| Author |
Message |
KrisRoberts
Joined: 21 Jan 2007
Posts: 115
Location: San Diego
|
| Posted: Mon Oct 04, 2010 6:08 pm Post subject: Photoshop image analysis for uniformity? |
|
|
Long ago I recall someone posting about using something in photoshop to exaggerate non-uniform areas of their image as a tool for getting it more even.
The gist of it was they had defocused blue and got good greyscale tracking at the center of the screen, but were frustrated that the defocus was not consistent across the screen.
They would take a picture of an 80IRE blue screen, run it through their photoshop filter/process/whatever and it would show clearly which areas of the blue were darker/brighter than the center. Then adjust the mg focus for those areas to compensate.
After my last pass at setting up my projector and doing some more color calibration this weekend I can see that my uniformity is not great and did some searches to find the thread about this method but came up empty here and on AVS.
Does anyone recall this or have other suggestions for a process to get good results adjusting for better uniformity?
|
|
| Back to top |
|
 |
WanMan
Joined: 19 Mar 2006
Posts: 10270
|
| Posted: Mon Oct 04, 2010 6:26 pm Post subject: |
|
|
Sounds like one could simply adjust the gamma in the software to exaggerate the delineation.
_________________
Trust no one. Absolutely no one. Advice of the board.
|
|
| Back to top |
|
|
perisoft
Joined: 29 Aug 2007
Posts: 2920
Location: Ithaca, NY
|
| Posted: Mon Oct 04, 2010 6:30 pm Post subject: Re: Photoshop image analysis for uniformity? |
|
|
| KrisRoberts wrote: | Long ago I recall someone posting about using something in photoshop to exaggerate non-uniform areas of their image as a tool for getting it more even.
The gist of it was they had defocused blue and got good greyscale tracking at the center of the screen, but were frustrated that the defocus was not consistent across the screen.
They would take a picture of an 80IRE blue screen, run it through their photoshop filter/process/whatever and it would show clearly which areas of the blue were darker/brighter than the center. Then adjust the mg focus for those areas to compensate.
After my last pass at setting up my projector and doing some more color calibration this weekend I can see that my uniformity is not great and did some searches to find the thread about this method but came up empty here and on AVS.
Does anyone recall this or have other suggestions for a process to get good results adjusting for better uniformity? |
You'd have to compensate for hotspotting, otherwise all you'll see is a big hotspot.
First step, get photos (identical) of each channel, load 'em in, convert to grayscale.
You could do that by using the green / red (which don't vary luminance with focus) to generate a baseline luminance map, and then invert it and add it to the blue. After that, any luminance variation in the blue ought to be due to focus.
Caveats: If you have a non-1.0 gain screen, there might be color shift in the R/G, meaning the hotspots would be in different locations (from the camera's point of view). You could shift the hotspot maps to match, but obviously that would create an area off to one side where there isn't any data; you could compensate for that by mirroring L/R (which are presumably identical). So, find the center of the hotspot, mirror the larger side across, do the same for R (or G), composite them to get a bit of an average, find the center of the blue hotspot, and put the inverted composite on.
After that, copy merged layers so you get one bitmap with the whole deal, and auto-levels that to pull out the variance.
Easy-peasy!
_________________
|
|
| Back to top |
|
|
virusc
Joined: 11 Apr 2007
Posts: 358
Location: Massachusetts
|
| Posted: Mon Oct 04, 2010 7:17 pm Post subject: |
|
|
what projector do you have?
there is uniformity control in some projectors for uniformity specific issues.
If you mean evening out wear (16:9 or un centered raster but not text burn) I do not think it is worth the trouble. I have tried it and only got marginal results.
also screen, lenses, and physical setup can effect uniformity if not correct or optimal. Average light falloff at far edge of phosphor and lenses is measured at about 50% drop off average in most cases(widescreen review study some years ago).
|
|
| Back to top |
|
|
perisoft
Joined: 29 Aug 2007
Posts: 2920
Location: Ithaca, NY
|
| Posted: Mon Oct 04, 2010 7:46 pm Post subject: |
|
|
| virusc wrote: | what projector do you have?
there is uniformity control in some projectors for uniformity specific issues.
If you mean evening out wear (16:9 or un centered raster but not text burn) I do not think it is worth the trouble. I have tried it and only got marginal results.
also screen, lenses, and physical setup can effect uniformity if not correct or optimal. Average light falloff at far edge of phosphor and lenses is measured at about 50% drop off average in most cases(widescreen review study some years ago). |
He's talking about neither of those things. He's saying that his uniformity is off because some parts of the blue are focused and some not; the not parts are brighter. What he wants to do is figure out more specifically which parts are defocused by doing image processing on a photo of the screen.
_________________
|
|
| Back to top |
|
|
dvh99
Joined: 25 Dec 2009
Posts: 2158
Location: nederland
|
| Posted: Mon Oct 04, 2010 7:56 pm Post subject: |
|
|
easiest thing to do is perhaps measure the luminance from photos taken close to the screen taken at the parts that allow for focussing of course.
then when you have all this data take the average and try to get the average luminance value everywhere.
a webcam or videocamera might be easier for this.
_________________
1 answer always poses multiple questions.
marquee 9500ultra HD10L moome hdmi1.3 v3+ some mods.
|
|
| Back to top |
|
|
dvh99
Joined: 25 Dec 2009
Posts: 2158
Location: nederland
|
| Posted: Mon Oct 04, 2010 7:58 pm Post subject: |
|
|
hfq 900 lenses are the best they give the absolute best available for uniformity.
_________________
1 answer always poses multiple questions.
marquee 9500ultra HD10L moome hdmi1.3 v3+ some mods.
|
|
| Back to top |
|
|
|
|
