1. Join Now

    AVForums.com uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Accuracy?

Discussion in 'Projectors, Screens & Video Processors' started by Gordon @ Convergent AV, Jun 25, 2005.

  1. Gordon @ Convergent AV

    Gordon @ Convergent AV
    Distinguished Member AVForums Sponsor

    Joined:
    Jul 13, 2000
    Messages:
    13,999
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    166
    Location:
    Living in Surrey, covering UK!
    Ratings:
    +2,803
    There is a great thread over at the crt forum about what gives 3 dimensionality to a projected image. This was brought up at the end of it and rather than let that thread degenerate, thought I'd just post this info for folk.

    Colour accuracy. Our television system in the UK has a colour standard specified by the EBU. They state the co-oridnates on the CIE chart that red, green and blue should be, along wiith yellow, cyan, magenta and white. Our tv system is basically made up of a black and white image that has low level colour detail slapped on top. In order for us to see the image as accurately as possible our displays should adhere to this standard as closely as possible. Seems straightforward.

    Over in the USA where they live in NTSC land they have similar system. Except the standards for red, green, blue, and the secondaries are different than for PAL.....there standards were set by SMPTE.

    Hey but wait......the standard for HDTV in USA is also different......

    So...in order to be able to replay all these standards as accurately as possible you would need a display that used three different sets of settings, each designed to display the gamut of colour for your desired source as accurately as possible. Does such a display exist......

    You see for phosphor based technology you tend to find that the manufacturer will not use phosphor with the correct properties. At least, not all the phosphor will be the correct spec. This may be due to the physical property of the phosphor they do decide to use....ie one gives off more light for a given hit of energy, or one decays faster than another. Sometimes in order to get closer to a desired spec filtering of the light may be employed. You read pretty often about colour filtered lenses on CRT projectors for instance. It's a tricky business.

    The ideal would be to have a display that was able to produce a gamut of colour larger than all the above standards but that could then be electronically manipulated to hit the exact specs dependant on incoming signal type. There is currently only one device that I know can do this. It's a single chip DLP from Samsung...yep, Samsung. Yamaha have this capability but I am unsure if they have big enough gamut to cover it all. I suspect there must be other DLP's, possibly commercial in nature that might be able to attempt this too. There are scalers that can do this remap of colourspace as well. The Teranex Xantus can have this feature loaded on to it but again you need to have a large enough range of colour in order to benefit. From speaking with Richard Ansell at Snell and Wilcox they have processors that can do this too. It's an area I'd expect we'll see more of in future hi-end display products.

    When you see those COLOUR MANAGEMENT settings in your DLP's and in some plasma's now you know what they are for. They alter the colour gamut and how it is displayed. It can be a uesful tool if you know what it's for.

    All the above is about accuracy. Of course the end result has to be something you want to watch. It's prefectly plausible that you may enjoy looking at R1 material from your DVDplayer on a display that is set up for reference EBU material (pal/r2). It may not be the most accurate though and indeed if it was set up for accuracy you may find you enjoyed it more.....so there we go.

    I think they above post was a long winded way of saying that if you read someone saying that colours are most accurate on product X you need to know what specification they are talking about for accuracy....cos if it's correct for one its wrong for the rest unless something clever is going on somewhere else......PHEW.....need a drink.

    Gordon
     
  2. Jase

    Jase
    Well-known Member

    Joined:
    Mar 16, 2001
    Messages:
    9,469
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    136
    Location:
    South Coast
    Ratings:
    +893
    The Panasonic AE700 has a Colour Management System that lets you individually adjust the Saturation, Hue, and Brightness of up to eight selected colours and save them in one of three user defined settings. This is in addition to the Contrast & Brightness adjustments for R, G & B. It also has High, Mid & Low Gamma adjustment. Other standard adjustments are available for Contrast, Brightness, Hue, Colour, Sharpness & Color Temp and different Picture Modes.

    Should be able to get a decent image with that lot (or hoof it right up!) although we've only just got ours so I'm still sussing it out. It stores different settings for interlaced or progressive sources. It might even go as far as storing settings for each signal type, Resolution etc but I can't check that until the new player arrives. :)
     
  3. Gary Lightfoot

    Gary Lightfoot
    Distinguished Member

    Joined:
    Apr 6, 2001
    Messages:
    12,436
    Products Owned:
    1
    Products Wanted:
    1
    Trophy Points:
    166
    Location:
    Surrey. UK.
    Ratings:
    +2,096
    Nice post Gordon, thanks.

    I've seen various CIE charts for DLPs, and they seem to vary a bit, The most common element seems to be green slightly towards cyan - I don't think I've yet seen a DLP CIE that had green at the target. I managed to get my old HT1000 to have red and blue primaries almost spot on, but green was a tad away as usual. How does this compare to CRTs generaly? I've posted a pic of the CIE so others can see what you're talking about.

    For those not sure - the CIE 'shark fin' is the full colour gamut that the eye can see, Within that gamut is a black triangle that corresponds to the standard being used - PAL, NTSC, HDTV for example. The white triangle is what is meaured. The gamut doesn't have to extend to the human eye's range as more saturation isn't required IIRC - the eye/brain is easily fooled so full colour can be imitated within a lesser gamut. What is important is that the standard of D65 is adhered to as close as possible for colour accuracy when reproducing a movie. That ensures accurate skin tones, grass etc etc.

    I think that if I'd spent some time adjusting the primaries within the NEC, it might have been possible to get it even closer.

    Gary,
     

    Attached Files:

  4. Gordon @ Convergent AV

    Gordon @ Convergent AV
    Distinguished Member AVForums Sponsor

    Joined:
    Jul 13, 2000
    Messages:
    13,999
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    166
    Location:
    Living in Surrey, covering UK!
    Ratings:
    +2,803
    In Garry's post the black triangle is the target and the white is what his device has measured.

    You'd think that the ideal would have been to set the standard with the points at the outside edges of the fin. However when they were set there was a limit on the technology available.The phosphor that emits that deep green light has a long decay time. So if it was used in a TV you'd end up with green folk reading the news if you'd just been watching the footie......on a Saturday afternoon. Remember the radar screen you'd see with the blip that would stay lit up then decay.......that's the same idea....not good for a TV set.....

    Gordon
     
  5. Gary Lightfoot

    Gary Lightfoot
    Distinguished Member

    Joined:
    Apr 6, 2001
    Messages:
    12,436
    Products Owned:
    1
    Products Wanted:
    1
    Trophy Points:
    166
    Location:
    Surrey. UK.
    Ratings:
    +2,096
    Interesting stuff Gordon. Is there a reason why the Digital Cinema gamut is not much greater and has an odd shape (not triangular)?

    Gary.
     
  6. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,064
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,150
    Gammut is one area that most display devices are rarely 100% accurate on. This includes most professional displays. Pretty much all of them have gammut limitations : even the best ones.

    EBU is somewhat similar to rec.709 (hd) from memory.
    SMPTE is a somewhat smaller gammut although it does not map entirely inside either of the others.

    EBU chromaticity primaries are/were the ones I'd normally use to setup a display . However having on occasion setup displays using smpte as well (broadcast monitors in this case) I'd have to say the actual differences in the chromaticity primaries is somewhat made redundant by other factors such as the coarse colour temperature of the display ( although not unrelated) and non-linearities existent in even the most well calibrated displays.These have a far larger effect on the percieved display than the differences between the chromaticity coordinates I find . For example the amount of time a display has been switched on , the aging of the tube/lamp and even things as obscure as the quality of the power supply would all produce effects more readily apparent than the differences between rec.709 and EBU maybe less so in the case of SMPTE but I doubt even this on most domestic kit
    .
    To be honest I wouldn't necessarily expect any display device even approaching affordability to have a particularly accurate or consistent gammut even between displays of the same model (sony FW900 type monitors don't for example...some are better than others but none are really consistent , some shades of blue just don't resolve well regardless of how well the displays are calibrated , they should do as they fall within the EBU chromaticity footprint but they just drop out on display).

    I suppose it depends how thin you want to slice the onion but even calibrating a display to a certain chromaticity standard does not guarantee it will be capable of rendering the entire gammut correctly.
     
  7. Gordon @ Convergent AV

    Gordon @ Convergent AV
    Distinguished Member AVForums Sponsor

    Joined:
    Jul 13, 2000
    Messages:
    13,999
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    166
    Location:
    Living in Surrey, covering UK!
    Ratings:
    +2,803
    Good work Keith. I was hoping you'd join in. I agree that there are other things that can have a far greater effect on perceived image quality. I was lucky enough to see JoeKane's demo of his little Samsung DLP that could instantly swap between all the gamuts and there was a definate obvious dfference when he did. Of course you are probably correct that none of them were perfect on all points and gamma etc.....nbut he seemed pretty confinced that this unit, at least, did what it said.

    EBU and HDTV are virtually identical with green beng more towards yellow, fractionally in UBU. smpte's NTSC has green much more towards yellow and red towards yellow as well. It's slightly skewed as you say so it's not entirely within the EBU gamut. For others following this youshoulkd be aware that all these standards share D65 as being the point on the chart that white to black should be. (ie grey). As colour is overlayed on the balck and white image it's critical that black and white is the correct black and white so that all the overlayed colours are not skewed out of whack.

    Garry: I am not sure why TI have a digital cinema gamut that is that weird diamond shape. I suspect that's one that requires some research if Keith doesn't know. We'd want to play back material in the system it was mastered for. I have a client right now who has instructed me to find a device that has a huge gamut and that is accurate in all aspects, including gamma and greyscale....and that has top HD resolution.....etc....I am unsure if such a devce exists. The search is on.

    Gordon
     
  8. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,064
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,150
    I don't know myself. Maybe its an analogue of film gamut I'll have to check.
    The dp100 doesn't manage film gamut as far as we are concerned and thats with it being fed 10bit log uncompressed digital negative rather than the Dcinema format.
     
  9. Gordon @ Convergent AV

    Gordon @ Convergent AV
    Distinguished Member AVForums Sponsor

    Joined:
    Jul 13, 2000
    Messages:
    13,999
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    166
    Location:
    Living in Surrey, covering UK!
    Ratings:
    +2,803
  10. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,064
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,150
    Its certainly a lot bigger than rec.709.
     
  11. Gary Lightfoot

    Gary Lightfoot
    Distinguished Member

    Joined:
    Apr 6, 2001
    Messages:
    12,436
    Products Owned:
    1
    Products Wanted:
    1
    Trophy Points:
    166
    Location:
    Surrey. UK.
    Ratings:
    +2,096
    Will the eye be able to spot the extra extension in the gamut? I just wonder if after all these years with PAL/NTS and HDTV if more colours are necessary other than a marketing thing?

    Gary.
     
  12. Gordon @ Convergent AV

    Gordon @ Convergent AV
    Distinguished Member AVForums Sponsor

    Joined:
    Jul 13, 2000
    Messages:
    13,999
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    166
    Location:
    Living in Surrey, covering UK!
    Ratings:
    +2,803
  13. Welwynnick

    Welwynnick
    Well-known Member

    Joined:
    Mar 16, 2005
    Messages:
    7,085
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Location:
    Welwyn, Herts
    Ratings:
    +825
    I've seen these charts many times before, but does anyone know why the chart is that shape rather then, say, a circle of a triangle?

    I understand the primary colours as corresponding to certain specific wavelengths, and all others as being a blend of them. But can anyone explain what do the x- and y- scales represent?

    Nick
     
  14. Mr.D

    Mr.D
    Well-known Member

    Joined:
    Jul 14, 2000
    Messages:
    11,064
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    133
    Ratings:
    +1,150
  15. kurtz

    kurtz
    Active Member

    Joined:
    Feb 15, 2004
    Messages:
    171
    Products Owned:
    0
    Products Wanted:
    0
    Trophy Points:
    21
    Location:
    last place you'd look
    Ratings:
    +21
    Thanks, Mr D. Nice link. I'd always taken the x-y thing on trust, but its nice to see the explanation.
     

Share This Page

Loading...