looking for icc profile for Rec.601 /625/ D65

Post by Thomas Richter
Hi folks,
could someone help me to locate and/or create an ICC color profile for
ITU-T 601 primaries & gamma using the 625-line version (PAL) of the
color space, with D65 white-point.
Unfortunately, I wasn't able to find anything suitable on the net, but
probably one of the experts here has something suitable?
Background of the question is the color-handling for JFIF images. The
original JFIF spec specifies the Rec.601/625 color space, though most
applications render the color as sRGB, and I want to get an impression
how much the images would differ. sRGB colors are identical to Rec.709
as far as I know, but the latter is close, though not identical to
Rec.601. The green primary of 601 is minimally different from the 709
green primary, and I wonder how much it makes a difference.
Thanks,
Thomas

If I'm not mistaken, and you have Photoshop, the SMPTE-C profile has the
right specs. If you don't have it I can send it to you.

HTH - Steve

Thomas Richter

2010-07-20 20:56:45 UTC

Post by NotMe
If I'm not mistaken, and you have Photoshop, the SMPTE-C profile has the
right specs. If you don't have it I can send it to you.

Does that use the chromacities Gernot is listing in his post? If so, I
wonder why that differs from the chroma coordinates I have in the specs?
Light adaption? I would need it for a D65 white point.

So long,
Thomas

Gernot Hoffmann

2010-07-20 16:21:17 UTC

Thomas,

you're busy in this field now since one year. Where did you find any
information about primaries in ITU-Rec.601? Do you own the original
specifications? I don't.

Some facts:

sRGB...icc
Rx 0.648451
Ry 0.330850
Gx 0.321199
Gy 0.597840
Bx 0.155887
By 0.066040
TRC G=2.4 + linear slope

PAL_SECAM.icc
Rx 0.648449
Ry 0.330855
Gx 0.311449
Gy 0.599700
Bx 0.155896
By 0.066048
TRC G=2.2

SMPTE-C.icc as recommended by Steve
Rx 0.638864
Ry 0.340203
Gx 0.331032
Gy 0.592086
Bx 0.162906
By 0.079015
TRC G=2.2

Extracted by ICC Profile Inspector from profiles
as available in Adobe Programs.
No responsibility for the validity of these numbers.

Chromaticity comparison for PAL/SECAM versus sRGB:
http://www.fho-emden.de/~hoffmann/gamuts08072002.pdf

Best regards --Gernot Hoffmann

Thomas Richter

2010-07-20 20:54:51 UTC

Thomas,
you're busy in this field now since one year.
Where did you find any
information about primaries in ITU-Rec.601? Do you own the original
specifications? I don't.

Yes - while I haven't tried, you can surely download it from the ITU as
well, but probably not the latest revised version. However, here are the
chromacity coordinates of the ITU-T 601 primaries - note that they are
*almost*, but not completely identical to the sRGB 709 primaries:

r: x=0.6400 y=0.3300
g: x=0.2900 y=0.6000
b: x=0.1500 y=0.0600

Note that the only difference is the x-coordinate of green, which is 0.3
for 709 aka sRGB primaries. (The above is, to be precise, the 625 line
version for ITU BT 601, aka "PAL", there is also a 525 version with
different primaries specified for NTSC - yuck!)

Post by Gernot Hoffmann
sRGB...icc
Rx 0.648451
Ry 0.330850
Gx 0.321199
Gy 0.597840
Bx 0.155887
By 0.066040
TRC G=2.4 + linear slope

Up to rounding this is what I have as well, yes.

Post by Gernot Hoffmann
PAL_SECAM.icc
Rx 0.648449
Ry 0.330855
Gx 0.311449
Gy 0.599700
Bx 0.155896
By 0.066048
TRC G=2.2

Yes, and this one somehow irritates me. I would believe that this
*should* be 601-625, but it isn't. Green is off (yes, I'm sure Gx should
be 0.29, not 0.31!) What's the source of the above?

Post by Gernot Hoffmann
SMPTE-C.icc as recommended by Steve
Rx 0.638864
Ry 0.340203
Gx 0.331032
Gy 0.592086
Bx 0.162906
By 0.079015
TRC G=2.2
Extracted by ICC Profile Inspector from profiles
as available in Adobe Programs.

Hmm, Adobe you say? Ok, then at least I know whom to ask (or whom to

Post by Gernot Hoffmann
http://www.fho-emden.de/~hoffmann/gamuts08072002.pdf

Thanks,
Thomas

Gernot Hoffmann

2010-07-21 05:25:34 UTC

Thomas,
you're busy in this field now since one year.
Where did you find any
information about primaries in ITU-Rec.601? Do you own the original
specifications? I don't.

Yes - while I haven't tried, you can surely download it from the ITU as
well, but probably not the latest revised version. However, here are the
chromacity coordinates of the ITU-T 601 primaries - note that they are
r: x=0.6400 y=0.3300
g: x=0.2900 y=0.6000
b: x=0.1500 y=0.0600
Note that the only difference is the x-coordinate of green, which is 0.3
for 709 aka sRGB primaries. (The above is, to be precise, the 625 line
version for ITU BT 601, aka "PAL", there is also a 525 version with
different primaries specified for NTSC - yuck!)

Post by Gernot Hoffmann
sRGB...icc
Rx 0.648451
Ry 0.330850
Gx 0.321199
Gy 0.597840
Bx 0.155887
By 0.066040
TRC G=2.4 + linear slope

Up to rounding this is what I have as well, yes.

Post by Gernot Hoffmann
PAL_SECAM.icc
Rx 0.648449
Ry 0.330855
Gx 0.311449
Gy 0.599700
Bx 0.155896
By 0.066048
TRC G=2.2

Yes, and this one somehow irritates me. I would believe that this
*should* be 601-625, but it isn't. Green is off (yes, I'm sure Gx should
be 0.29, not 0.31!) What's the source of the above?

Hmm, Adobe you say? Ok, then at least I know whom to ask (or whom to

Post by Gernot Hoffmann
http://www.fho-emden.de/~hoffmann/gamuts08072002.pdf

Thanks,
Thomas

Thomas,

yes, I forgot to mention the Chromatic Adaptation Transform (CAT).
Please compare p.7 and p.8 here:
http://www.fho-emden.de/~hoffmann/colcalc03022006.pdf

The physical values for sRGB are on p.7: sRGB without CAT.
These are 'round' numbers:

Rx = 0.64
Ry = 0.33
Gx = 0.30
Gy = 0.60
Bx = 0.15
Bz = 0.06

So far for working space white D65 and reference white D65,
everything physically correct.

On p.8 we have working space white D65 but reference white D50.
I didn't ever understand the purpose of this modification, but then
one
needs formalistically a CAT, though no human adaptation is involved.
This CAT leads to slightly modified numbers as quoted.

The Bradford CAT is described here, p.9:
http://www.fho-emden.de/~hoffmann/cielab03022003.pdf

Bestt regards --Gernot Hoffmann

Steve

2010-07-21 19:28:09 UTC

Post by Thomas Richter
Hi folks,
could someone help me to locate and/or create an ICC color profile
ITU-T 601 primaries & gamma using the 625-line version (PAL) of
the color space, with D65 white-point.
Unfortunately, I wasn't able to find anything suitable on the net,
but probably one of the experts here has something suitable?
Background of the question is the color-handling for JFIF images.
The original JFIF spec specifies the Rec.601/625 color space,
though most applications render the color as sRGB, and I want to
get an impression how much the images would differ. sRGB colors
are identical to Rec.709 as far as I know, but the latter is
close, though not identical to Rec.601. The green primary of 601
is minimally different from the 709 green primary, and I wonder
how much it makes a difference.
Thanks,
Thomas

Thomas,
you're busy in this field now since one year.
Where did you find any
information about primaries in ITU-Rec.601? Do you own the
original specifications? I don't.

Yes - while I haven't tried, you can surely download it from the ITU
as well, but probably not the latest revised version. However, here
are the chromacity coordinates of the ITU-T 601 primaries - note that
they are *almost*, but not completely identical to the sRGB 709
r: x=0.6400 y=0.3300
g: x=0.2900 y=0.6000
b: x=0.1500 y=0.0600
Note that the only difference is the x-coordinate of green, which is
0.3 for 709 aka sRGB primaries. (The above is, to be precise, the 625
line version for ITU BT 601, aka "PAL", there is also a 525 version
with different primaries specified for NTSC - yuck!)

Post by Gernot Hoffmann
sRGB...icc
Rx 0.648451
Ry 0.330850
Gx 0.321199
Gy 0.597840
Bx 0.155887
By 0.066040
TRC G=2.4 + linear slope

Up to rounding this is what I have as well, yes.

Post by Gernot Hoffmann
PAL_SECAM.icc
Rx 0.648449
Ry 0.330855
Gx 0.311449
Gy 0.599700
Bx 0.155896
By 0.066048
TRC G=2.2

Yes, and this one somehow irritates me. I would believe that this
*should* be 601-625, but it isn't. Green is off (yes, I'm sure Gx
should be 0.29, not 0.31!) What's the source of the above?

Hmm, Adobe you say? Ok, then at least I know whom to ask (or whom to

Post by Gernot Hoffmann
http://www.fho-emden.de/~hoffmann/gamuts08072002.pdf

Thanks,
Thomas

Thomas,
yes, I forgot to mention the Chromatic Adaptation Transform (CAT).
http://www.fho-emden.de/~hoffmann/colcalc03022006.pdf
The physical values for sRGB are on p.7: sRGB without CAT.
Rx = 0.64
Ry = 0.33
Gx = 0.30
Gy = 0.60
Bx = 0.15
Bz = 0.06
So far for working space white D65 and reference white D65,
everything physically correct.
On p.8 we have working space white D65 but reference white D50.
I didn't ever understand the purpose of this modification, but then
one
needs formalistically a CAT, though no human adaptation is involved.
This CAT leads to slightly modified numbers as quoted.
http://www.fho-emden.de/~hoffmann/cielab03022003.pdf
Bestt regards --Gernot Hoffmann

Getting back to original topic: it appears that the PAL_SECAM.icc profile,
which comes with Adobe applications, has the correct primaries. Sorry I
originally I missed the reference to PAL you made. If you have Photoshop
you can check the values by:

Go to "Color Settings..." Then select the RGB profile you want to check.
Next from the same dropdown list select "Custom RGB...". This should open a
dialog showing gamma, white point, and RxRyGxGyBxBy primaries. Of course
you can create your own here too. Just change the fields you want and give
it a descriptive name and click OK. To save it make sure to select "Save
RGB..." from the same dropdown list. HTH - Steve

Thomas Richter

2010-07-21 21:15:44 UTC

Post by Steve
Getting back to original topic: it appears that the PAL_SECAM.icc profile,
which comes with Adobe applications, has the correct primaries. Sorry I
originally I missed the reference to PAL you made.

Thanks - no, I don't have Photoshop, sorry. However, the primaries in
the PAL_SECAM profile are not identical to the 601 published primaries.
Am I correct that this is because the white point is different? I always
assumed that the white point of 601 is also D65, but is it D50, thus are
we assuming that the white-point adaption is a different one and thus
one should use the Bratford adaption formula to convert the D50 adapted
primaries back to D65 ones?

Sorry for sounding stupid, but I really want to understand what is going
on here.

Thanks,
Thomas

Gernot Hoffmann

2010-07-22 06:12:18 UTC

Thanks - no, I don't have Photoshop, sorry. However, the primaries in
the PAL_SECAM profile are not identical to the 601 published primaries.
Am I correct that this is because the white point is different? I always
assumed that the white point of 601 is also D65, but is it D50, thus are
we assuming that the white-point adaption is a different one and thus
one should use the Bratford adaption formula to convert the D50 adapted
primaries back to D65 ones?
Sorry for sounding stupid, but I really want to understand what is going
on here.
Thanks,
Thomas

Thomas,

here comes a screenshot of ICC Profile Inspector's interpretation
of PAL_Secam.icc:
Loading Image...

MediaWhitePointTag: numbers are that of D65
Illuminant: numbers are that of D50

MediaWhitePoint = Working Space White for those RGB-spaces
which are usable as working spaces (sRGB, AdobeRGB,
Prophoto-RGB...).

Illuminant D50 is in this case (opposed to CMYK profiles) a
fictious parameter, which confuses the interpretation.
I had already mentioned this for sRGB.
The physical primaries of a monitor have nothing to do with
'Illuminant', and the choice of the monitor white (MediaWhite)
doesn't shift the primaries.

Please download the free ICC Profile Inspector. Perhaps we
have good luck and the true chromaticity coordinates are
meanwhile available. Otherwise one has to do the calculations
by one's own programs (as I did e.g. for sRGB).

Best regards --Gernot Hoffmann

Thomas Richter

2010-07-30 21:07:54 UTC

Post by Gernot Hoffmann
here comes a screenshot of ICC Profile Inspector's interpretation
http://www.fho-emden.de/~hoffmann/pal-secam.png

Thanks to all, got what I needed. I found that the primaries PAL_SECAM
profile is indeed only one Bradford transformation away from the 601
primaries I found in the ITU document. Thus, indeed, the difference is
the CAT.

If anyone cares, the difference of using the 601-D65 adapted colorspace
compared to an sRGB colorspace is indeed pretty minor. It is barely
visible, and probably too large for a fully profiled workflow, but small
enough for the casual observer.

So long,
Thomas

Gernot Hoffmann

2010-07-29 17:36:27 UTC