Rendering with vimeo settings loses contrast

Hi, all...Seems to me that when I render my video using the vimeo settings on the deliver page (except that I change Quality to Automatic, not limiting the bit rate), the video loses contrast. There's a big difference between what I see during color grading in Resolve and what I see in the rendered output.

Actually, now that I think of it, it's only marginally reduced in the output file, but noticeably reduced once I view it on Vimeo, especially on my TV with the Vimeo Apple TV app.

Can someone suggest what might be going on?

Thanks

QuickTime Gamma Shift:

https://jonnyelwyn.co.uk/film-and-video ... o-editors/

http://forums.creativecow.net/thread/8/1153069

http://www.videocopilot.net/blog/2008/0 ... mma-shift/

Hi, thanks for your reply, but I'm not able to get anything out of the links you provided. Or at least they have not given me a clear path. My video has a nice contrast when I play it in Resolve but is washed out when I export with the Vimeo 1080p settings and play it with QT Player. Also wash out when I upload to Vimeo and play it back. When I play it with VLC, the contrast is nice (unchanged), but it's a little oversaturated. I'm stumped!

Can someone please recommend a way to preserve the contrast when exporting from Resolve for publishing on Vimeo or other web sites?

History of the content: I ingested and edited the video in Premiere Pro and exported to Resolve in the Rec.709 color space. Finishing it in Resolve.

Thx

Steve

Can someone please recommend a way to preserve the contrast when exporting from Resolve for publishing on Vimeo or other web sites?

Important safety tip: Vimeo looks different from YouTube, and they both look different from other playback engines. Mac OS looks different than Windows. Chrome looks different from Safari, Safari looks different from Edge, Edge looks different from Firefox.

The number of possible combinations get fairly astronomical, fairly quickly. Me, I go through Frame.io and tell my Mac clients to watch on the current Safari, and it seems to work adequately under the current Mac OSX. Beyond that, it's "punt."

I was part of an effort about 13 years ago at Sony Pictures' ad/pub department where we spent 6 months testing various delivery methods in order to get consistent playback results on YouTube and Vimeo with different gamma and contrast settings, different color spaces, and other parameters. After 6 months and hundreds of tests... they gave up and just posted Rec709 2.2 gamma and let it go. Their reasoning was that there was such a lack of standardization on playback screens (laptops, tablets, phones, desktops, etc.) that that was a bigger problem than getting the playback engine correct per se. You can't control the world around you -- you can only put out the best picture you can under controlled conditions, then send it out and hope for the best.

Audio is even worse.

Hi, thanks for your reply, but I'm not able to get anything out of the links you provided. Or at least they have not given me a clear path. My video has a nice contrast when I play it in Resolve but is washed out when I export with the Vimeo 1080p settings and play it with QT Player. Also wash out when I upload to Vimeo and play it back. When I play it with VLC, the contrast is nice (unchanged), but it's a little oversaturated. I'm stumped!

It's not a Resolve problem.
Something in the video pipeline from video player, operating system to GPU is changing the levels or gamma.

Ack!

Since my delivery mechanism will be Vimeo, I'm tempted to dupe the timeline and exaggerate the contrast, almost crushing the blacks, so when it renders on Vimeo the contrast looks like it's supposed to. Nuts?

Or I might create a new project, import the .mov output file from my original project, adjust contrast as I mention above and output.

Thx

Steve

Ack!

Since my delivery mechanism will be Vimeo, I'm tempted to dupe the timeline and exaggerate the contrast, almost crushing the blacks, so when it renders on Vimeo the contrast looks like it's supposed to. Nuts?

If it looks good in Resolve then that is the best you can do.

What you encounter does not seem to be like relatively minor differences but levels and gamma shifts, this indicates that your viewing environment outside Resole is not setup correctly.

After thinking about this. I don't think it would be to difficult for BM to add a some built in correction in the 'Deliver' sequence so that Resolve can compensate for the colour encoding of these chosen outputs. So that when you chose an output for Quicktime, Resolve adds correction to your program to make it look as you intended.

Just my thoughts,

Steve

I assume you properly monitor Resolve preview over BM card, not in GUI preview on monitor?

Some people do "correction" for Vimeo/Youtube only viewings (even on 1mln$ projects), but this is technically not correct approach.
Whole thing is unrelated to famous term "QT gamma shift" as modern MOV files don't use gamma header.

Steve, it might take some experimentation but just using Rec.709 with gamma 2.4 in Resolve might make a quick improvement. That’s what I use and no one has complained viewing on their computers or television. My viewing screen on the iMac 27” is also defined that way.

But everyone is right here. WYSIWYG only applies on your system. Once you hand the file over to YouTube or Vimeo, they transform It to their preferences. And everyone viewing it on their equipment will see it differently anyway.

Sent from my iPhone using Tapatalk

Whilst I agree Rick, that it only applies on your own system. I do believe that these changes created by the different delivery options must be quantifiable. My feeling was that for those of us without a broadcast monitor a system like this could be quite useful!

I had already got Resolve set up as you suggested and have now set the MacBook Pro up with Rec.709 as well so I will have a quick play.

Cheers

Steve

This is just a reality of web platform delivery. I have been working on a short with very dark lighting, since it was all shot in the middle of the night, and though it looks good on my reference monitor, it's absolutely terrible on YouTube on a cheap TV.

Jack, you mean it looks like much of Game of Thrones? Shooting very dark scenes with almost no lighting seems to be very popular now. I’m assuming this stuff looks great on 108” OLED 4K or plasma screen but on my modest 32” LCD HDTV, I really have to imagine what’s there as I see almost nothing. Show me the Dragon Queen!


Sent from my iPhone using Tapatalk

Whilst it has always been true that you have no control over the equipment other people will watch you production on, I still think that I should expect some parity on the same computer! I do not recall this being a problem when using other software such as FCP, Premiere, Avid, and Lightworks. So why Resolve?

Steve

I do not recall this being a problem when using other software such as FCP, Premiere, Avid, and Lightworks. So why Resolve?

Why do you assume the problem is Resolve and not the viewing environment.

The test is simple, take the rendered result back into resolve, if it looks the same as what you graded then Resolve is not the problem.

Been there, done that Cary (see my other thread with examples), and there is a difference.

viewtopic.php?f=32&t=75337

There are other threads by users as well.

Steve

Been there, done that Cary (see my other thread with examples), and there is a difference.

viewtopic.php?f=32&t=75337

There are other threads by users as well.

Steve

In that thread, I understand you compare Resolve output with Quicktime.
That is not what I was saying, you got to bring the rendered footage in Resolve, compare it and if it is the same Resolve is not the problem.

It's like saying: "I went to the supermarket and got exactly one kilogram of sugar, but now I check it on my $5 home scale and it is a bit less". And then asking: "what is wrong with the scale at the supermarket?".

You cannot blame Resolve for what happens after the video is rendered.

I did take the Quicktime footage back into Resolve and compared the two using the scopes. There is a difference.

Steve

I did take the Quicktime footage back into Resolve and compared the two using the scopes. There is a difference.

That would be a Resolve bug, not a problem caused by the environment.

Could you make a little test: create SMTE chart in the timeline, encode the video and dropbox the resulting file together with a screenshot of the original scopes in Resolve? Use no LUTs and Resolve YRGB.

Hi Cary,
I hope this helps. The Scopes on the left is the straight SMPTE file, which was then exported via deliver to quicktime. The resulting file was then imported into the media page and then onto the timeline next to the original and its Scope reading is the right hand one.

Steve

SCOPES.jpg (271.91 KiB) Viewed 3602 times

Hi Cary,
I hope this helps. The Scopes on the left is the straight SMPTE file, which was then exported via deliver to quicktime. The resulting file was then imported into the media page and then onto the timeline next to the original and its Scope reading is the right hand one.

Steve

SCOPES.jpg

Looks fine to me, I don't see any gamma or levels shift.

Yes, I couldn't see any difference on the timeline either, so I will continue testing. What do the spikes represent in the right hand shot?

Steve

Yes, I couldn't see any difference on the timeline either, so I will continue testing. What do the spikes represent in the right hand shot?

I think it is chroma subsampling.

Yes, this is related to chroma subsampling (and if file is compressed then also to compression). Even intermediate codecs do overshoots on high contrast areas.