about RGB or RAW is RGB Necessary

[tj williams]

When working with the RED in ENG EFP situations the biggest problem with using 35mm image size is DOP second largest is size of glass needed to cover range expected in regular HD lenses.

Now that Red RAw 4K and 2K are supported on the Finalcut desktop I have revised my opinion about shooting these forms in RAW. under some circumstances. If the client edits on FC then the final output to HD or SD is probably about as quick as RGB? Is that somewhat true?

As far as the actual shooting is concerned:

If you put up a wide zoom say 17-85mm IS this is in the range that is hand holdable. Then if you need more telephoto and would use the doubler on an HD cam. Instead just switch the camera to 2K mode. Instantly the VF jumps to fill the VF with the 2K window. Now the 17/85 35mm lense is looking at only the center or 2K window. It is effectively 4 times longer. (approx 64mm to 340mm in equiv 2k or S16 lens) to The DOF is comparably equivalent to the same image from the same distance in HD as the 2K sensor is only slightly larger than the 2/3" chip targets in an HDCam.

Zoom focus etc. etc is from the handle by Birger so is similiar to eng efp cams. weight is similiar and low light capability in the window/doubler mode is similiar to HD cams.

When I consider shooting with this model the only place where RGB will help me is where I need instant delivery to a system which will only accept rgb HD.

[I Bloom]

When working with the RED in ENG EFP situations the biggest problem with using 35mm image size is DOP second largest is size of glass needed to cover range expected in regular HD lenses.

Now that Red RAw 4K and 2K are supported on the Finalcut desktop I have revised my opinion about shooting these forms in RAW. under some circumstances. If the client edits on FC then the final output to HD or SD is probably about as quick as RGB? Is that somewhat true?


My understanding was that the RGB format options would be downsampled (i.e. scaled) from 4K, not cropped off the sensor RAW. (Rob??).

There are several different algorithms for debayering a 4K image to 4K, some work better than others for different images, and its good to have choice in post. However once you start downsampling a Bayer image you use a scaling algorithm to scale the Bayer Pattern into smaller RGB pixels (so in a way there is no debayering per say since by debayering we really mean interpolating between different colored sensors). Thus the existance of RED RGB. RGB = Downsampled Bayer. (Rob am I talking out of my ass again?)

There is some question for me as to whether the RGB formats will have compression or not or if we are trading compression for scaling. It would be nice if the camera could do DCT compression and deliver various standard compressed HD formats to quicktimes on the drive. Just for those days when I'm slumin shooting a CEO and dropping footage to the drive of someone who has never heard of RED. (Some of these people are still "thinking" about HD).

My assumption is that this won't ever happen because this kind of feature might require a "rewiring" of the camera's field programmable gate arrays and this is probably the most difficult part of their software developement task. This is also the reason I'm guessing that the announced RGB features will be last out of the gate.

IBloom

[Stokestack]

So far we've been told that RGB is the only answer for people who want 2K-or-smaller images but want them downscaled from the full sensor, not cropped.

I'm curious as to why you'd want to switch to non-wavelet compression just because the image is RGB.

Also, I wonder if a future update will be able to downscale the 4K raw data and record it as 2K raw, keeping the full latitude. I seem to remember an earlier post saying that this isn't impossible, but hadn't been explored yet. I'd think that this would take priority over getting RGB working, given that it might render the RGB mode unnecessary.

[GlennChan]

I'm curious as to why you'd want to switch to non-wavelet compression just because the image is RGB.

I think you mean non-RAW? AFAIK, the red website indicates that redcode RGB will be compressed (it's implied in the format table).

As I understand it, the point of RGB is that you can do a decent job of de-bayering in camera. It would be a faster workflow than having to wait for stuff to go through Redcine. You also get in-camera processing like knee and detail (useful for live production, live switching). The Red team has indicated that in RAW mode, the camera will have limited menu settings suggesting that the RAW mode doesn't have as much in-camera processing/manipulation of the image.

[Daniel Gourley]

I think you mean non-RAW? AFAIK, the red website indicates that redcode RGB will be compressed (it's implied in the format table).

As I understand it, the point of RGB is that you can do a decent job of de-bayering in camera. It would be a faster workflow than having to wait for stuff to go through Redcine. You also get in-camera processing like knee and detail (useful for live production, live switching). The Red team has indicated that in RAW mode, the camera will have limited menu settings suggesting that the RAW mode doesn't have as much in-camera processing/manipulation of the image.
That's not my belief as I have understood it you have the option to load a LUT recorded from REDCINE into the cam to use as an output for monitoring, but obvously it will record RAW, but with Metadata to the associated LUT.

NE1 can confirm my slow memory???

[I Bloom]

So far we've been told that RGB is the only answer for people who want 2K-or-smaller images but want them downscaled from the full sensor, not cropped.

I'm curious as to why you'd want to switch to non-wavelet compression just because the image is RGB.

Also, I wonder if a future update will be able to downscale the 4K raw data and record it as 2K raw, keeping the full latitude. I seem to remember an earlier post saying that this isn't impossible, but hadn't been explored yet. I'd think that this would take priority over getting RGB working, given that it might render the RGB mode unnecessary.

Well, I think most people who own a RED are going to have a version of FCP that they can cut Red footage natively or use REDCine to take it wherever they want. However, lots of people are slow to change and I don't want to be a salesman for a specific proprietary format. For some productions, I just want to drop off a drive with the footage in a format that they specify, and that usually means a version of HD based on DCT compression. If RED can do that and save me the effort of running it through REDCine then I'm all for having that OPTION.

Downscaling from 4K raw to 2K raw doesn't make much sense. Remmember RAW doesn't just mean 12bit, it also means that each pixel only has one out of three colors sampled. There would be no advantage to keeping the Bayer pattern after scaling.

http://en.wikipedia.org/wiki/Bayer_filter

I don't think its impossible for the scaled RGB to be 12bit linear? I'm curious as to why?

IBloom

[Nick Shaw]

I don't think its impossible for the scaled RGB to be 12bit linear? I'm curious as to why?

Stuart (and others I think) has said previously that it will be possible to record REDCODE RGB unprocessed, so it still retains a lot of the benefits of RAW. I assume that means it will still be 12 bit linear.