things can get very complicated.

To quote John Rofrano (of Creative COW fame):
You would only use 32-bit pixel format when working with 10-bit YUV input/output or when using xvYCC/x.v.Color media or if you are experiencing color banding when working with 8-bit media and you need to smooth it out. Otherwise there is no reason to use 32-bit pixel format.

The 32-bit floating point settings allow greater precision for processing video, but require significantly more processing power than working with 8-bit video so you should only use it if you "need" to, not because you "want" to.
Click to expand...

If you are working on a feature and need to precisely match color for broadcast, or if you are having trouble matching 2 cameras or something like that, I would use 32-bit to give you more power and precision while editing. I also probably wouldn't be using Vegas personally, but that's neither here nor there (I use vegas for quick jobs that have fast turn around, and Premiere for work that I intend to spend more time on due to its more powerful controls and AE integration). The bottom line is, as you pointed out, your souce is 8-bit color, and the end result (youtube) is also back to 8-bit.

Meanwhile, here is my workflow for a lecture series in which quick turnaround is key. Every Monday I record 4-6 short talks and render them out for upload to the client's youtube channel.
Using Vegas 9, Project Properties:
8-bit color.

Effects:
Color Corrector and/or Curves
(for simple contrast and CC tweaks).

Add titles, cut between cameras, save.

Render:
WMV, in appropriate HD preset
Upload wmv to youtube

Why WMV? Because I found that Vegas' built in mp4 renderer doesn't play nicely with Youtube. After upload, Youtube still re-encodes it and often added a nasty gamma shift. So, yes- rendering out to Cineform gets around that, but Youtube does not accept cineform uploads so you need to convert it with handbrake.
That's all fine and good, but my personal experience (and feel free to argue with me) is that rendering out to WMV is fast, retains all the color and definition youtube needs, and is acceptable as is for upload.

Again, this may not necessarily yield the BEST results. But for quick edits with youtube as the final destination, I think it works great.
 
Thanks for responding, man. I just did a quick test of 8-bit vs 32-bit video levels and the resulting video looked noticeably better with 32-bit so I guess I'll stick with it for now.
 
SuperSet said:
Thanks for responding, man. I just did a quick test of 8-bit vs 32-bit video levels and the resulting video looked noticeably better with 32-bit so I guess I'll stick with it for now.
Click to expand...

Noticeably after upload to Youtube? Because they are crushing it back down to 8 in the end.
 
M. Gilden-Cool, didn't realize youtube didn't need to re-encode WMV. I'll have to check that out!

Superset-be careful with levels in Vegas, especially 32 bit. Before you commit to one workflow, search the SCS Vegas forums because this topic, comes up OFTEN.

kairosmatt
 
kairosmatt said:
M. Gilden-Cool, didn't realize youtube didn't need to re-encode WMV. I'll have to check that out!
Click to expand...

Oh, it DOES need to be re-encoded, as Youtube uses mostly H.264 under the hood, and pretty much re-encodes everything you upload anyway. Once upon a time you could get away with encoding locally into the exact format Youtube wanted and avoid a re-encode, but it seems that isn't the case these days. So there isn't much of an advantage over one format versus another, at least, not in that regard.

Now, there are certain formats Youtube won't accept, such as Cineform. So, the workflow would require something like what the OP is describing, where he would output a final render to Cineform for archiving in high quality, then convert to mp4 with an external app. You might be asking, why not just render to mp4 directly from Vegas?

There are 2 reasons I am aware of:
1) This way you have a high quality render for converting into other formats in the future (DVD, mobile, etc) without another tedious render off the timeline.
2) For some reason, YouTube tends to slightly alter the colors and/or gamma when encoding mp4 files from Vegas directly. I never could quite figure out exactly what causes it, but I found many of my darker videos would lose a lot of their details in the shadows after YouTube got a hold of them. A particular guitarist I recorded once ended up having purple-tinged skin tones that clearly WEREN'T in my original render either.

I found a few mentions of this shifting bug on the Youtube Google support forums, but it didn't gather enough noise for anyone to have really dealt with it as a problem.

My theory is that it might have something to do with Studio vs Computer color spaces being mismatched in the final render versus what Youtube is expecting. But after much fruitless experimentation, I gave up trying to make Vegas-rendered MP4 files work perfectly with youtube, and moved on to other formats. I found that a high quality WMV render went fairly quickly off my timeline, was decent for local viewing and uploaded nicely to Youtube without anywhere near as many weird color shifting from the original. So it seems like a decent alternative that I've just stuck with for lack of a better option.

Someone smarter than me probably can (and probably should) get to the bottom of this, but for now I'm pretty content keeping my upload format in Windows Media.
 
Thanks for that description. Its a PIA wrestling Vegas' colorspaces just right, and I hadn't even noticed (re: payed attention to) the final output in youtube. One more thing to monitor....

kairosmatt
 
kairosmatt said:
Thanks for that description. Its a PIA wrestling Vegas' colorspaces just right, and I hadn't even noticed (re: payed attention to) the final output in youtube. One more thing to monitor....

kairosmatt
Click to expand...

Yep. I wonder if the reason there isn't much chatter about this is because so many Vegas users don't notice or care. My first choices are usually Vimeo for sharing and editing in Adobe CS, it wasn't until I uploaded a music video from vegas for a client and noticed the shadow details and skin tones had shifted.
 
Based on some digging in the Sony Vegas forum, there's quite a bit of discussion about YouTube and Vimeo players (not the sites themselves) expanding the color space. Hence, the gamma shift that you mentioned. And from there, I've learned to use the Histogram to stay within Studio RGB (16-235) so when this shift takes place, my highlights and shadows don't get clipped. I haven't tested this with WMV but will do it to see if it makes a difference.
BTW, one of the reasons that I encode in DNXHD and then Handbrake is that X264 tends to hold higher detail at lower bitrates than other codecs.
 
SuperSet said:
BTW, one of the reasons that I encode in DNXHD and then Handbrake is that X264 tends to hold higher detail at lower bitrates than other codecs.
Click to expand...

Interesting.
Just to make sure we're all on the same page, X264 is not a codec.
X264 is the name of a particular software library that some open source encoders use to create H.264 files. Sony Vegas can also export H.264 directly, however one may argue that Sony's encoding of it isn't as efficient as the GNU version.

That workflow makes a lot of sense if I were hosting videos on my own website- I'd want the most detail for my targeted bitrate. But if I'm having youtube convert it into their own format anyway, it doesn't matter how efficient the file that I'm uploading is. The WMV files I upload are generally 3-4x the size of the h.264 mp4s I would use for local uploads or draft samples. Youtube will compress it down to their own format and targetted bitrate anyway.

Unless of course you're dealing with a painfully slow upload connection (3G, etc). Otherwise the time it takes to upload the larger file might be less than the time it takes to export in one format and convert it again and THEN upload.
 
You must log in or register to reply here.
Back
Top