-
Kona’s interpretation of gamma differences between 601 and 709.
(precursor my reference point is from a PAL perspective)
I’m not convinced Kona has the gamma interpretation between 601 and 709 space correct. This is amplified even further when you throw RGB and YUV space conversions in to the mix.
My understanding is there is almost no perceptual difference between 601 and 709 (gamma values of 2.20 versus 2.22). This opinion was also supported by a color scientist that I recently met, stating that the difference is so small many facilities work to a 709 spec regardless of SD or HD finish.
I’ve attached 2 images which shows the Custom LUT graph in Kona. If you open custom LUTS (command ‘L’) you can access this and make sure you uncheck the checkbox to use Custom LUTs leaving the window open in the finder and you can see what the Kona is doing to the image as you change the gamma from 601 to 709 or change your video i/o option in the software controlling Kona.
Kona 601 Gamma Curve
Kona 709 Gamma Curve
As you can see if you are familiar with working with curves for doing color correction there is a considerable difference to they way Kona is manipulating the image output. This results in a fairly noticeable gamma adjustment to the image.
Is this correct?
Anyone?I don’t think it is. For mine, it seems that it’s the 709 that is wrong. I’ve never had any issues with anything going out SD (using the Kona 601 output).
On the theory there is no perceptual difference (0.02 on the gamma scale) a workaround is to force the Kona gamma setting in the control panel to 601 for a more accurate HD preview.
Is this a problem, is it something Kona need to look at, has anyone else found this and a workaround or is it 100% correct and I’m worrying about nothing and there is a massive difference between 601 and 709 space and if you are converting between you need to do a global conversion that replicates the difference between these 2 curve graphs (could bake a LUT to achieve this).
The interesting variable is over the years Quicktime updates can influence this behaviour and I reckon there was one point in time that this large variation didn’t exist.
Opinions, theories, someone from Kona?
-
4 Replies
-
Grant,
I can offer this. We are a third party using the AJA hardware and have been for about 8 years now and we have found all aspects of the hardware to be 100% accurate. Feature film and television is very demanding and MUST be accurate and this is where we sit with Rave (the product). A problem would have exposed itself many years ago and in fact when one had, it was not the AJA hardware.
Now I can also say that if you are doing your comparisons using quicktime, especially when it comes to RGB, you must be very careful, as QT has a track record of its issues.
If you are doing these comparisons thru software scopes, you must also be careful there so as to not add to in-breading to your testing. Hopefully you are not and your using an external one.
These are my two cents and rather than posting like this I would recommend sending it direct to AJA to see what they can have to offer.
We did a comparison many, many years back between 601 and 709 and it showed exactly as it should have. Will try to dig it up, but no promises.
Cheers,
Ramona -
Grant,
I am unable to dig out the comparison, like I said many years ago. In-breading should be inbreeding (sorry, late night response to the post).
The other consideration I did not add was the understanding of full range and SMPTE range 64-940 for the rec 701 files and knowing what you are exactly getting with each file to make an accurate comparison.
And like I said before, use an external scope and you should figure out what is going on.
Cheers,
Ramona -
Thanks for your input Ramona.
-
Anytime. Hopefully it is enough information (I know rather vague) to get you some answers so your not chasing your tail.
Cheers,
Ramona
Reply to this Discussion! Login or Sign Up