Test results showing how FCP7 running on 10.6 messes with gamma

Question

I've been running some simple tests to determine how FCP7 handles gamma when running 10.6.4.
Note that 10.6.4 now uses a gamma of 2.2 natively (as do PCs).
I created three images in targa, png and tif (also tried exr, but FCP won't read that) that contained squares colored in different shades of gray, solid colors and mixed colors.
This download link is good for 7 days:
https://www.yousendit.com/directDownload?phi_action=app/directDownload&fl=SWhZek ZpTk1IcWR1a29CcFA0aU9SMHVDeVNFRkF0Qm1kRmM2aXU1dg
When I brought these into FCP, the results were not good.
FCP changed the gamma considerably on them, so that all the colors were WAY off (127,127,127 became 145,145,145)
I was able to ALMOST fix this by applying a gamma filter (NOT the gamma correction filter, totally different) with a value of .82.
Now, in a 0.255 scale, my values were only off by one.
However, that is a massive color shift for anyone dealing with color management.
1. Is there a way to bring in an image sequence so that FCP does NOT tamper with the gamma?
2. If I must gamma correct, is there a 3rd party plug-in with more significant digits? I need more digits than .82.
I also did some testing on Quicktime.
QT10 is a total mess. When I output to a non-self-contained mov (after gamma correcting), it was WAY off.
QT7 is on the mark and matches what VLC shows.
The same goes for outputting to H.264. QT7 = good, QT10 = BAD.
I suppose I'll experiment with Compressor now to see the proper values to get ALMOST back to my original colors.
Hopefully I'm missing some hidden setting for 'give me back my original gamma'.
BTW - you can open the images I supplied in Photoshop to see the correct values.
Using the 'digital color meter' included with 10.6 is a quick way to determine the RGB values that are displayed on the screen in FCP, QT and other programs.
And a value of 127,127,127 is just an absolute value, doesn't really matter if you have a $10K reference monitor to display it on (although, I'm sure the same RGB values would look different on the reference monitor).
BTW, FCP7 and 10.5 also mess with gamma.

Answer 1

As I had such dismal results with getting h.264 from FCP7/Compressor to match my input colors when viewed in VLC and the Quicktime player, I decided to try using Premiere CS4.
Both Premiere CS4/AME and FCP7/Compressor produced h.264 that was pretty far off when viewed in VLC and the Quicktime player7, although CS4/AME was a bit closer.
As my final output is often intended for the web, I decided to try the same files when displayed using the latest Flash player.
The results were interesting.
The Premiere CS4 results were all within 1 value of being correct. 127,127,127 would end up as 126,126,126 - and the same for the other shades of grey and color combinations.
But the FCP7/Compressor display was a disaster. 63,63,63 displayed as 81,81,81.
127,127,127 displayed as 144,144,144.
So, at least for Flash viewing of videos on the web, FCP/Compressor just couldn't produce accurate colors, they were all WAY off.
I'll have to try the same tests for DVD and BluRay.
BTW - this is the very simple image I used for testing (this display is a png, I used a tga for the actual tests to eliminate any possibility of a color profile getting embedded by mistake).
http://i34.tinypic.com/29lbgo7.png
This link with the original files is valid until aug 12:
https://www.yousendit.com/directDownload?phi_action=app/directDownload&fl=SWhZek ZpTk1IcWR1a29CcFA0aU9SMHVDeVNFRkF0Qm1kRmM2aXU1dg