Large adjustments, should I be worried? Epson 8350

Just did the first run of calibration of Epson TW-3200 (aka 8350 in the US) using latest AVSHD, PS3, SpyderII and ColorHCFR. After reading several reviews and people saying how well the projector is calibrated out of the box, I was surprised to end up with quite a few settings to get things right.

Here my settings - is there anything that jumps to you as odd?
Basic settings: Cinema, Iris hispeed, eco mode, skin tone 3 (default), NR off, super white off, HDMI range Expanded. I got BTB and WTW to show up, obliterated everything outside 16-235 with Brightness and Contrast.
Brightness -11
Contrast 11
Color Saturation -11
Tint -7
Color temp 7500k
Offset R -3
Offset B 1
Gain R 17
Gain B 3
Green offset/gain 0
Gamma 2.2 (default)

With this I got very good tracking of gray levels: deltaEs well below 4 from 30-100% greys, 20% was around 6. Grey was very close to 6500k, Cyan was quite close to reference, and red Y was ~21% of white Y. There are no clipped colors, and white and black points are good.

So although I had to make quite a few adjustments, they ended up really nicely. The main thing I'm concerned is whether there's something I should be concerned about with the large tweaks, or am I fretting about nothing?

I started with the color temp, perhaps I could get lower adjustments with calibrating with default 6500k - I might try that over the weekend.

edit: forgot advanced settings

About that... My screen is actually a custom-made artists's canvas, so it was never even meant to meet any standard. It cost 105 EUR four years ago, a fraction of the cost of a proper screen, and works beautifully - but clearly is not optimal for accurate color reproduction.

For my second run, what's a good point to start? Should I reset color temp to default 6500k before setting Brightness/Contrast or grey levels. In other words, which of the basic settings is most likely to result in unwelcome color shifts, banding or clipping?

Excellent tips on skin tone (wasn't sure what it does) and on gamma, thank you! I had an inkling there was something going on when I had to set temp to 7500k.

I'm not sure I agree on the 255. From what I've read there are practically no BDs which have any usable data above 235: BD spec is 16-235 after all. If I calibrate pure white to 255, actual whites from BD sources would be light gray, right?

I'll run a second calibration today or during the weekend.

I've read numerous threads, and it boils down to BD content being encoded with only 16-235 of useful information, everything else is either noise or a very rare exception. Therefore I don't see what would be the point of calibrating a PJ to show anything above 235 for BDs.

There might be a point if one uses a PC as main source, and the entire I/O chain supported 0-255.

Echoing alluringreality, I'm not talking about computer playback, but BD from a PS3. I merely mentioned PCs as a potential reason to calibrate for 0-255 for those who use it as their main source and whose entire I/O chain supports it.

Since anything outside of 16-235 is out-of-spec for BDs, I still don't see the point of calibrating to 255, as that would leave "BD white" ever so slightly grey. This is assuming something in the chain doesn't expand it to 0-255, which I assume some GPUs would do - in this case whites and blacks could be clipped if the PJ expects a 16-235 signal.

While my current calibration (done to 235) looks good, I haven't scrutinized the whites thoroughly. I'll try to find some movie scenes from my BDs with plenty of whites during the weekend to see if there is any benefit from calibrating to 255.

My understanding is that PS3 outputs all BD/DVD content with YCbCr, and that Limited/Full setting doesn't impact BD/DVD output. I just did a quick test with the AVSHD calibration disk, and didn't see any difference on black or light levels between Limited and Full.

I just tested white levels with BBC's Planet Earth - Ice Worlds BD, which has plenty of whites all the way to specular. The 235 calibration looked good, but the one I calibrated to 252 had slightly grey whites with no additional detail. This in my view confirms that 235 is the correct calibration - or is there another explanation?


That's my understanding as well, and that's what I've had it set up for years.


How can you force games to YCbCr? Is that Deep Color output off in Display Settings?

Ok, that's starting to make sense I'm more familiar with (still) photographic color spaces, but I don't know YCbCr.

When I calibrate to 235 I don't get any clipping according to the color bars showing clipping on the AVS HD disc (third pattern on Misc. Patterns IIRC). This would imply they don't "have excursions" beyond 235, and as such are of limited value.


My bad, I used inaccurate language. What I meant by "grey" is that the whites have less luminosity when calibrated to 252, than the same setting calibrated to 235. The only difference between the two is contrast setting. The 252 is based on my 235 calibration; I didn't bother to recalibrate the 252 from the ground up. But based on what you said above, I will do that tomorrow - I might be able to get higher than 252.

Thank you, excellent post I haven't seen before. Pretty surprising results shown there. How accurately were you able to get the primaries and secondaries to track with the CIE targets with your 8350? Did you use the guide you linked to to get there?

I haven't worked much on getting primaries dead on, yet. I did check them early on, and all of them were oversaturated at 100 IRE (or was it 75) - haven't checked calibrated results.

I used the method described by Curt Palme (and others), where you set Red channel Y to 21% of pure white Y at 75 IRE (had to drop saturation quite a bit), and Cyan to REC 709 reference values of x=0.225 / y=0.329 (I only minimized the sum of deltas as I wasn't able to get to reference with Tint control alone). I did some quick checking with The Tourist and it looked good - but it could be just because the movie is useless as a reference disk since all you do is stare at Angelina Jolie :P

BTW, I re-calibrated with 6500k, skin tone 0 and gamma set to 2.4 as you suggested earlier - you were correct, it resulted in gamma of 2.2 and good calibration. So thank you! Next steps are to re-calibrate with 16-25x (as high as it goes), and go deeper into the CMS to see if I can get the primaries to line up along the CIE chart as per your link.

I just finished the first full calibration which took around 2 hours, and am very happy with the results DeltaEs around 1 on grayscale, tracking very nicely with 2.2 gamma, 6500k temp, colors and even saturation. I was surprised how well-behaved the CMS was: I set all the saturations at 75% saturation, went back to grayscale and gamma to tweak it back to reference. After returning back to 75% saturation targets I had to make only very little adjustments.

The only issue I have is with green saturation (post here), but I think I can live with that.

Thanks again for all the help, especially Droid6!

Now it's Blade Runner time

edit: BR looked gorgeous, deep blacks, faithful skin tones, and bright whites. That movie never gets old.

edit2: I ended up with going back to 16-235 despite recommendations above. I re-calibrated with 16-235, and as expected whites are now brighter than 16-252, there's no clipping of colors according to the color gradients, and the calibration required less tweaking. Calibrating to 16-235 instead of 16-25x should also improve contrast.