Mr D. Gamma of 1.5

Discussion in 'Projectors, Screens & Video Processors' started by ROne, Aug 16, 2002.

  1. ROne

    ROne
    Active Member

    Joined:
    Dec 21, 2000
    Messages:
    781
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    18
    Ratings:
    +8
    Mr D, I have tested my LCD projector to have a gamma of 1.5, I know you say it ought to be 1.0 but as I have suspected it ain't linear.

    My question is what maths or system do I use to compensate to 2.2? At the moment i have access to the overlay and am using 1.1 gamma setting and this seems reasonable.

    Any pointers - thanks.
     
  2. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,190
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,229
    OK first things first. this is where it gets tricky.

    That "1.5" is a combination of the display gamma and the system gamma( including material) and : assuming you have the black point and white point correctly set. ( lets assume it actually is an end gamma of 1.5 it might well not be indicative of that but we'll get there later)

    We need to figure out a couple of things. However we know that for video material an end gamma value of about 1.2 is desirable : 1.0 tends to look a little bit flat for most people( so 1.5 is not bad if its a real value ).

    Here's the theory and sorry but its based on CRT methodology which is intrinsically linked to video. Most CRTs have a gamma of around 2.5 (2.2-2.8 in the real world) . Video has the inverse gamma of this : around 0.45 giving an end to end gamma on display of 1.1 -1.2 or there abouts ( 1.0 is found to be a bit flat looking for most people even though its technically correct).

    So all your 1.5 is telling us is that your projector when fed from your source gives an end to end gamma of 1.5. If your 1.1 gamma setting is accurate on your overlay ( which is doubtful unless the whole system has been calibrated right through the hardware) then we infer that the display gamma of the LCD is about 1.363... so to get it nearer to 1.2 you'd need to drop your system gamma to about 0.9. If its accurate. What you should do is tweak the gamma on your overlay until you get the 1.2 patch to match. Assuming you don't have a gamma option on the projector ( which will complicate this cause then you are playing with two unknowns .. but we sort of are anyway)

    Another very very important issue is what the gamma chart you use is actually saying : for example a patch labelled 2.2 : is this the actual end result value in which case if you were to look at a display with a gamma of 2.2 and feed it a system gamma of 1 you'd see a match at this patch .
    Or is the patch value supposed to relate to the end to end gamma of display and source combined which ideally you want to be 1.2 or there abouts. I'm pretty sure the Avia gamma pattern is supposed to be correct at 2.2 (although its actually representative of an end gamma of 1.1 -1.2 you see confusing isn't it because it assumes your display will be a gamma of 2.2) Still with me?

    Additionally your LCD may well have gamma correction applied to the incoming signal to correct for the fact that the display gamma of the panel is linear ie 1 or thereabouts whilst your source is designed to display on a gamma of 2.2.

    How would I do it?

    Firstly I'd start with the scoped settings for my hardware off the theatertek site and verify by using my CRT monitor that these values are correct for an analogue display with a gamma of 2.2.
    This way if I hook my HTPC up to a digital display and it has gamma correction on it to effectively take it 2.2 then the image should look correct: verify with the Avia gamma pattern which is correct when 2.2 matches ( or thereabouts). If I hook it up to a display that is other than 2.2 then I'd firstly ensure the black and whitepoints on the projector were correct without touching the overlay settings on the PC. Then I'd check the gamma with the avia pattern and either adjust gamma on the projector ideally if available or my overlay if thats the only option. Then I'd do colour.

    If your gamma pattern is the Avia one and it matches at 1.5 then your system is wonky and you need to recalibrate.

    If its any other pattern I haven't got a clue what its telling you.
     
  3. ROne

    ROne
    Active Member

    Joined:
    Dec 21, 2000
    Messages:
    781
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    18
    Ratings:
    +8
    I'm going to absorb that - but my gamma reading is based on the AVIA patch for the overlay and a desktop pattern which both indicated 1.5.

    Just going over the details.
     
  4. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,190
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,229
    You need to recalibrate then. Just verified that 2.2 is definetly the aim on the avia pattern.Do black and whitepoint first then gamma.
     
  5. jrwood

    jrwood
    Guest

    Products Owned:
    0
    Products Wanted:
    0
    Ratings:
    +0
    Mr D. interesting stuff indeed :)

    You say you should aim for 2.2 does this apply to any projector ?. I had a read of http://www.photoscientia.co.uk/Gamma.htm which was interesting, they also have a gamma 2.2 pattern to test your system on. Is this a good pattern to test a system on or is it wrong like most other so called calibration patterns that the site owner so eloquently talks about!.
     
  6. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,190
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,229
    Displaying an image file of a pattern is not the same as displaying the gamma pattern off Avia ( to be honest its the only reason I have Avia as I prefer Video Essentials). You need to know what your system is going to do to material sourced off dvd. So I still recommend using the Avia pattern.

    2.2 should be the aim from the gamma chart on Avia : any other pattern I can't comment on.
     
  7. ROne

    ROne
    Active Member

    Joined:
    Dec 21, 2000
    Messages:
    781
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    18
    Ratings:
    +8
    Just a couple of things to tidy up:

    When you refer to be having an aim of end to end gamma of 1.2 for video how does this relate to having a match in AVIA of 2.2.

    My AVIA chart will only show 1.5 for me. There is no way with my digital panel/system I can get near anything like 2.2 with any combination of gamma/brightness/contrast.

    the new drivers for the radeon 7200 now support pre-determined gamma curves rather than values. They are 0.85, 1.0,1.1, 1.2. 1.45, 2.0.

    I find 1.1 lifts the shadows out a bit more than 1.0. whereas with the 1.0 setting my black bars are clipped in AVIA.


    Thanks.
     
  8. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,190
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,229
    The information relating the 2.2 gamma match on the Avia pattern was from Guy Kuo:


    . Gamma Check

    Display AVIA's Gamma pattern and find the gray patch which best matches the background in brightness. An NTSC display should be between 2.2 to 2.5. Otherwise images will appear too light or too dark even if black point and white point are correctly set. If a display has gamma problems, I avoid it because you usually cannot correct the problem.

    This indicates to me that the gamma patch named 2.2 is actually relating to an overall end to end gamma correct for video material displayed correctly on a video display device :ie the end gamma is 1.1-1.2 and the patch correlates for this on the 2.2 patch.

    I'll have a play around with my system tonight and let you know what I find: I've not run the Avia patch through it yet (although I did with the HTPC hooked up to a CRT monitor). Mainly because I find Avia really painful.

    It is quite a tricky pattern to evaluate. Try squinting at it or defocus it to better see what patch relates to the non-dithered brightness level.

    I actually played around with some gamma pattern images off the net at work yesterday ( where I have a very accurate monitor and very accurate gamma control of the incoming signal).

    Selecting a system gamma of 1 gave me a match in the area between 2.2 and 2.5 which is bang on. and sliding the gamma control around gave me the predicted response on the pattern. The avia pattern would behave slightly differently if I pulled it in . I believe I'd have to set a gamma of around 1.2 to get the 2.2 patch to match. Thats assuming its taking video gamam into account. The problem with the other patterns is that they don't necessarily do this.

    I'll have a fiddle and get back to you.
     
  9. Gordon @ Convergent AV

    Gordon @ Convergent AV
    Distinguished Member AVForums Sponsor

    Joined:
    Jul 13, 2000
    Messages:
    14,214
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    166
    Location:
    Living in Surrey, covering UK!
    Ratings:
    +3,057
    Keith,

    Being a glasses wearing freak all I have to do is take my specs off and it becomes quite simple to do the gamma patch comparison! Any other glasses wearing readers should try it....

    Gordon
     
  10. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,190
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,229
    Same here Gordon: although my right is way better (worse) than my left.


    Had a play last night.

    Checked the gamma patch on my CRT TV using my normal panasonic A350 player: correlation was around 2.5-2.8.

    Checked my projector using my HTPC (using scoped settings for the card) gamma correlation on the patch was around 1.8

    Went back and using the Avia test patterns I reset black and whitepoint on the projector (not touching the HTPC) Also had a look at colour. Projector (ae100 ) now has the following settings.

    Picture: Normal
    Brightness: +2 (before filter this was a negative value)
    Contrast -17 (dropped slightly relative to previous but I have also upped some gains)
    Colour temp: Warm
    Red: +5 (down on previous )
    Green:+31 (maxed out previous value was around 20)
    Blue:+31 (maxed out same as before post filter)

    So I've swapped some red for green and dropped the overall whitepoint (probably into the peak of the red hump non-linearity in the grayscale hence the cancelling of the extra green)
    Colour looks excellent on all Avia tests : blue is bang on: red is nearly there(flashing patches still slightly visible) green similar to red. Put the slight deficiencies in red and green down to the lesser accuracy of the red and green filters relative to blue. (there as good as they get with regard to varyng saturation and hue levels)

    So I'm happy with the projector settings now to tackle gamma.
    The scoped setting for the HTPC gives the gamma value as 81. This produced good results on my CRT monitor so I know that any deficiencies in this area are down to the projector rather than the HTPC. Although I can't adjust gamma on the projector as far as I'm aware (the picture mods are probably some sort of custom non-linear curve response rather than a gamma) I can modify gamma on my HTPC (hopefully without introducing banding).

    So I drop the gamma on the HTPC all the way to gauge behaviour. The gamma chart shows the patches overall getting darker and bringing the correlation up to nearer 2.2 so the gamma control is good. The patch itself is a little problematic as you are dealing with nonlinearity in the grayscale which makes brightness perception a bit trickier ( hint try the green filter from Avia to remove colour casting to better gauge intensity). Again the patch is a little problematic in use as you end up with fairly subtle intensity changes around your correlation area. Also its imperative that black and whitepoint are set accurately as this will also have a bearing on where the correlation is beyond the gamma control. Take your time and try to look at average brightness not solidity. Initially with the gamma slider as far down as possible I though I was getting a correlation around 2. As I played with the slider it gradually became apparent that the correlation point was actually a little higher than I initially percieved it so I was able to raise the value to 27 (from 81 on scoped). Correlation is now somewhere between 2-2.2.
    (remember my CRT gamma : I think as well I perceived the correllation a little higher than it was whitepoint probably off)

    So how does it look: Colour is good but it always has been really: although I 'm now confident the filter isn't damaging the colour response too much. I know I'm not losing any intensities near black or white ( I went back and checked after gamma tweak and the values were still good) The picture now looks a bit punchier but also smoother: more like an analogue picture rather than a digital one.

    I do look at the image and notice what I'm confident are the disadvantages of video compared with film: ie more clipped whites and way less detail in the blacks relative to film. However I don't think that the projector is losing anything from the video source and indeed I notice the exact same properties on my CRT monitor. (Avia , Optimode and VE all show a healthy intensity response)

    I estimate the display gamma on the ae100 to be around 3.3 meaning you need to feed it something with a gamma of the inverse of this (0.3 and a bit) as opposed to around 0.45 to feed a CRT type device. ( this bit of cod estimation is based on assuming my HTPC is near 0.45 for the scoped settings and the initial low correlation on the projector 1.5 approx)

    I also then went back and recalibrated Dscaler (minimal differences although Dscaler quality on my system is only what I would describe as functional rather than good: OK for gaming).

    My HTPC settigns are now:
    Brightness :8 (scoped)
    Contrast :100 (scoped)
    Saturation 104 (scoped 100)
    gamma: 27 (scoped 81)
    Hue:5 (scoped 0)

    Radeon VE (6025 driver: W2K)

    Hope this helps
     
  11. ROne

    ROne
    Active Member

    Joined:
    Dec 21, 2000
    Messages:
    781
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    18
    Ratings:
    +8
    That's very useful and interesting.

    My projector exhibits similar behaviour to that, bringing my gamma on the HTPC down to 0.85 (curve) rather than a slider value gives a gamma level of about 1.8-2.0 in AVIA and thats the best I can get.

    It's plainly or painfully obivious that my more expensive 16:9 LCD doesn't have as accurate gamma as the Panasonic - which from a visual perspective having had both side by side is what I saw. It seems to me I have to tweak the hell out of my unit to get a decent picture but the Panasonic is not bad from the box.

    However I tried some further tests, i have to run my HTPC through a transcoder to YUV (my projector has problems with the RGB of the signal it has to be 720p). And this little box clips my blacks and whites at a gamma of of lower than 1.0. So this creats a right old mess in the chain as there is no adjustment on the box.

    So I have to forsake VGA card accuracy by altering the black and white level on the radeon a few percent. I get my levels back again and try for accurate gamma and 1.8 is pretty much the best I can get. So I summise that my projector is incapable of dealing with a accurate gamma response, and that's pretty much it.

    But your set ups have been very useful and I'm going to get the panasonic back again to try along side.

    Interestingly I notice with my gamma set "as accurately as possible" I would have subjectively conceeded that shadow detail was lacking and there is a temptation to bring the gamma up to boost this which has been throwing me off.

    Do you have LOTR to give me an idea of what you see in the dark scenes and how happy you are with them?

    Cheers
     
  12. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,190
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,229
    I do and I'll have a look but it may not be for 8 weeks as I doubt I'm going to be anywhere except work for any meaningful length of time.

    All I will say is that the gamma pattern will not be anywhere near accurate if the black and whitepoints are not set correctly.

    This would exhibit itself as clipping when you manipulated the gamma slider.

    Also don't be too hard on your projector ( I wouldn't call 3.3 an accurate gamma for the panny) and the gamma pattern itself is fairly tricky to read.
     

Share This Page

Loading...
  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice