Messages

Install module by selecting in module menu then restarting camera.

Komputerbay 64GB 1066x CF on Canon 5D3.123 @ 24fps

Komputerbay 64GB 1066x CF on Canon 5D3.123 @ 24fps

For now, I'm guessing we'll need 1200x (180MBps theoretical read speed) card to eke out reliable 100MBps write???

Could offer common selectable values (like 1024/2048)/etc.) and allow for user input. This would provide for quick selection of known black levels and ability to input any value for unforeseen and creative circumstances

No doubt

Just so you guys know: the minimum screen resolution for DVR is 1680x1050, and that's written in the DaVinci config guide.

This resolution works too

Suggest search forum, many posts on this with solutions.

Sandisk Extreme 64gb 120mb/s card...

Is that an SD or CF card? If SD, the write speed of the SD card data bus in the 5D3 is only 20MBps, you need to get a fast CF card (1000x/1066x) to record continuous 1080p. The CF card data bus is 100MBps.

If it's a CF card, then what resolution and frame rate are you trying to shoot at?

See here for camera write speeds: http://www.magiclantern.fm/forum/index.php?topic=6215.0

See here for card write speeds: http://www.magiclantern.fm/forum/index.php?topic=12630.msg121741#msg121741

Don't wait for official stable release, else you may never install. The nighties are very stable, sometimes minor glitch, but fixed quickly.

No problem on complete ML uninstall including boot flag

Have you set the Macbook display settings to more space (1920x1200)? DVR doesn't support the standard retina resolution.

Works

Yep, thats what I meant, now that it is out of beta I had hoped it is fixed, but it is still doing this! Am I the only one who is annoyed by this?

Yeah, beta 1 was worse, it happened every time you changed major views, not just within the color view like beta 2. But @fisawa found good solution.

@RamsesIIV

What is the current format of the image you are trying to view in camera? JPEG, PNG, TIFF...?

Canon cameras output RAW (CR2) and JPEG image files, so if not currently JPEG, then try converting it to JPEG then load onto card and try to view. If that doesn't work, take the old image (and convert to JPEG if not already) and modify EXIF data to match that of a JPEG taken with your Canon camera.

Best I can think of.

Retouching images manually is the only thing I can suggest also. One approach is to count how many rows of pixels are corrupted then divide by 2 and copy that many lines of pixels from just above the corruption and paste to top part of pink stripe, then copy that many lines from just below the corruption and paste to bottom half of pink stipe, then blend using the healing tool or such.

Try this technique on a few frames and test how it looks, maybe at full speed playback it will look alright.

UPDATED: Also, if pink stripe frames are not consecutive, then try frame-blending the previous and subsequent frames to create a new frame and then replace corrupted frame.

Video:

"But the researchers came up with a second technique, using the "rolling shutter" typical on digital video cameras, to extract enough information from the video to still recreate high-frequency audio."

Find article here: http://www.theguardian.com/technology/2014/aug/05/mit-crisp-packet-eavesdrop-audio-sound

Yeah, I've resorted to taking phone captures too, just easier

So, at least you have some valid benchmarks to work with, be aware though it'll be a little slower with Global Draw on even of it's just for drawing crop marks. We benchmark with Global Draw off as a 'standard' to make it easier for comparison.

Can understand that.

If you are going to use ML again, I would suggest getting any old SD card and running ML from that. So much easier just to use freshly formatted CF card (no boot flag or ML) for your shoots without worrying about updating every CF card you get with boot flag and ML. Also, should do in-camera format on CF card before important shoots to help ensure good images, when ML is on CF card, you tend not to format it as often.

Was the CF card with the new nightly build bootable?

Either way, is there any specific compelling reason you are not using the SD card to boot with ML?

Unless you're using an eyeFi card or something, then I would suggest doing it this way, it's been a very stable approach for many 5D3 users.

Difficult and impractical, yes, impossible, no, worthwhile, maybe (depends on individual).

The initial compilation of current lens t-values look up table, though arduous, is achievable. The active adding of new lenses in a timely manner, improbable

My question is, what would you do with such a value on a through-the-lens metering device (camera)? Use it for reference/research? Or are you planning on using external light meter?

...hash my Bitcoins at the same time.

Now that's a feature I can get behind

I would say that the f number is an objective value without regard to the lens' light transmission efficiency, while T stop value is subject to the 'quality' of the lens.

Also, I have to disagree with you regarding film cameras not having 'through the lens metering', see here: http://en.m.wikipedia.org/wiki/Through-the-lens_metering

T-stop values are typically used when relying upon external light meters for either digital or film cameras. Also, T value for a lens informs you to the real amount of light passing through it, valuable for comparing lenses when low light conditions call for faster/fastest lens as two different lenses with the same maximum aperture may in reality pass through different amounts of light.

@DFM

Thank you for clarify that Osiris LUTs are film emulation, not simply color correction for normalizing footage. There are times I've converted footage to log space did some primary color correction and then applied a color grading LUT (like Osiris) and was happy with results and stopped there.

That being said, a common workflow is primary corrections (including normalizing LUT like VisionLog) > apply grading LUT (like Osiris) > secondary corrections (like grading with a mask).

Version 11 beta 2 fixes this (mostly), now it happens just on the color grading view when toggling the node view.

But yes, it's still annoying when trying to work quickly

What do you mean 'blank track'? And how are you trying to sync?

There's an auto exposure module which is pretty phenomenal that might do the trick, but I've never used it and maybe only works in photo mode.