Here's a weekend project I put together using ML RAW on a 5D3 -> RawMagic -> Resolve -> FCPx
I was also curious to try out a template from http://www.motionvfx.com/
https://www.youtube.com/watch?list=UUJyhFSRpgV8lgL1EuVNDlng&feature=player_detailpage&v=AjL144GI0dA
"There needs to be a rule about posting test vids."
There needs to be NO rules. Wrong choice of words. Rules suck :)
Blackmagic made some changes in DaVinci Resolve to make it work with the 14bit files produced by raw2dng . You need new v9.1.5 for this.
http://www.blackmagicdesign.com/support/detail?sid=3948&pid=11735&leg=false&os=win
Here is the trailer for my short movie "Elementarschaden".
It was shot entirely on 5DII RAW with Zeiss Primes.
I loved shooting and editing this movie. feel free to head over to my thread at
for some feedback. :)
The power to control the highlights and shadows is amazing. Great skin tone color as well
Important - Huge leap in DR on Canon 5DMKIII/7D with ML RAW
A1ex over at ML has performed a trick with the Canon 5DMKIII & 7D that samples half of the sensor at ISO 100 and the other half at ISO 1600 to provide greater dynamic range - basically alternating ISO during sensor readout. For example. If you mix these two, you can get almost the entire dynamic range the sensor is capable of (around 14 stops).
There are no motion artifacts; not even a difference in motion blur for the two exposures.
So far in my own tests on the 5DmkIII its pretty radical.
It is really strange claim, may be they just found some raw preprocessing. Alternating ISO (and it must change for each line) must provide little value.
As a1ex mentions, "This technique is very similar to the “zero noise” “technique of the 4 f-stops“ developed by Guillermo Luijk (http://www.guillermoluijk.com/article/nonoise/index_en.htm), where he uses only two bracketed images, at 0 and +4 EV, to achieve great interior shots without the ”radioactive“ HDR look. Emil Martinec gives a hint on how to shoot at ISO 100 and 1600 simultaneously, but this requires two separate amplifiers fed from the same sensor data."
A1ex gets around it with a simple software trick and a little math. "The ISO (analog amplification) is alternated for every two scanlines, between two user-defined values (say ISO 100/1600), so half of the image is exposed for highlights and the other half is exposed for shadows." He's been developing this code for the past few weeks hence little work anywhere else on ML.
I'm editing some test videos at the moment for upload later. One point to note is that there is a little loss in vertical rez.
a1ex wrote up a PDF describing the approach in-depth.
every time some news like this about the MK3 pops up it makes hard not to want one.
There are no motion artifacts; not even a difference in motion blur for the two exposures.
It's one exposure, not two.
Incidentally, this is akin to what the BMCC and BMPCC sensors do, except they do it for every single photo site, with a pair of amplifiers and 11-bit ADCs on each with two different gain settings.
In fact they claim that Canon sensor (and ADC) is special. It is also seems quite clear that work had been done with some Canon help and under order of some big client.
It looks like you're new here. If you want to get involved, click one of these buttons!