F5: Wanted Beautiful cinematic F5 4k footage.

Not enough F5/55s in the wild yet to really even start to comment on how it looks. I didn't see anything in Dig or Mahout that was earthshaking (if anything it made me think harder about a used F3) but I have to admit the form-factor/ergonomics of the F5/55 appeal to me. Maybe in 6 months, after the RED case has been dismissed and people have shot miles of stuff on both cameras there will be constructive dialogue.

*I take it all back. I finally got the time to download the 1080p version of Dig and watch it on the 46. I don't know what anyone's definition of beautiful and cinematic is but I think the F55 has an incredible look, my word for it would be "breathtaking". The RED crew have good reason for concern when people have the choice to rent epic or f55.
 
Last edited:
Sharp is never a problem, especially if you shoot out of focus a lot. (meant to be funny joke)

The "video look" as many of us thought for years was due to the lack of sensor dynamic range, but the more I learned to control my lighting on-set and work within limited contrast ratios it looked "better" but still didn't seem to shake that "video look".

I think the real cause behind the "video look" is just plain old poor highlight management. Sure higher and higher sensor dynamic range is going to help capture a little more information in your bright tones - but there is something else internally being processed in Alexa and maybe even the F35 that no one is talking about.

Watch this video comparing EPIC, F55, and Alexa with some shots +4 stops over exposed. The Alexa and F55 do an amazing job keeping those highlights under control, far better than the EPIC in this test:

raw-battle_zpseebb7de0.png

Very interesting, if the Sony has 14 stops compared to Alexa's 14.5 (or thereabouts), then it must go a couple of stops deeper into the shadows, because the Alexa looks to hold a good 1.5-2 stops over the new Sony in the highlights.
 
Very interesting, if the Sony has 14 stops compared to Alexa's 14.5 (or thereabouts), then it must go a couple of stops deeper into the shadows, because the Alexa looks to hold a good 1.5-2 stops over the new Sony in the highlights.

I thought in this set of tests most of the cameras would have looked better with less exposure. On the F3 s-log dropped middle grey from 68 to 38. A lot of people over exposed at first as they didn't know how much of a drop there would be in middle grey. (How the heck did the red show deep blacks and the others didn't? Red raw should be all grey.)

I'd like to see Mitch or Andy give us the machine test of DR with middle grey marked like they've done with the F3 and other camers.
 
I thought in this set of tests most of the cameras would have looked better with less exposure. On the F3 s-log dropped middle grey from 68 to 38. A lot of people over exposed at first as they didn't know how much of a drop there would be in middle grey.

For people not being careful about it, the temptation is to VERY much overexpose Slog, especially version 2.

With S-Log2, middle gray is supposed to be at 30%. No joke. You apply the LUT in post and it brings it up to 50.

I personally see it all the time, because in many, many situatiotns (flat cloudy day, interiors, especially lit interiors) if you put mids at 30, you will have nothing in frame that even comes close to 100, so the temptation is to expose to the right and open up. That's why shooting with a LUT helps, it makes your old habits work.
 
Is the F55 between RED and Alexa?

Is the F55 between RED and Alexa?

I believe the "Alexa look" goes far deeper than the number of stops a camera has. Take for example D21 with such a limited dynamic range of 10 stops, from the posted video test it is quite clear and obvious to me that Alexa is does additional internal processing to the highlight information data coming off the sensor before it is recorded and encoded into Log format.

Think of it as an audio compressor/limiter just like what's used in sound recording studios, it maximizes the signal very close to 0db gain but compresses it enough so it never exceeds or clips. In this case of the video signal it is applied to the highlights of the signal, like a self-adjusting knee on the signal. In the video the Alexa whites never clip and always remain smooth and appear to "transition" from extreme whiteness.

RED doesn't seem to do any internal highlight management like Alexa but its also not a log format, it is linear gamma so in post you can actually pull down all the brightness and recover lost detail as well as smoothen out your highlights.

This test is excellent at showing how highlight clipping is handled from an in-camera processing point of view, but misrepresentative comparing log gamma to linear gamma and how much more flexibility you would get from RED (linear gamma) in post.

What's interesting however is the F55 highlight clipping test! It seems to almost combine a bit of both elements from Alexa and RED from a technical standpoint?!? ie. in-camera highlight processing, yet it is linear gamma encoded so you will still benefit from extensive highlight latitude in post.

Sony, I'm in.
 
I believe the "Alexa look" goes far deeper than the number of stops a camera has. Take for example D21 with such a limited dynamic range of 10 stops, from the posted video test it is quite clear and obvious to me that Alexa is does additional internal processing to the highlight information data coming off the sensor before it is recorded and encoded into Log format.

Think of it as an audio compressor/limiter just like what's used in sound recording studios, it maximizes the signal very close to 0db gain but compresses it enough so it never exceeds or clips. In this case of the video signal it is applied to the highlights of the signal, like a self-adjusting knee on the signal. In the video the Alexa whites never clip and always remain smooth and appear to "transition" from extreme whiteness.

RED doesn't seem to do any internal highlight management like Alexa but its also not a log format, it is linear gamma so in post you can actually pull down all the brightness and recover lost detail as well as smoothen out your highlights.

This test is excellent at showing how highlight clipping is handled from an in-camera processing point of view, but misrepresentative comparing log gamma to linear gamma and how much more flexibility you would get from RED (linear gamma) in post.

What's interesting however is the F55 highlight clipping test! It seems to almost combine a bit of both elements from Alexa and RED from a technical standpoint?!? ie. in-camera highlight processing, yet it is linear gamma encoded so you will still benefit from extensive highlight latitude in post.

Sony, I'm in.

What's the D21 has got to do with it? Was it in that test? If it was I missed.
The D21 is an older camera and it has nothing to do with an Alexa. Or are you saying because the D21 is also an Arri and only has 10 stops of DR, the Alexa has to be doing some trick and can't have real 14 stops of DR?
 
Back
Top