Fine Pixel adjustment question - AVS Forum
Forum Jump: 
 
Thread Tools
post #1 of 19 Old 09-21-2013, 08:06 AM - Thread Starter
AVS Special Member
 
curlyjive's Avatar
 
Join Date: May 2004
Location: Abington, PA
Posts: 1,336
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 44
I'm wondering if the fine pixel,adjustment feature many projectors has affects picture quality or 1:1 pixel mapping. I'm referring to the feature that allows for fine panel convergence. Since this must be done electronically, vs physically moving the panels, does it have any impact or degradation on picture quality? Do you if you shift blue up of example do you lose pixels on the bottom? If the increment is finer than 1 pixel how do you not lose 1:1 pixel mapping?
curlyjive is offline  
Sponsored Links
Advertisement
 
post #2 of 19 Old 09-21-2013, 09:24 AM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,210
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 604 Post(s)
Liked: 580
Try doing a little research on the forums. smile.gif Obviously if you shift the entire image up one line of pixels, you will lose one line of pixels. If you shift a partial pixel you will lose some resolution and sharpness. The prevailing wisdom by industry pixels are shifting a few tenths of a pixel is OK, shifting one half pixel is the worst case. Shifting more than one half is ridiculous, since you can shift by a full pixel and then back up by less than one half. Using 4K, I find the image is so dense and whatever you do has little adverse impact. Partial pixel shifts are made available by projector manufacturers to stop consumer poor panel alignment claims. Partial pixel fixes hide errors, they don't fix them.

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #3 of 19 Old 09-21-2013, 10:58 AM - Thread Starter
AVS Special Member
 
curlyjive's Avatar
 
Join Date: May 2004
Location: Abington, PA
Posts: 1,336
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 44
I did search but all I found was people's experience in using that feature, nothing about its impact on pq. So on my rs40 if I shift up I believe it is a full pixel. What your telling me is I lose a line of blue pixels at the bottom, which is no big deal since I have a little overshoot. But does it affect 1:1 pixel mapping or adversely affect the image since it is a full pixel? It does improve the convergence a bit overall, but I'd rather not use it if it is not advised to do so.
curlyjive is offline  
post #4 of 19 Old 09-21-2013, 11:05 AM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,210
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 604 Post(s)
Liked: 580
There really is no adverse impact and only positive impacts with full pixel shifts. You will never miss the line as you say and the convergence will be noticeably improved. regarding partial shifts, obviously you start using two pixels to do the job of one and you know incorporate the space between pixels. together this causes a loss of resolution and sharpness. A ad is considered acceptable and to overall improve the image. remember minor misconvergence can not be seen from normal viewing distances and even this is further ameliorated by going to 4K.

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #5 of 19 Old 09-21-2013, 11:56 AM - Thread Starter
AVS Special Member
 
curlyjive's Avatar
 
Join Date: May 2004
Location: Abington, PA
Posts: 1,336
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 44
And just to double check the rs40 uses 4 full pixel shifts correct?

I know they are ver visiable, the only time I can tell is if I look directly at the beginning of a black bar on a scope aspect film I can see a tiny bit of blue at the edge of the image. That is from sitting distance. I can live with that over impacting the pq, but if it doesn't hurt anything unlike pixel increments, then it does help. I uses the small crosshatch dot pattern on the avs calibration disc to verify the overall improvement.
curlyjive is offline  
post #6 of 19 Old 09-21-2013, 01:47 PM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,210
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 604 Post(s)
Liked: 580
Of all the many JVCs I have tested, I don't think I ever saw one that would be improved by a full pixel shift. I don't remember how many full are available but from the factory the shifts if I rember correctly are set near the middle of the range.

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #7 of 19 Old 09-21-2013, 01:52 PM
AVS Special Member
 
zombie10k's Avatar
 
Join Date: Dec 2007
Posts: 7,704
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 339 Post(s)
Liked: 392
Quote:
Originally Posted by curlyjive View Post

And just to double check the rs40 uses 4 full pixel shifts correct?

I know they are ver visiable, the only time I can tell is if I look directly at the beginning of a black bar on a scope aspect film I can see a tiny bit of blue at the edge of the image. That is from sitting distance. I can live with that over impacting the pq, but if it doesn't hurt anything unlike pixel increments, then it does help. I uses the small crosshatch dot pattern on the avs calibration disc to verify the overall improvement.

you could be seeing some CA from the lens, the blue would have to be off by a good amount to be visible from seating distance.

A good test is to put on familiar content and watch it from 1/2 distance to the screen. If you don't see any objectionable fringing on the edges of high contrast scenes, then leave it alone and stop staring at the convergence patterns.. smile.gif
zombie10k is online now  
post #8 of 19 Old 09-21-2013, 01:54 PM
AVS Addicted Member
 
darinp2's Avatar
 
Join Date: Sep 2003
Location: Seattle, WA
Posts: 21,220
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 70
Quote:
Originally Posted by mark haflich View Post

Of all the many JVCs I have tested, I don't think I ever saw one that would be improved by a full pixel shift. I don't remember how many full are available but from the factory the shifts if I rember correctly are set near the middle of the range.
This may depend on how the projector is setup. For instance, upside down on a ceiling mount versus right side up and overhead may have only one of those cases match the way it came from the factory. I do recall one where the best pixel adjustment was different between two different setups that used different amounts of lens shift, but don't recall the model.

--Darin

This is the AV Science Forum. Please don't be gullible and please do remember the saying, "Fool me once, shame on you. Fool me twice, shame on me."
darinp2 is offline  
post #9 of 19 Old 09-21-2013, 02:01 PM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,210
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 604 Post(s)
Liked: 580
All my tests were on table mounted JVCs which I was testing out for others and having calibrated for them.

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #10 of 19 Old 09-21-2013, 09:41 PM - Thread Starter
AVS Special Member
 
curlyjive's Avatar
 
Join Date: May 2004
Location: Abington, PA
Posts: 1,336
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 44
CA is probably it. When cold, having the blue shifted one looks completely off. So I'm just going to leave it.
curlyjive is offline  
post #11 of 19 Old 09-21-2013, 09:56 PM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,210
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 604 Post(s)
Liked: 580
Its not CA. CA is a function of how light bends when it exits the glass to the air in front of your lens. RGB have different frequencies and bend at different angles. This csan be ameliorated by the lens desin (front element) and the front element lens coatings. Thearea where it is minimized is in the center of the element, the flat sweat spot. As one uses more lens shift and goes towards close through, more of the image exits from ther curved areas of the lens. CA looks like misconvergence but isn't. CA is not a constant error like convergence is. As one goes more into the curvature it worsens on those areas of the screen where that light hits.

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #12 of 19 Old 09-22-2013, 02:23 AM - Thread Starter
AVS Special Member
 
curlyjive's Avatar
 
Join Date: May 2004
Location: Abington, PA
Posts: 1,336
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 111 Post(s)
Liked: 44
Quote:
Originally Posted by mark haflich View Post

Its not CA. CA is a function of how light bends when it exits the glass to the air in front of your lens. RGB have different frequencies and bend at different angles. This csan be ameliorated by the lens desin (front element) and the front element lens coatings. Thearea where it is minimized is in the center of the element, the flat sweat spot. As one uses more lens shift and goes towards close through, more of the image exits from ther curved areas of the lens. CA looks like misconvergence but isn't. CA is not a constant error like convergence is. As one goes more into the curvature it worsens on those areas of the screen where that light hits.

I understand the difference between the two and what causes them. Given that the error is most pronounced at the lower 1/4 of the screen where the lens curves more I think there is some CA going on due to lens shift. Probably a bit of both actually as there is no perfect convergence in a 3 panel system.

It's academic really as the errors are small for the most part and are rarely visible. On on an occasional sharp line that is out of focus from the source can I see actual fringing. And only if I'm looking for it. But it is helpful to better understand the issues with using pixel adjustments. That's really what I was after here.
curlyjive is offline  
post #13 of 19 Old 09-22-2013, 03:11 AM
AVS Special Member
 
JonStatt's Avatar
 
Join Date: Aug 2001
Location: London, UK
Posts: 2,260
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 45 Post(s)
Liked: 47
I have written a few posts detailing precisely the loss of resolution with fine convergence adjustments, particularly on JVCs. I did a search and it came up without too much difficulty.

I even explained how to see it yourself without any special tools on a JVC. All you have to do is set one of the colours to +-8 in either axis (try vertical). Then go to the built-in test screens and display the convergence test pattern which shows R, G and B lines all separated. You will see the colour you adjusted now uses 2 rows of pixels instead of 1. You have now halved the resolution in that colour!

JVC panels do shift as they warm up. You should set convergence for a warm projector, not a cold one. If it looks wrong when its cold, its only 10-15 minutes like that and then it should be sorted. Due to the way the panels are mounted, blue is the one that normally moves the most. Blue is also the least offensive if deviated from the rest. Red/Green convergence is the most importarnt.

CA is normally accompanied with loss of focus in that colour in part of the screen. It's easy to test for CA by using lens shift and seeing if the problem reduces/increases as you move back or forward into the direction concerned.
JonStatt is online now  
post #14 of 19 Old 09-22-2013, 06:47 AM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,210
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 604 Post(s)
Liked: 580
Jon. As you know, 8 is the absolute worst case one can use. One should never use higher, because by shifting one full pixel (harmless), one can go the other way and get under 8 instead of over 8. Basically, and its a judgmental balancing of eliminating the artifacts of panel misconvergence by giving up resolution and sharpness, probably up to about 4 will yield a net positive. I am working on a modification to my PQ meter to take the subjectivity out of this evaluation.

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #15 of 19 Old 09-22-2013, 11:18 AM
AVS Addicted Member
 
Charles R's Avatar
 
Join Date: Nov 2000
Location: Indianapolis
Posts: 10,207
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 218
Quote:
Originally Posted by mark haflich View Post

Of all the many JVCs I have tested, I don't think I ever saw one that would be improved by a full pixel shift.

 

This is my experience from using a half dozen or so JVC projectors. At least after warming up. It makes me wonder if they are set at the factory... of course you might improve one area of the image but you'll damage others as much or more. Now I have zero experience with sub-pixel adjustments.

Charles R is offline  
post #16 of 19 Old 09-22-2013, 11:25 AM
AVS Special Member
 
Ron Jones's Avatar
 
Join Date: Jul 2000
Location: Florida and West Virginia, USA
Posts: 5,956
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 199 Post(s)
Liked: 211
The JVC RS40 does not offer a really fine (less than one pixel) convergence adjustment. Rather you can shift the entire sub-image (e.g., red) one full pixel at a time and this has no negative impact on picture quality. If I recall correctly, the DILA chips actually have a few extra pixels both vertically and horizontally to accommodate the pixel shift feature. The fine convergence adjustments (with less than one pixel per step) offered on the some of the JVC, Sony, Epson, etc. projectors can impact image resolution if used excessively.

Ron Jones
Blog + Reviews + Articles: projectorreviews.com
Ron Jones is offline  
post #17 of 19 Old 09-22-2013, 12:30 PM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,210
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 604 Post(s)
Liked: 580
They impact it if used at all. Its a question of trade offs and being able to judge when the positive benefits are outweighed by the negative impacts. A subjective judgment call until I perfect my PQ meter.

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #18 of 19 Old 09-22-2013, 12:55 PM
AVS Addicted Member
 
darinp2's Avatar
 
Join Date: Sep 2003
Location: Seattle, WA
Posts: 21,220
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 70
It seems to me that the algorithms for JVC's E-Shift could allow for a 1/2 pixel shift (in 1080p pixel size) and not have that shift affect resolution with E-Shift enabled. I doubt they do though.

I've felt for a while that one advantage of 4K projectors would be the ability to do zone convergence basically within the 2K space.

--Darin

This is the AV Science Forum. Please don't be gullible and please do remember the saying, "Fool me once, shame on you. Fool me twice, shame on me."
darinp2 is offline  
post #19 of 19 Old 09-22-2013, 03:12 PM
AVS Special Member
 
JonStatt's Avatar
 
Join Date: Aug 2001
Location: London, UK
Posts: 2,260
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 45 Post(s)
Liked: 47
Quote:
Originally Posted by mark haflich View Post

Jon. As you know, 8 is the absolute worst case one can use. One should never use higher, because by shifting one full pixel (harmless), one can go the other way and get under 8 instead of over 8. Basically, and its a judgmental balancing of eliminating the artifacts of panel misconvergence by giving up resolution and sharpness, probably up to about 4 will yield a net positive. I am working on a modification to my PQ meter to take the subjectivity out of this evaluation.

You are absolutely right Mark. 8 is a worst case on JVCs. If you go to 4 then it is proportional where you would lose 25% resolution of that colour. An adjustment of 2 would be 12.5% and so on. If you only adjust the convergence in specific areas of the frame then its effect is of course smaller again.

To be clear though, I DO think there is value in fine convergence adjustment. It should be used towards the periphery of the image only to minimize or eliminate visible fringing from seated. In the central area, it should be whole pixel adjustments only.
JonStatt is online now  
Reply Digital Hi-End Projectors - $3,000+ USD MSRP

Tags
Jvc

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off