Re: grayscale

Subject: Re: grayscale
From: Pierre Abbat (phma@oltronics.net)
Date: Sat Apr 21 2001 - 17:53:59 CDT

• sorted by: [ date ] [ thread ] [ subject ] [ author ]
• Next message: Pierre Abbat: "Re: refcounting images (was Re: Inline Images??? [ was: Re: Graphic Images ])"
• Previous message: Jesper Skov: "Bug status week 16"
• In reply to: Patrick Lam: "grayscale"
• Reply: Pierre Abbat: "Re: grayscale"

On Sat, 21 Apr 2001, Patrick Lam wrote:
>Note that we use the formula (red+green+blue)/256 for grayscaling (at
>least in PS output), but that's not the correct ratio that corresponds to
>human colour perception. What is the correct ratio? A quick google
>search didn't turn up the result.

Here are the conversion matrices between RGB and YCbCr as used in JPEG:

( RGB goes from 0 to 255 each. YCbCr go normally from -127 to 128.
  Conversion matrices :
   0.299 0.587 0.114
  -0.168735891648 -0.331264108352 0.5 RGB to YCbCr
   0.5 -0.418687589158 -0.0813124108417

   1 0 1.402
   1 -0.344136286201 -0.714136286201 YCbCr to RGB
   1 1.772 0

    299/1000 587/1000 114/1000
   -299/1772 -587/1772 1/2
      1/2 -587/1402 -114/1402

      1/1 0/1 1402/1000
      1/1 -25251/73375 -209599/293500
      1/1 1772/1000 0/1

  The largest value of Cb for Y and Cr 0 is 0x48; of Cr for Y and Cb 0 is 0x5b. )

phma

• Next message: Pierre Abbat: "Re: refcounting images (was Re: Inline Images??? [ was: Re: Graphic Images ])"
• Previous message: Jesper Skov: "Bug status week 16"
• In reply to: Patrick Lam: "grayscale"
• Reply: Pierre Abbat: "Re: grayscale"

This archive was generated by hypermail 2b25 : Sat Apr 21 2001 - 17:49:06 CDT