Messages

Most color graders will tell you to apply the LUT first, then adjust levels, and I agree with this approach, but either is acceptable if you end up with footage you want/like.

In Davinci Resolve you MUST change your project settings to match the footage at the beginning of a new project BEFORE any footage is imported into the project, selecting 'change project settings to footage' (or whatever that pop up says) with the pop up window after adding footage does not do the trick, you'd think it would like AE or PP.

So, if that's what your doing, then it'll never be correct when you try to render in the delivery tab until you create a new project, update the project settings to 2.35:1 (in 3 places on the pop up project settings form to be safe, I think only 2 really matter for rendering) and then add your footage to the project.

Not necessarily, but recommended as the nightlies are very stable (certainly not completely or as official release or alpha builds) and have the latest improvements

But! Older nightly builds actually have greater horizontal resolution, up to 1888 I believe (1880 for sure) compared to latest of 1846 (I think, I'm on a 5D3 these days, so don't exactly remember, plus could be back to larger now for all I know).

So, yeah, latest nightly should serve you well as long as those few extra lost pixels don't bother you

@ghostlayer

Thank you!

My fault for not catching this sooner, when doing card speed test you need to be in MOVIE mode, PLAYBACK mode doesn't give valid benchmark test results for using ML raw recording, movie mode does

Secondly, the other factors in benchmarking write speeds for comparing card speeds (and in actual raw record speeds) is the frame rate setting in the Canon menu, 24/25/30 and if Global Draw is on/off. These settings make a real difference in actual raw recording performance and are reflected in the benchmark results.

So, not saying this to 'guilt' you into posting more/valid test results, just letting you know

The 5D2 will never shoot 1080p as its bottleneck is the camera's CF card write speed, but shoots very close to it and can be upresed to 1080p without concern

As for 5D3 vs 5D4, well I'm hearing the 5D4 is going to shoot 4K, so probably gonna cost like $4K too but hopefully will knock down the price of 5D2/5D3 a little

Get the RED Scarlet and rent it out then use the money to get more Canon lenses which will fit both the RED and 5D3 then you'll have everything and still be able to rent the RED

But also, what do you need any of these killer raw recording machines for, besides rental?

@ghostlayer

Would you mind also posting your card speed test results here: http://www.magiclantern.fm/forum/index.php?topic=12630.0

Looks good!

Yes, remove 's' from 'https'

No doubt the scenario your shooting will play a big factor in your decision. I like using viewfinder when shooting with shoulder rig and outdoors on locked down tripod shots. External monitor is good for things like steadicam and dynamic tripod shots.

KomputerBay is good IF you get a good copy, but some KB cards are just outright bad, like the KB  128GB 1000x, way slower than advertised. Also, watch out for fake cards, some guy in another thread got like 400x cards when he paid for 1000x. Transcend and Lexar are probably most reliable, but more expensive.

As far as stability goes, find a stable build and stick with it for paid work and play around with latest nightlies for less important shoots. But either way, before committing to pro/paid work with ML, just use as much as possible to learn limitations, quirks, etc.

Can you open the file in an audio editor and really see what's what? Like Adobe Audition or even some free one like this: http://audacity.sourceforge.net

If you want, post a link to one of the 'empty' wave files and I'll take a look at it.

You say batteries, are you using a 2-battery grip? If so, try just one battery without grip and with battery door back on (then another one in case chip in one is bad or something).

Well, for quality yes, external mic is much better, but as far as diagnosing your problem it was just info needed to help figure things out.

Maybe the gain was too low for the audio source you were trying to capture. Can you get your hands on external mic and try? Either way, built in mic is really only good for syncing with audio from external mic and recording verbal footage notes.

Check it out

See here: http://photo.stackexchange.com/questions/27923/distortion-quality-when-using-a-full-frame-lens-on-a-crop-sensor-camera

Basically, it'll be less in the center of the lens and it can be removed in post. And 24mm distortion isn't as bad as say 11mm or 17mm

You can use full-frame EF lenses on crop sensor cameras, but cannot use EF-S lenses on full-frame cameras.

I would suggest favoring getting EF mount lenses for a couple of reasons. One, they'll work on crop and full frame sensor cameras, so if you upgrade to say 5D/6D you won't have to buy new lenses. Two, crop sensor cameras use only the center of full-frame EF mount lenses, this is the sharpest part of the lens and you get very little (if any) light fall-off/vignetting, see here: http://www.the-digital-picture.com/Canon-Lenses/Canon-Lens-Vignetting.aspx

Caveats are the focal length changes using full frame lenses on crop sensor camera, must use multiplier (1.6 usually, some cameras 1.3) to determine new effective focal length. For example, a 35mm EF lens on 1.6x crop sensor camera would effectively be 56mm. Also, EF mount lenses are usually bigger/heavier (for primes, not really an issue, for zooms it can be) and more expensive.

It's a bit disappointing to hear the max SD sped is 21 Mbyte/s Any links to info about this?

-> http://www.magiclantern.fm/forum/index.php?topic=6215.0

What resolution and frame rate are you trying to record for each camera?

It's been a while for me with that camera, but for sure I either used bitmap crops marks or the framing provided by mlv record.

I do use the info bars, top mostly for sound level, bottom mostly for exposure values.

No, but just about everything else is! 

Only using GD for framing, all other niceties turned off, but MLV info is on

No sound though

I would suggest something like Canon does, namely display an indicator only if buffering is problematic. You may have never seen it because you should not (unless you have a slow card). Splattering text would be good enough I think, because it is abnormal operation and should be noticeable and explicit.

"Slow card: maximum speed reached"      (say, 0.5% idle time)
"Slow card: <n> frames dropped"

I have seen the Canon buffering, when accidentally shooting ALL-I to the slow SD card I have ML installed on (forgetting to put fast CF in), if I remember it's vertical and on the right side of the LCD screen.

That's not a bad idea, just use colored status if all is mostly well, then additional info./warning as it becomes a concern.

@dmilligan
I personally have shot raw on a lesser camera, the 650D, and that's actually where I got used to relying upon the bit-rate and idle time, because you're really threading a needle when trying to record continuous raw 720p@24fps on that camera.

I like the comfort level bit-rate and idle time give me while I'm shooting, but as long as status is green, everything is Ok, so like you say, less important than remaining time/space (turn on for debuging would be good though).

On PC during card format (right click on it to bring up options). Mac IDK

Single module with MLV sound switch in MLV record menu sounds like good idea, take care of mlv recording all in one place and default to 'on'

I would prefer that it was set to 'on' by default also, but I think for unified versions it's set to 'off' due to the performance hit recording raw takes with audio enabled.

FYI, I don't think the forum is sending emails anymore

Also, are you sure you're right clicking on the imported footage in Adobe and changing the main footage fps to 23.976? Not just in the composition settings.