colorsys: RGB to YIQ transform matrix

[Edward C. Jones]

In colorsys.rgb_to_yiq, the matrix of the linear transformation used is

.30 .59 .11
.60 -.28 -.32
.21 -.52 .31

The places I checked on the Internet use

.299 .587 .114
.596 -.275 -.321
.2.1 -.523 .311

Why does colorsys use a 2 decimal place wersion of the array? What is
the actual definition of the RGB to YIQ transform?

[David Berry]

Edward C. Jones wrote:

In colorsys.rgb_to_yiq, the matrix of the linear transformation used is

.30 .59 .11
.60 -.28 -.32
.21 -.52 .31

The places I checked on the Internet use

.299 .587 .114
.596 -.275 -.321
.2.1 -.523 .311

Why does colorsys use a 2 decimal place wersion of the array?
Pass
What is
the actual definition of the RGB to YIQ transform?

A good place to start is here,

http://www.poynton.com/ColorFAQ.html

which used to contains details of RGB->YIQ and much much more on color
space transforms.

Dave

[A.M. Kuchling]

On Mon, 10 Nov 2003 22:53:58 -0500,
Edward C. Jones <ed******@erols.com> wrote:

Why does colorsys use a 2 decimal place wersion of the array? What is
the actual definition of the RGB to YIQ transform?

I'll wager there's no technical reason. Maybe whoever wrote the module
used a reference that only quoted the matrix to two places. You could submit
a patch on SourceForge to correct this.

(Hm, looks like there's no test suite for the module. That would be worth
fixing.)

--amk