FAQ: Viewing Gamma Choices Explained

[R Neil Haugen]

Viewer Gamma Choices Explained for SDR Color Work

I don't know if you're on a Mac without reference modes, a Mac with Reference modes, or everything else. This will cover everyone basically. For reference, I work for/with teach pro colorists and have worked in both Premiere and Resolve for over a decade.

Rec.709 standards require the screen/monitor to apply a display transform to the video data of essentially gamma 2.4. All broadcast specced systems, PCs, TVs, streaming, Android, and Macs with Reference modes set to HDTV  ... follow the proper standards.

ONLY Macs without Reference modes use a different, and very much lighter, display transform!

For some unfortunate reason, Apple uses essentially gamma 1.96 on all Retina Macs without Reference modes. Which presents a light image in the shadows of the same file data. They may also have a separate issue in mapping Rec.709's sRGB data to their P3 color space screen. As the image is also low in saturation.

This happens with any app that allows the Mac OS to manage color. Such as QuickTime player, and Chrome and Safari browsers.

VLC player, Potplayer, and Firefox do not allow the Mac OS to do color management to Rec.709 media, so they show a more correct image than QuickTime player on Macs without Reference modes set to HDTV.

The image therefore looks more like on all other systems outside the "Macs without Refererence modes" ecosphere. It's a good way to check the results using the two very different transforms.

Premiere 25.x has Viewer Gamma Choices!

Viewer gamma 1.96/Quicktime will show an image within Premiere's Program monitor, that is very similar to QuickTime player, Chrome and safari outside of Premiere ... but only on Macs without Reference modes. It's not useful for anything else.

And note, when you export the file to play it back ... the image on all other systems ... the vast majority of screens out there ... will be the darker image, than the light image on Macs without Reference modes.

Viewer gamma 2.2/web is what most users should be using, not because "web", but because if you are grading in a normally 'bright' room the standards require grading with a display gamma of 2.2. And most editors don't work in the really dark "proper" colorist's room lighting.

Viewer gamma 2.4/broadcast should be used only if you are in a very dark, near to but not quite, blackened room. That is the required setup for broadcast quality grading with Reference monitors and a 'bias' light on the wall behind the Reference monitor.

[Kevin-Monahan · Accepted solution]

Thanks for the FAQ, Neil!

Kevin

[R Neil Haugen]

Most welcome sir. The rabbit hole on all this can be incredibly deep, yet, for most users, the practical steps are pretty simple.

[Shebbe]

Hey Neil,

I appreciate your efforts in explaining, but I think you're not quite accurate here.

"Rec.709 standards require the screen/monitor to apply a display transform..."

"ONLY Macs without Reference modes use a different, and very much lighter, display transform!"

The term display transform means something completely different from what you're referring to, display characteristics and/or calibration. The standard also doesn't require anything to be applied by the dispay. It's a bit weird to explain it that way.

"For some unfortunate reason, Apple uses essentially gamma 1.96 on all Retina Macs without Reference modes."

Source? The non XDR retina displays have 2.2 gamma last time I checked.

"Viewer gamma 2.4/broadcast should be used only if you are in a very dark, near to but not quite, blackened room. That is the required setup for broadcast quality grading with Reference monitors and a 'bias' light on the wall behind the Reference monitor."

The specification requires a dim surround not dark. Dark surround is for cinema.

[R Neil Haugen]

I've been through hours of discussions, presentation, read several deep texts and papers. Which all list the specs as including a display transform in order to keep the legacy view of the time we used CRT monitors which naturally had a "shape" to their response similar to the gamma 2.4 function that was adopted to match. Ergo Bt.1886.

You're not suggesting that Bt.1886 doesn't exist? 

You can go into the rabbit hole here quickly, and get incredibly technical and parse terms forever. I look for the practical effect ... and everything I've ever seen always lists that display transform for Rec.709 media.

In fact, some heavy duty color calibration and management types have papers out positing that since we haven't had CRT's in a long while, the Bt.1886 transform is something that should get dropped. However, they don't have a way to do so without shall we say, a fair amount of issues viewing media "built" for the Bt.1886 transform and media not built for that transform on the same device.

As to the Mac gamma issues. I've seen tons of noted experts from Steve Shaw through major colorists and color scientists ... all stating that the Rec.709 process in Retina's without Reference modes uses "essentially" a gamma 1.96 view. I've never, ever seen that disputed. Anywhere.

And of course, if those Macs actually used the same display transform as the specs, there wouldn't be the entire Mac "gamma display" issue, would there? The difference between a display gamma of 2.2 and 2.4 isn't enough that most people would even notice.

As to your final point, I am puzzled. Because I clearly expressed a difference between a totally dark room and one specced for Rec.708's "semi-darkened" requirements. Looking back at it, I cannot see where you missed that.

I am of course totally aware of the difference between a black room, for cinema projection grading only, and that of the "semi-darkened" room specified, and the data for the measure light levels in that room.  

As someone who has setup the room to meet the standards, including all light metering, it's pretty darn dark when you have total control of light sources and all. Which is why the use of small localized light sources/lamps is so needed by colorists for their workspace so they can see 

[Guillaume30883528kcuh]

Hello,

I'm french so excuse my english please.

There's something that I really don't understand with viewer gamma option. See images below, they show my premiere settings, also the program monitor (upper image) and the quicktime look after export (lower image).

I've made 4 tests with the same shot, changing the viewer gamma and also the media color space setting.

As you can see, when the "source clip" section is set to "use media color space", then "viewer gamma" 2.4 or 1.96 change the look of images inside premiere's display but does not affect the exported file. BUT, if for some reason we want to override media color space of the clip (here set to sRGB instead of rec709) then we get in the reverse situation : changing "viewer gamma" from 2.4 to 1.96 won't do anything in premiere display, but the exported files will be different.

It seems like overriding color space of the media is applied after the gamma viewer correction, wich is very weird logic to me.

Am I missing something ?

Thanks