CCU control controlling LUT vs camera itself

Recently, someone made the claim to me that using the ATEM switch and adjusting the color of a BMD camera only does applies a color correction of sorts but doesn't actually adjust the sensor performance. Does the adjustment merely apply a LUT or does it actually adjust the sensor?

That is info that only BMD would know for real. But i guess the processing of the sensor. Because a color correction will be in the recording and in the output. While normally you can apply a different LUT to the output.

But processing nowadays is complete different then back in the day. When cameras had 3 chips you would play with the analog gain of the chip of each color.

These days it is a correction to the de-bayern pattern formula's to create a color offset. Because it is 1 sensor instead of 3 sensors that are responsable for Red Green and Blue.

If the Atem had an input color corrector it wouldn't be a secret.

As far as the studio cameras are concerned, the output is definitely changed. You can change the settings then disconnect the SDI and the settings are maintained in camera until power cycle.

How CCU control affects in-camera recording Vs. output I can't say definitively. Raw would be upstream, Prores would be downstream, Braw, I don't know, could be either.

Good Luck

Edit: I re-read the question. Yeah kinda, Any adjustments are post debayer, post Rec709 matrix, if they happen upstream of the 10bit 4:2:2 encoding is anyone's guess.

To be clear, I'm not talking about a hidden color corrector: I'm well aware that the changes are being made in the camera. How they are being made, however is quite important to the discussion I'm having.

I dont think you will ever get the real answer.. But i would place my bet on the processing of the Sensor and not the LUT.

To be clear, I'm not talking about a hidden color corrector: I'm well aware that the changes are being made in the camera. How they are being made, however is quite important to the discussion I'm having.

Which camera?

To be clear, I'm not talking about a hidden color corrector: I'm well aware that the changes are being made in the camera. How they are being made, however is quite important to the discussion I'm having.

Which camera?

I'm primarily referring to their broadcast and studio cameras but also the Mini Pro would be relevant.

Yeah, I don't know. The Ursas can all record in log 12bit 4:4:4 so if the CCU is using an intermediate X, Y, Z space/LUT it's got to be big enough to encompass XQ which makes the quality question largely moot. 16/12 bit sensor space or 12 bit 4:4:4, same smell, different data, 4:4:4 is less efficient, that's about it.

Broadcast cameras are going to make most adjustments in the DSP which is mostly 14 bit. Functionally I think it goes White Shade, Gamma, Gain, Knee, White Balance, Matrix-ish.

For the Rec709 Studio Cameras I can see where some functionality could be "unlocked", my personal testing has proven that the CCU isn't going to do it.

Good Luck

Yeah, I don't know. The Ursas can all record in log 12bit 4:4:4 so if the CCU is using an intermediate X, Y, Z space/LUT it's got to be big enough to encompass XQ which makes the quality question largely moot. 16/12 bit sensor space or 12 bit 4:4:4, same smell, different data, 4:4:4 is less efficient, that's about it.

Broadcast cameras are going to make most adjustments in the DSP which is mostly 14 bit. Functionally I think it goes White Shade, Gamma, Gain, Knee, White Balance, Matrix-ish.

For the Rec709 Studio Cameras I can see where some functionality could be "unlocked", my personal testing has proven that the CCU isn't going to do it.

Good Luck

Can you elaborate here. Is the DSP post-sensor? Is a LUT essentially being applied with these adjustments?

The DSP is the part that takes the 14 bit linear from the sensor A/D and mashes it down to 10 bit 709. On a broadcast camera you have control of most data pre encoder at the DSP stage. That's the whole deal with a knee and why it's such a big deal that BM cameras don't have them.

The DSP is the part that takes the 14 bit linear from the sensor A/D and mashes it down to 10 bit 709. On a broadcast camera you have control of most data pre encoder at the DSP stage. That's the whole deal with a knee and why it's such a big deal that BM cameras don't have them.

So, it's fair to assume that they are doing the same with their cameras as a broadcast camera? All those black, white, gamma, etc levels are at the DSP end?

I don't think so. Mainly because there are so many different sensors and so many different types of processing. The Fairchild 4.6 and HD are dual 11 bit rolling, Cmosis is 4K global, Aptina is 4K rolling. I think only the Aptina has old fashioned analog gain. My guess is an XYZ intermediate space.

An easy test would be to put a G2 in 709 mode and plug it in, point it at a light, turn the contrast down, if you see a clipped shelf the adjustment is post encoder. If there's detail as you turn the contrast down then the adjustment is pre-encoder.

Good Luck

You guys know there only 1 FPGA behind the sensor that is ingesting all the data from the sensor and outputs to SDI and flash card etc..

So there wont be many DSP processing going on.. As the data is handled by the FPGA..

Sure, we're taliking about the when,where, possibly how in the image pipe the CCU adjustments are applied on BM cameras given their lack of DSP. I gave an easy method to test it a few posts up, if only I knew somebody with and Atem and some Ursai.

Good Luck