I can't imagine that it wouldn't have this scaling option. But I performed some tests.

I took a blankclip:

blankclip(color=$0000f0,pixel_type="yuy2")

coloryuv(analyze=true) # gives the values: YUV = 39, 234, 111

Then I made the following script as source (conversion to RGB24 done by AviSynth itself):

blankclip(color=$0000f0,pixel_type="yuy2")
ConvertToRGB24

I made two mpeg1 clips with TMPGEnc: bc_def.mpg without the option "output YUV as Basic YCbCr and not CCIR601" checked (which is the default). Importing (as YUY2) and analyzing in AviSynth gives:

DirectShowSource("e:\temp\test\bc_def.mpg",fps=25)
coloryuv(analyze=true) # YUV = 40, 233, 111

conclusion: clip is untouched. The second clip is with the option "output YUV as Basic YCbCr and not CCIR601" checked. Importing (as YUY2) and analyzing in AviSynth gives:

DirectShowSource("e:\temp\test\bc_ycbcr.mpg",fps=2 5)
coloryuv(analyze=true) # YUV = 27, 248, 108

conclusion: the values are changed. Some math:

The inversion of y = x/255*( 235-8 )+8 gave the incorrect values! So I used:

y = x/255*(235-16) + 16 [(0,255)->(16,235)]

The inverse is given by

x = 255/(235-16)*y - 16*255/(235-16) [(16,235)->(0,255)]

y = 39 (luma of blankclip) results indeed in y = 26.8 = 27.

conclusion: the scaling (16,235)->(0,255) is performed when the option "output YUV as Basic YCbCr and not CCIR601" is checked.

btw, I see that I misread Jorel:
I'm very sorry. I will perform some tests, to see whether this is true.

Great, you find a way to do the check ! I search for it yesterday but never found how to do that

That's a great info. I'm glad to never have the option checked

In fact, that's me that told this.

I'll do also (now that I have a way to perform it )

Hi all !
Great results Willbert !
I suspected that "output YUV as Basic YCbCr and not CCIR601" option checked is performing scale conversion but didn't find way to prove that .
After this one we have to find what is wrong (in some movies ) with colors when black turns to be dark blue or green . This problem is mentioned many times in KVCD forum but NOONE could solve this problem .
Even if we check "output YUV as Basic YCbCr and not CCIR601" option colors become more natural but still there is some visual difference .
I suppose that they are using some kind of filter for color scale conversion that doesn't affect DVD MPEG2 stream but after encoding with TMPG color lose right values .
That will be great If we could find how exectly prevent this unwanted effect .
bman

@Dialhot,

I did some more tests:

# source:
# blankclip(color=$000010, pixel_type="yuy2") # YUV = 18, 135, 127
# ConvertToRGB24

I opened this script in TMPGEnc and checked "output YUV as Basic YCbCr and not CCIR601", resulting in bc2_ycbcr.mpg. It has YUV values: 2, 136, 127 (I did it this way, because I don't know how to make this in AviSynth):

# DirectShowSource("e:\temp\test\bc2_ycbcr.mpg",fps= 25)
# coloryuv(analyze=true) # YUV = 2, 136, 127
# ConvertToRGB24

For some strange reason TMPGEnc crashes with the script above, so I loaded bc2_ycbcr.mpg directly in TMPGEnc. Now I unchecked "output YUV as Basic YCbCr and not CCIR601" and encoded to bc2_ccir601.mpg. I made the following script:

DirectShowSource("e:\temp\test\bc2_ccir601.mpg",fp s=25)
coloryuv(analyze=true) # YUV = 16, 128, 128

My conclusion is that all YUV values with Y<16 is clamped to (Y,U,V)=(16,128,12 when leaving "output YUV as Basic YCbCr and not CCIR601" unchecked.

I guess all values with Y>235 are clamped to (Y,U,V)=(235,128,12, but I didn't check this.

Conclusion: Jorel is correct!

Could you give some links, and also upload a short vob where this problem occurs?

That is not Jorel that told you that, that's me

So as I told a long time ago, bye bye limiter in the scripts, is useless

Thank you Wilbert for having done all these tests.

I see, sorry

Of course we are only talking about TMPGEnc here ...

Of course. I don't do any encodign with other tool.

But if were going to try another encoder like CCE then we should keep Limiter() in to help with the black background blocks, Right?

Don't know, I never used CCE. Accordingly to the CCE-faq at doom, it has a luminance level option. Maybe someone can do the same two tests (as I did above) to see what this option does?

Yes, CCE has this luminance level 16 to 235 or 0 to 255. I think by default it is at 0 to 255. Now I need to read again on what the two differences do