Magic Lantern enables 14-bit RAW CinemaDNG on 5D

What about SDXC cards?

These are rated at 90MB/s.

Price seems better than CF cards and I have a feeling the speeds on the SDXC cards seem to be increasing quickder and with larger capacity with the new UHS-II rating
 
Last edited:
Kholi said:
Would it be fair to say that people are underexposing the raw footage right now? That's been my hunch for a while now.
Click to expand...

Yeah probably, they're still thinking like they're shooting H.264. Took me a while to make the adjustment, I only just figured out the clean ISOs are dirty now.
:badputer:
 
I remember you talking before about 1250 being the cleanest in H264 in your previous tests. What do you think explains the difference. And I'll dig around a little with some of the creators of vids where we seem to be seeming a lot of the vert FPN and see if they know what ISOs they were using.

FYI, if you're convinced these numbers are correct, I'd make this the start of a thread. Very useful, if confer intuitive info!

Edit: and helps explain why this guy's 3200 looks so good (though this is on the Mk2): https://vimeo.com/66722244
 
squig said:
Yeah probably, they're still thinking like they're shooting H.264. Took me a while to make the adjustment, I only just figured out the clean ISOs are dirty now.
:badputer:
Click to expand...

Well, permit me to throw something out there: I'm willing to bet that the exposure tools are mostly useless now, aside from maybe the spot meter if it actually gives you a sensor read. Waveforms wouldn't help at all, and would just force you to undermine the sensor.

Also, it may be a case of people assuming that say 12 stops automatically means nothing should be blown out. Combine that an improper set of exposure tools for the format and you're going to see a ton of FPN and worse, unrecoverable color or odd colors. I'm only slightly confused by my own guess because, if it IS linear data, then color shouldn't suffer so drastically...

Before I forget, are you using IR NDs?

If people are going to use this seriously, you should likely go out and do brackets for each ISO in a controlled environment. A setup that you can over and underexpose at will. A sort of hack-kneed high-noon test could work if time's taken to make notes.

Oh, last thing, I may be way off about all of this but I do want to see someone figure out exposure. People were complaining about Macgregor's test and blown out windows, but his was the only one that I've seen so far what wasn't severely underexposed to my eyes. Could have been the grade.

Edit here: I haven't seen ALL of the footage out there, just commenting on what I have seen so far. Do not pitchfork me. xD
 
Last edited:
I've just got a schneider 4x4 and a hoya screw on ND, I think I've got a screw on IR somewhere. ML have developed a raw histogram. My thinking so far is it exposes a lot more like the 1DC (see Shane Hurlbuts tests). I'll give it a good daylight workout on the weekend.

The other thing is Kholi you don't know what kind of torture people are subjecting the DNGs to. I'm still trying get a good workflow going, by the time I get it right ML with have the CDNG conversion sorted. :cheesy: Prores422HQ doesn't look great with any underexposure. 4444 not too bad. And then there's the final H.264 transcode which always messes things up.
 
Last edited:
Great tests Squig!

I think this confirms what I've always thought: the native ISOs are 100, 200, 400, etc, and the 160x ones are achieved with a digital pull from 200x.

The vertical FPN, to me, is still a mystery. Some shots are useless because of it, some shots don't have it at all. Given how much variation there is between workflows, it may still take us some time to find the exact cause. It would be good if two people -one with horrible vertical FPN in his camera, and one getting clean images- could team up, harmonize workflows, and solve the mystery.
 
I think it's just tuning of the sensor, some may not be tuned as well as others, and underexposure of-course. Some people aren't doing a good job at denoising and grading either. A lot of people didn't even know about the FPN till the raw dev.
 
from limited experience I find AE workflow, the result is a bit statery, do you check Frame Blend in AE6 in composition settings?

On the other hand I find Quick time 7 Pro, changes the gamma, making it lighter for some reason..Overall I like so far Tiff, Quicktime path, it' s a bit quicker too.
 
Lots more great stuff being worked on at ML right now, all very early stage developments but 10/12 bit raw being explored, CinemaDNG, and external SSD recording. Like always nobody knows if these things will work but that doesn't stop them trying.
 
You must log in or register to reply here.
Back
Top