I *thought* that vimeo/you tube were expanding levels from studio to computer when they process uploads. I just tested this with a chip chart and it is not true anymore.
The levels range you upload is what you get when you download their processed video.
You remembered correctly because I got bit by this when uploading Computer RGB to YouTube and it made it much darker so I always uploaded Studio RGB after that. I'll have to do some test renders and see before I upload again. Thanks for the "heads-up".
Rocky, "The levels range you upload is what you get when you
That is completely correct. The PLAYER adjusts the levels contingent on the source (YUV or RGB), not Youtube or Vimeo. In the case of Youtube, all encodes are YUV (Y'CbCr), thus levels are mapped out accordingly on playback, right or wrong.
"They expect video that is encoded for that particular format correctly."
Meaning YUV 16-235 luminance and RGB 0-255.
Nothing has changed that I am aware of in that respect, and I expect my downloads to be exactly the same levels as the uploads.
BTW, Youtube employs Flash Player; the latest versions seem to be working as expected. When we talk about Youtube's encoding and Youtube's "Player," those are two different things.
Glenn, I'm looking at your chart and trying to figure how I should handle AVCHD on the timeline if I am rendering to DVD. In my case the source footage is 16-255. I have created a modified RGB to Studio preset in Levels that changes this to 16-235. Can I apply this at the video bus level? And woudn't this also take care of the illegal black in a 0,0,0 background event?
> "When we talk about Youtube's encoding and Youtube's "Player," those are two different things."
I guess that is what's throwing me off. They look darker when played back on YouTube but I guess the actual file is the same. It all depends on the player.
Jr,
Yep. Of course nothing is absolute, but if player, system, and video drivers are behaving, the results are predictable. A properly calibrated monitor doesn't hurt a bit. ;?)
Meaning YUV 16-235 luminance and RGB 0-255.
That's kind of misleading.
Vegas never works with Y'CbCr (which some people call YUV) directly. It doesn't make too much sense to talk about it since Vegas sends and receives RGB to/from codecs. Some codecs expect 16-235 RGB while other codecs expect 0-255 RGB. And that can change depending on whether you are in 8-bit or 32-bit mode. It's messed up and illogical but that's how it is.
Every other program out there handles the levels stuff for you. It's only Vegas which is messed up... you have to handle things manually. And Vegas is one of the few video playback applications that often does not display video correctly... in many cases it does not decode video in a way that displays the picture correctly.
Almost all video cameras record illegal levels above white level but that is a different issue.
Glenn, I'm looking at your chart and trying to figure how I should handle AVCHD on the timeline if I am rendering to DVD. In my case the source footage is 16-255. I have created a modified RGB to Studio preset in Levels that changes this to 16-235. Can I apply this at the video bus level? And woudn't this also take care of the illegal black in a 0,0,0 background event?
You could do things that way and that could work.
However, there are some subtleties to watch out for:
A- If you render to a new track, the levels on the render will be off. You have to remember to turn the video bus FX off and on.
B- Some formats (the ones that decode to computer RGB) like still images will require a different levels conversion.
Here's a quick check to see whether 16 and 235 luminance in the uploaded file end up mapped to 0 and 255 or not. Watch it at different resolutions, on different screens, and both embedded and on YouTube, as you may get different results. When I surveyed it a year ago, only about 2 out of 20 could see anything but white and black in this video. I doubt recent Flash Player updates have changed fundamentally in this behaviour, but it's possible.
Nick's test clip still looks as expected with the latest Flash version, meaning no differentiation.
And here's my free dual range 1080 gray scale for Vegas that can be used for checking workflow from timeline through delivery. Its use should be self-explanatory.
Just right-click and "Save link as."
Thanks MusicVid, I was looking for that pattern to test with - thought I'd saved it before.
I just tested using this pattern and the vimeo player is not expanding levels but the youtube player is. I'm trying to deliver a file that will play correctly on both but it's not to be.
Rocky,
Sorry, I'm not seeing that here. Jerry just uploaded a new copy of Nick's quick test to Vimeo, and the results are just the same as Youtube; iow, 235 is mapped to 255, 16 is mapped to 0.
One "can" see an imperceptibly slight shade in the 235 swatch, undoubtedly due to a 1-bit encoding shift.
Since Vimeo and Youtube use the same Flash Player, it makes sense that the results would be the same. Settings in some Nvidia drivers can negate this, we've found.
You need to understand this:
-Long story short... the scopes in Vegas are not reliable.
-The preview window in Vegas may be wrong. (!!!)
-In Vegas, you need to manually wrangle your levels.
-With other NLEs on the market, you do not need to bother with any of this stuff because they handle levels conversions for you.
It's not that Vimeo is "expanding" your levels. The preview window in Vegas expects computer RGB levels while many of Vegas' codecs expect studio RGB levels. This is why so many people on this forum are confused. What you see in Vegas is OFTEN NOT WHAT YOU GET.
Vegas is the problem... not Vimeo or Youtube. You need to manually intervene to fix things, and/or use something other than the Video Preview window for an accurate image.
Put Music Vid's chart on the time line and rendered Sony AVC as an MP4 at 16 Mbps and 1920x1080-30p.
Uploaded to vimeo and you tube.
Play it back (in chrome) from the website on my std monitor and look to see where the blacks or whites are crushed to being the same. With Vimeo 0-16 and 235-255 are all the same with no discernible line between them. On You tube all of the levels were visible.
Doesn't vimeo have a choice of flash or html5 player somewhere? That could be the difference.
Rocky,
Can you upload the original clip somewhere? Like dropbox or google drive?
Also links to your Vimeo and Youtube versions?
Are you applying the Studio RGB levels correction to both?
Results between the two are the same on this end.
Your browser/flash/html5 versions should make no difference.
I just uploaded a test clip (Mainconcept AVC) to both Youtube and Vimeo.
Both look the same to me. Anybody see a difference? https://vimeo.com/53947478
It's worth pointing out that you can only download a processed video from Youtube, while Vimeo allows one to download the original upload. Is that where you're seeing a difference, Rocky?
IE and Firefox give correct view on both vimeo and you tube.
Chrome on Youtube gives the correct view
CHROME/Vimeo gives a different view for the thumbnail vs. playing. That is, when you first open a video before clicking play the view of the thumbnail is correct. That is, blacks 0-15 and whites 236-255 are all the same for the original full range file. However, once I press play then those same ranges are now visible and step from value to value.