AVS Forum banner
Status
Not open for further replies.
1 - 16 of 16 Posts

·
Registered
Joined
·
373 Posts
Discussion Starter · #1 ·
I've been noticing an image problem with my MT8 that I dont remember seeing on my 12S1, though I didnt get to spend much time with that one.


Problem is most noticeable on peoples faces if the face is moving across the screen. The skin tones take on a banding effect as the face shifts left or right. Its as if DLP can not keep up, because pausing the image returns the face to normal. This happens in HD and DVD sources. The face goes from being made up of say, 256 colors, to 16 and the difference in light level on the face goes from smooth gradations, to a dramatic difference in colors like individual bands. I have Firmare 2.9 BTW. Do all DLPs do this?


Mark
 

·
Registered
Joined
·
11,213 Posts
Get thee over to the special guests forum. Pete Broas from TI just owned up that virtually all single chip DLPs are 8b. Leaves little room for error for A/D conversion/calibration of your video. Even though they are higher bit gamma tables - this is used to spatially/temporally dither the signals to try to effectively increase the bit depth by fooling the eye. To save bandwidth motion scenes drop the dithering - loosing the effective bit depth. Or even worse the scene motion interferes with the temporal dither and you get a flat fuzzy mess rather than the detailed deep picture.


And faces are a real problem - they are broad areas of minute detail that the brain sees but that is not digitized well. Is that cheek moving or sitting still, is that a freckle or a dither!


So it is TI's fault - not your DLP's fault. The best you can do is calibrate still grey/colour ramps (sub brightness/contrast RGB) to eliminate any banding caused by poor A/D conversion to avoid reducing the dynamics at the source. But then you may have a flourescent rather than daylight coloring - so it is a tradeoff between dynamics and true color.


Cinema 3chip DLP has the bandwidth to add more real bit depth to the display drivers - so dithering, banding and motion should not be a problem.
 

·
Registered
Joined
·
373 Posts
Discussion Starter · #3 ·
Thanks Krag, If A/D has some effect on this, Ill hav e to test my new DVI DVD player and see how much improvement it makes. Thanks again.


Mark
 

·
Registered
Joined
·
6,436 Posts
I noticed that on my HT1000, but only certain DVD players did it, and it was a pretty rare occurance at that.


In one case, I ran one particular scene over and over ten times with the same player, and it only did it twice. Another player did it every time...


I haven't had a chance to try my HTPC with the disks that I noticed it on, but I'll have to try that out of curiosity...


Gary.
 

·
Registered
Joined
·
1,588 Posts
I have not noticed this effect using my Panny XP30 DVD player in either Interlaced or Progressive mode on my SP7200.


It is possible that a combination of precise calibration and the particular DVD player used have an effect on this situation, or else I just don't have as keen an eye as you.


I'll keep a look out for this "artifact" and see if I can spot it when looking for it specifically.
 

·
Registered
Joined
·
6,436 Posts
It's almost like contour lines you see on a map, and most noticeable on facial skin tones.


Not sure if it's related to how well the player/display can cope with colour posterisation, but a Philips 963 was the worst offender for this in both the banding/contour lines that Mark mentioned, and colour posterisation in general.


A good test of your player/display is the earle scene in Titanic where the sub sinks into the depths to photo the wreck - as the sub descends, the lights of the sub form a circle of light which is bright blue on the inside, and should gradualy fade to darker blue on the outside.


A bad player will show distinct bands of colour, while a good one will show a gradual fading from light blue to dark. The Philips had very distinct bands on both this test, and on the faces like Mark described.


All other players had minimal banding, or none at all.


HTH


Gary
 

·
Registered
Joined
·
373 Posts
Discussion Starter · #7 ·
I think I first noticed this when I was watching the Oscars in HD. They showed all the old past winners, faces were panning right to left on the screen. The colors looked fine when the camera stopped panning, but as soon as the camera panned right, that persons face would turns into bands, every face did it, and this was 1080i HD into the projector. It was not nice at all, and VERY easy to spot.


Mark
 

·
Registered
Joined
·
6,436 Posts
I wonder if calibration can make a difference?


I had problems calibrating the Philips DVD player, so maybe that was why it was so bad with this.


Maybe calibrating for the source of HD would help? Are there any test patterns broadcasted in HD?


Gary.
 

·
Registered
Joined
·
11,213 Posts
Calibration makes a big difference to minimize the effect. But you must do it for each analog source because the voltages may all be different! If you have a 8b DAC (DVD player) into 8b ADC (projector) it can be tough to get perfect dynamic range - especially if each is adjustable.
 

·
Registered
Joined
·
6,525 Posts
I noticed this on many HD1 projectors but haven't seen it nearly to the same degree on the HD2 units. The sharp 10000 had *no* evidence of color-banding that I could discern like I've seen on the older 9000.


-dave
 

·
Registered
Joined
·
339 Posts
Could there be any relationship between this and the noise reduction feature? The only reason I wonder is that you can choose to not apply noise reduction to skin tones. There is a brief mention in the manual I think about keeping the look more natural. I know next to nothing about how the noise reduction works, but thought it curious that InFocus singles out flesh tones for special processing.


I have not seen this on my new MT8, but I've only had it for a few days.
 

·
Registered
Joined
·
4,040 Posts
Hi,


This is classic. Keep blaming the source or whatever.


Bottom line: avoid DLP at ALL costs. Some say even 3-chipers has the same issue.


regards,


Li On
 

·
Registered
Joined
·
6,436 Posts
Li On,


The effect varies from source to source - possibly something to do with matching them perhaps, but one player is very pronounced in th effect, one minimal, and two didn't have a trace of it. So the source is in fact a factor.


I prefer DLP myself, and although it has it's pros and cons like all display devices, the cons of DLP are less of a concern to me than the cons of LCD (and CRT for that matter), which presently I wouldn't really consider. YMMV of course.


The industry seem to have a similar line of thought perhaps, considering that we are now seeing DLP in commercial theaters, and not LCD.


Gary.
 

·
Registered
Joined
·
4,040 Posts
Hi,


Just checked the TI thread in the AVS Special Guest forum.


TI official comment in respond to a KBK question is:

As for bypassing the dithering, No. The dither algorithm helps gives the necessary bit depth required for high-quality video images.


The bad news is the motion dithering is still a feature on DLP.


The good news is I don't need to bother with DLP in the near future.


In the meanwhile you guys can keep the classic going by blaming the source as if TI haven't told you countless times about this feature is in the heart of every DLP engine...


regards,


Li On
 

·
Registered
Joined
·
6,436 Posts
I guess LCD doesn't have any features of it's own then? ;)


Gary.
 

·
Registered
Joined
·
4,040 Posts
Hi Gary,


LCD has many picture artifacts such as screendoor, vertical banding, color bais, motion scanline, weak contrast, poor black level. But I haven't heard anyone says these "features" actually help gives a wonderful picture. Let alone from a official comment.


What really scare me is what if the TI rep is actually right? What if those DLP buyers actually enjoying these dithering artifacts? And can't live without it?


regards,


Li On
 
1 - 16 of 16 Posts
Status
Not open for further replies.
Top