Originally Posted by Gotchaa
I've done Stage with Superwhite, Standard Color Profile, with Standard HDMI levels, and No Color Profile as well, 3 reports. I need to compress them before posting the PDFs, PM me your email and I will send it over.
Just to update those interested in this, Gotcha has kindly emailed me his results (pdfs too big to be attached), and I can confirm that stage doesn't track better regarding saturations than the profiles tested by Jonstatt. So it looks like the saturations bug is still there on all the profiles when CMS corrections are applied (standard/THX track almost perfectly when uncorrected).
Gotchaa hasn't had the time yet to try a calibration from profiles off, and might be able to do so over the week-end, but as profiles=off doesn't seem to be tracking properly uncorrected (undersaturated at 75% sat and below), it's very unlikely this will not get worse when a substantial amount of CMS correction is applied. It also makes it a not so great candidate as the starting point for a calibration with an external VP, which is bad news as it's narrower than in last year's models (which is good news). Also, as before, all the advanced controls (sharpness, CMD, color temp) are greyed off when profiles are switched off, which makes it very inconvenient.
For those who don't calibrate themselves, there is no need to panic, a good calibrator should find an acceptable compromise (they just won't be able to show you a nice looking report with all the targets spot on and a minimal dEs, unless they cheat!).
It doesn't mean that the picture will not be close to the standard in reality, even if I know that many might struggle with this concept. What it means is that it will not be as close to the standard as a calibration which is spot on at 100% sat and tracks perfectly at lower levels (an ideal rarely reached, even with excellent external VPs such as the Radiance), but it will be closer visually than a calibration that looks spot on at 100% sat (which is where the rec-709 calibrations are usually made) and is strongly undersaturated at 75% and below, which is unfortunately the case with the internal CMS since the rs50/rs60 when aiming for 100% of rec-709. The former gen (rs20-35) was not perfect, but the undersaturation wasn't as pronounced as it is on the new models (rs50-55). As far as I can see from these measurements, and both Gotchaa and Jonstatt seem to agree, it is about the same on the rs55 as it is on the rs50.
For those who calibrate, like last year, it means that you are unlikely to get a good calibration from any profile if you aim for rec-709 (so autocal with a non modifiable target gamut is unlikely to work well. It will show you perfect graphs and charts, but the picture itself, if you compare it with THX/standard, is likely to be quite undersaturated). The best compromise is to select a wide gamut (like stage or anime1), and aim for oversaturated targets at 100% sat (it doesn't matter if you are calibrating at 75% stim or 100% stim) in order to compensate for the undersaturation at 75% and below. Calman allows you to define a custom target gamut, so you can save the right targets once you have found them. I don't think Chromapure allows this (saving a target gamut) at this stage.
If you are not experienced enough to do that, you might able to get better results (by this I mean visually) when using a target gamut at 75% sat instead of 100% sat (the default). This is what Chromapure allows you to do when you select the 75% of rec-709 option as the target gamut. In Calman, you would have to define a custom gamut at 75% sat as I don't think the 75% sat option is there. The result is not as good as when you find the best compromise manually, but in my experience (on the rs50) it is better than when you calibrate at 100% sat, which is the default (whether you use the 75% stim or 100% stim patterns).
I have asked Gotchaa to confirm that the picture is visually undersaturated on his stage calibration when he targets rec-709. It should be as the undersaturation at 75% sat and below is as pronounced as it is on the X7 when doing the same.
Hopefully Gotchaa and others can confirm this (and confirm that the profiles=off doesn't track properly either), for the CMS saturation bug to be (un)officially confirmed as not fixed.
I attach three screenshots from Gotchaa's reports, one showing the way saturations track when stage is corrected at 100% sat with the internal CMS, and the other showing the way saturations track whith profiles = off (uncorrected), so native gamut. The last one shows saturations for standard/THX as a reference, to show that uncorrected profiles track almost perfectly (unfortunately, it's slightly undersaturated so uncorrectable). Of course there might be some unit to unit variations, so hopefully more users will confirm whether it is the same on their unit or not.
Thanks all for all the good investigation work!