Scarlet sharpness workflow question

[Jim Coulter]

I posted on this issue as well. I had shot some aerials of an oil rig using an 18mm Red Pro Prime. When I looked at the footage in RCX at 100% I was shocked! I thought I had somehow shot everything 'out of focus' and I was really worried that my client would kill me since the chopper had already left the oil platform. Looking at the footage 'full frame' everything looked fine but at 100% it was definitely SOFT! When I got back to Houston I met up with a friend who owns an Epic and we did a test with his Zeiss compact primes. We loaded them into RCX and popped into 100% and guess what? His was just as soft. I felt vindicated that I hadn't shot my aerials out of focus but also very upset that the image was not sharp. I began to wonder if 100% on RCX was in fact a 'magnification' of the image in which case the softness would be understandable. But IF in fact I were looking at the true 100% image in all it's glory- I would be very underwhelmed.

Are you saying that 100% in RCX the image should be clean and sharp? I have also done some research after posting my experience and several people that seemed knowledgeable stated that it was normal and that was how RED's debayering affects the image. Now I'm not sure what to believe.

[Matt Ryan]

I posted on this issue as well. I had shot some aerials of an oil rig using an 18mm Red Pro Prime. When I looked at the footage in RCX at 100% I was shocked! I thought I had somehow shot everything 'out of focus' and I was really worried that my client would kill me since the chopper had already left the oil platform. Looking at the footage 'full frame' everything looked fine but at 100% it was definitely SOFT! When I got back to Houston I met up with a friend who owns an Epic and we did a test with his Zeiss compact primes. We loaded them into RCX and popped into 100% and guess what? His was just as soft. I felt vindicated that I hadn't shot my aerials out of focus but also very upset that the image was not sharp. I began to wonder if 100% on RCX was in fact a 'magnification' of the image in which case the softness would be understandable. But IF in fact I were looking at the true 100% image in all it's glory- I would be very underwhelmed.

Are you saying that 100% in RCX the image should be clean and sharp? I have also done some research after posting my experience and several people that seemed knowledgeable stated that it was normal and that was how RED's debayering affects the image. Now I'm not sure what to believe.

Soft is one thing but out of focus is another. You can work with a soft image to sharpen it but an out of focus shot is not salvageable regardless of what you do. The original posters question was about his specific images, which lack focus. Of course the image will be a bit soft since it's RAW but your usually able to get more detail when adding contrast back into the image. Look at Phil's image example, this is the sharpness I would expect and what I typically get from the camera.

When it comes to sharpness you have to remember these stills are coming from motion, typically at 24fps with a 180degree shutter which cause motion blur with any movement. These are factors that will cause the single image to appear "soft". Then you have to consider user error with focus. Even a fraction of a turn off and in 4K it's bound to look soft...

This is one of the many reasons Cameron and Jackson are pushing the 48fps and 60fps argument if I remember correctly.

[Jim Coulter]

Matt,

Thanks for the input. Is 100% in RCX a magnification or the actual full size image? If so maybe my explanation was off. The image was more pixelated than I would have thought.

[Peter Majtan]

Matt,

Thanks for the input. Is 100% in RCX a magnification or the actual full size image? If so maybe my explanation was off. The image was more pixelated than I would have thought.

It should not be pixelated at all. Make sure that when changing to 100% (1:1) you also change the de-bayer quality to full - they go hand in hand and affect each other...

[Peter Majtan]

Also - for all those that missed this - this is official word from Graeme:

Sharpening isn't necessary, but can be applied if necessary. Practically all cameras sharpen as a matter of course, but being RAW-centric, we leave sharpening to the user as a choice. One of the sound reasons for that is that sharpening is best applied with the destination format in mind, so if you're wanting to go out to a cinema screen and a web video you could well want, for optimum results, different sharpening parameters for each.

What we all want to avoid is that "bad DVD release look" where you see layers of concatenated sharpening ringing all over the image drawing thick childs' crayon lines in black and white around objects in the scene.

Given many people take the 4k or 5k and downsample, it is often the case that the downsampling filter will be sufficiently sharp to render post-sharpening unnecessary. Personally, I'd only look at introducing sharpening if I'm going for creative effect or displaying the finished movie at 1:1 from it's source resolution.

Graeme

That should clear this issue...

Peter

[Matt Ryan]

It should not be pixelated at all. Make sure that when changing to 100% (1:1) you also change the de-bayer quality to full - they go hand in hand and affect each other...

+1, you beat me to it

[Jim Coulter]

Yes the resolution was definitely Full. I would really appreciate it if I could look at an R3D sample from a user. That way I could bring it into RCX to make a comparison with a Full setting and punching in to 100%.

[Matt Ryan]

Yes the resolution was definitely Full. I would really appreciate it if I could look at an R3D sample from a user. That way I could bring it into RCX to make a comparison with a Full setting and punching in to 100%.

Here you go:
http://matthayslett.com/wp-content/u...2/MattsR3D.zip

[Jim Coulter]

Thanks very much Matt! This will help me a lot.

[Matt Ryan]

no problem! I can get you some more R3D's if you need them, just let me know.

Keep in mind that shot I gave you was shot using tungsten lights which mean theoretically it should be noisier than daylight.