My New Year’s Resolution to fix Final Cut Pro

[Jack Tunnicliffe]

I gave up a couple of hours of work this morning to do some tests. I can’t say that I came up with the results I expected. I had a ton of rarely used components in my quicktime library like microcosm, DivX, sorenson, flip4mac, etc. I thought removing all of those might get me back to some normal After Effects renders but in the end no change what so all.

I removed all the AJA components and rendered a Macbeth chart. If you aren’t familiar with a Macbeth it is a series of chips with RGB color values assigned plus a grey scale across the bottom to check gamma.

In the end there was no difference in the QT files rendered as either 10 bit uncompressed or as Apple Intermediate. They both looked virtually the same coming back into Final Cut. The apple intermediate render retained the correct RGB values, very similar to the source file. Certainly good enough for an HDV project.

The problem here is that while they display accurately on the Final Cut Canvas. I used the digital color meter to take the values, they display with a gamma shift when displayed on the color grading monitor. The canvas looks correct and the monitor looks washed out.

Dropping a v210 render into an HDV sequence results in proper output on the grading monitor. (I did not test the grading monitor output but cutting back and forth between AE output and FCP output show the same gamma) but now the canvas is far too dark in FCP.

So, in the end, I’m not sure that I accomplished anything this morning other than to give myself yet a bigger headache. I’m going to send my results to AJA and see what they have to say about this. I am not accustom to working in an HDV project, usually uncompressed 10 bit, but why can’t I when I need to. This particular client is producing a lower budget show for a cable network and is working in HDV. He would like to have the intro I am creating back as an HDV QT but how can I even begin to guess that my settings are accurate. Yeesh!

Sorry to open this can of worms everyone. Again appreciate any feedback.

Jack Tunnicliffe

[Jeremy Garchow]

Yikes! all you need to do is remove the AJAUncompressedcodec.component before you launch After Effetcs. I actually have a little script that does this for me. You might get some small color shift during the yuv>rgb>yuv conversion.

It works for me without the washed out monitor, I should add. Can you post a few seconds of that Macbeth chart on yousendit or something?

I have not done this in HDV, though.

Jeremy

[Jeremy Garchow]

Also, you’re on CS3?

[Jack Tunnicliffe]

I removed everything AJA as well as any other third party codecs for the renders out of AE only. This was to get back to QT square one. The washed out monitor existed in FCP with all the AJA components back in the QT folder. I may not have explained myself clearly.

I do a fair bit of beta testing so I have this need to always get back to factory settings and start putting things back in, if you know what I mean, so I have a touchstone to start from.

I forgot to mention that the v210 in and HDV timeline displays properly on the grading monitor but the RGB values are too low (crushed down) on the canvas.

Bring all these renders back into AE showed that the only one that went out of AE accurately was the v210. I brought them back in and stacked them on top of the original Macbeth chart and they all had lifted gamma compared to the source chart with the exception of the v210.

Here’s a link to a Macbeth chart. It’s pretty accurate in AE. Don’t measure its values on the desktop in preview as preview doesn’t resolve it properly. I’ve also posted a chart of what they should be. Actually 2 charts. I notice that chip 12 is out of spec. Not sure why.

Here it is:

https://web.mac.com/java009/iWeb/Macbeth%20charts/Macbeth%20charts.html

[Jeremy Garchow]

That’s the specific charts you used?

[Jack Tunnicliffe]

Yes it is. It’s just a jpeg but is very, very close to accurate values. I’ve used it before when speaking color language to both AJA and Apple engineers and they have not argued about it’s accuracy. It certainly displays the issues that we are seeing on the user end.

[Jeremy Garchow]

I can only save it as a web archive and not a jpeg. Have you tried this with video and not a low res jpeg set to whatever color space?

[Jack Tunnicliffe]

give me your direct email and I’ll send to you. That’s funny that you can’t download the original.

[Jeremy Garchow]

editstation at gmail dot com

[Jack Tunnicliffe]

I thought I would close the loop on this thread by informing you of some new information that has been forthcoming. First of all my sincere thanks to Jeremy Garchow of Chicago who worked with me through a battery of Final Cut and After Effects tests. The result of these tests is that gamma shifts are being caused in round trips from FCP through AE and back to FCP due to the AJA uncompressed codec component being installed in the QT library folder. Removing this component results in perfect round trip renders for both Uncompressed 10 bit and 8 bit files as well as for ProRes. It is no longer necessary to use the AJA v210 codec for round trips to compensate for colour and gamma shifts. AJA engineering is looking into this issue and will likely release a fix or update in the near future. It is safe to leave this component out of the QT library folder unless you are calling up a project that has legacy files. You may have colour or gamma shifts if the component isn’t installed but for new projects work with it removed.

Thanks everyone for your feedback and help on this.

Jack Tunnicliffe
Java Post Production