• AJA Video Systems

  Gamma shift issue with Aja-Io

  Posted by Uli Zentis on October 8, 2005 at 2:33 pm

  Hi,

  when putting out footage to tape with Final Cut Pro 4.5 and Aja Io, I experience different color/luma values which seem to be caused by an unwanted gamma shift.

  Comparing a psd/tiff-file filled with 50% grey dragged into the sequence with a 50% grey test pattern from our Deck, Io displays a brighter image.
  FinalCut-vectorscope shows 57%, an external waveform-monitor displays nearly 400mv instead of expected 350mv.

  I created a test-pattern (again in Photoshop) with different color-bars fading to black. Aja-Out again looks brighter, so I compared Aja-Out with SDI-Out of another box using DigitalVoodoo. When picking the colors in a Quantel Hal, Digital Voodoo shows nearly (+/-1) correct values, Aja-Io displays much brighter midtones.

  Depending on what footage is dragged into the timeline (differences in PC- / Mac-generated psd files, used codecs in movies), gamma-values differ the same way in Aja-Out and in the image shown in the sequence itself. Cross-Checks around our boxes show same results.

  The setup:

  G5 2x2GHz, 2gig RAM
  Aja-Io (tested out 1.3.1-2.0)
  tested Quicktime 6.5.2 – 7.0.2
  using IoDigitalPal-Drivers / sequence set to PAL CCIR601

  Tested with Tek WFM601A and 7.3.2b1 DigitalVoodoo and Quantel Hal.

  Does anybody know about that behaviour? I mailed to Aja- and Apple-Support a few days ago, but perhaps there is a quick solution out there to switch off all gamma shifting in 422-uncompressed sequences?

  Thanks,

  Uli Zentis

  Mark Beazley replied 20 years, 7 months ago 2 Members · 3 Replies
• 3 Replies

• 

  Mark Beazley

  October 10, 2005 at 12:28 am

  FCP applies a gamma curve to all imported footage when using DV, uncompressed 8bit or uncompressed 10bit. Run your photoshop through a gamma filter. I saw the same problem when I transitioned from Media 100 to AJA/FCP.

  Adam Wilt emailed me the following:

  Try applying a gamma correction to the test patterns (it’s either 1.2222 or 0.81818; I forget which one you use in which direction off the top of my head. Chris Meyers says he uses 1.228 and 0.824, FWIW). You should be able to get everything back on target this way.

  That should correct the problem.

  -mark
• 

  Uli Zentis

  October 10, 2005 at 5:24 pm

  Hi Mark,

  thanks for your comments! Putting a gamma filter with the given values onto our footage in FC will set nearly exact RGB-values compared wih our voodoo-card. FinalCut rounds up the last two digits of the gamma values, this could explain why some colors differ up to +/- 3 in each channel. I did not get exact values, but this will help.

  But one thing confuses me: I it a *bug* ? If yes, will it be fixed? Or are there just two worlds of editing, a bright and a dark one? I guess there are many Aja-Users out there who are not aware of their tapes differ from the rest of the world.

  Thanks a lot, Uli Zentis
• 

  Mark Beazley

  October 10, 2005 at 8:22 pm

  I don’t think it is a bug. If you digitize material with FCP and output from FCP you should not need any gamma fixes. If you bring clips into After Effecta and/or create graphics with AE, then you need to apply a gamma curve to the incoming AE video files and the subsequent file you make when you render something. At first it is a pain, but I’ve got the filters saved as Favorites in AE, so it really does not take much more time.

  I also believe any trans communication between Apple Pro apps you should be fine also.

  -mark

Reply to this Discussion! Login or Sign Up