Video in canvas plays darker, playhead stops image lightens*

. . The video in canvas plays darker, when I stop the playhead the image lightens* . I've come across others mentioning this before but can't remember the reason/fix!

Anyone know how to solve this?

Cheers,
WM

What codec? This happens if you are using a non-FCP codec like DNxHD or H.264. This can also happen if your hard drive is slow...FCP lowers the quality in order to not drop frames.

And you do know that the FCP Canvas and Viewer are not color accurate to begin with, right?


www.shanerosseditor.com

Listen to THE EDIT BAY Podcast on iTunes
[itunes.apple.com]

Quote

And you do know that the FCP Canvas and Viewer are not color accurate to begin with, right?

yes, thanks shane - I am aware of this. the light-dark stop-start issue has only been occurring since i've been working with footage transcoded for premiere, which as you say may explain it.

Cheers,
WM

Quote
Wanted Man

Quote
Shane Ross
And you do know that the FCP Canvas and Viewer are not color accurate to begin with, right?

yes, thanks shane - I am aware of this.

Not true. Not true. How can we stamp out this myth?

With a calibratable monitor and good calibration procedure, the FCP Canvas can be perfectly color accurate -- as perfect as your colorimeter. I've posted several times about this, like here, here, here, here...

I enjoy quite accurate BT.709 color on the FCP7 canvas on a consumer NEC PA301W monitor. It proved tricky because of flaws in the NEC SpectraView II software (or its interaction with the Mac). You must know a bit to calibrate. But you should also know a bit to say you can't calibrate.

Dennis Couzin
Berlin, Germany

dcouzin Wrote:
-------------------------------------------------------

> Not true. Not true. How can we stamp out this
> myth?

By showing the Canvas and the Viewer side by side with a properly calibrated broadcast monitor. Have a measuring tool measure the colors on both and show the difference.

> With a calibratable monitor and good calibration
> procedure, the FCP Canvas can be perfectly color
> accurate -- as perfect as your colorimeter.

A perfectly calibrated COMPUTER DISPLAY doesn't show the same colors as a properly calibrated VIDEO MONITOR. The signal being fed to the computer display is different than the one sent to an HDTV or broadcast monitor. If you take HDMI out of a laptop, for example, into an HDTV, and did full screen playback. And then did the same from a video IO device, like the AJA IoXT to the same HDTV...the colors would be different.


> I've posted several times about this, like here, here,
> here, here...
>
Doesn't mean that you were right. The HDMI out of the computers...or Thunderbolt out to a TB monitor...doesn't send a proper Rec 709 signal.

> I enjoy quite accurate BT.709 color on the FCP7
> canvas on a consumer NEC PA301W monitor.

You got Rec 709 on a computer display...displaying a computer signal (the FCP interface)? tricky...


www.shanerosseditor.com

Listen to THE EDIT BAY Podcast on iTunes
[itunes.apple.com]

Shane, as I've explained before, the calibration of the computer monitor must take into account the idiosyncracies of the "player", in this case FCP7, the OS, and the graphics card. My method of calibration involves putting a test clip into FCP7 and measuring the colors on the canvas. When the various Y'CbCr in the test clip yield correct colors on the canvas calibration is complete. Here is such a test clip for download. It is uncompressed 8-bit but can be converted to ProRes with no color change in the FCP7 canvas.

Quote
Shane Ross
The HDMI out of the computers...or Thunderbolt out to a TB monitor...doesn't send a proper Rec 709 signal.

"Proper Rec 709 signal" is a red herring here. The shape of the signal as measured with an oscilloscope, how the digital data are pulsed, line synchronization and all that, are matters of electronics (television) engineering quite separate from the color information conveyed in the signal. (In fact the BT.709 signal carries Y'CbCr information still needing to be decoded to make colors.)

Understand how BT.709 works with color information. BT.709 consists of some equations and some colorimetric constants which together indicate what CIE XYZ shall be displayed for any particular Y'CbCr in the video. (The one hole in BT.709 is the missing playback gamma; see EBU Tech 3320 which recommends 2.35.)

It turns out that FCP7 + the Mac OS + the graphics card do not make a horrible mess interpreting the Y'CbCr in the clip, which if they did would require a 3D LUT to correct for the monitor. It is enough to calibrate the monitor to emulate the BT.709 primaries and to have the desired playback gamma.

Quote
Shane Ross
By showing the Canvas and the Viewer side by side with a properly calibrated broadcast monitor. Have a measuring tool measure the colors on both and show the difference.

They don't have to be side by side if we have decent colorimeters. You can run the test clip on your system and measure the colors produced on your broadcast monitor in Los Angeles. I can run the same test clip on my system and measure the colors produced on the FCP canvas on my computer monitor in Berlin. And then we compare our numbers XYZ, and compare them with what BT.709 requires them to be for the 32 frames in the test clip.

Dennis Couzin
Berlin, Germany

I vaguely remember this issue.
I think it was with video levels too high in my case.
By applying a video safe filter it cured the issue.
but this is quite a few years ago.