Messages

Thanks 
Tried to watch you video but, said its Private 

Sorry About that, I Fix it

On the 5D2 the Max-with from the sensor is 1880
Canon upscale the image to 1920 for h264 with software.
It really depends on what your Project calls for.
Being that the image is very scale able meaning , re-framing, punching-in.
I like to use Max amount of the sensor then crop out sides .
the other reason i like it is because of the reduce height for a more Film look & feel !

It also easier to Write to the card @ 68.5 MB/s
That's at lest in my case, I think if with can keep frame data close to the best Hexadecimal 
it should write quicker, but i'm Not a programmer so that a guess.

But with 1856x1003 it very close to HD @ 1:85.1 AR
that would be the ticket, there again re-framing etc...
I will settle for 1856x1003 for  HD Work & for Documentary
I would use 1880x854 or even 1880x940 if you can do it.

Stop thinking about line skipping and all the stuff, you will always have problems
with that line skipping for panning etc..
But think of you camera as a "Film camera" to capture Positive Film Stock (but More)
It's Not a Video Camera, because with are capturing Frames like "true Film" Stock
I personally turn off everything on the camera & use manual focus,
use proper exposing Technics, "Use Meter on camera" i don't use the "ETTR" function.
but i have a Variable ND and  Stop it up or down for scene.         

I totally agree with your assessment, I film some quick shots of a creek behind my home area

I love the raw2dng.app .RAW converter, I'm wondering how accurate the white balance auto, can you give an option like we have with ProRes 4444 and Destination folder to turn off.

Ok,
my camera never when off "continuous ok" it always said "continuous ok"
But i understand what you are saying, 
i have always had better numbers then what i have read in this forum.
i will try the 1728x972 AR and see if this problems is still there.

But in the Short term i can record 1856x1003 for 14,000 frame ! that's 9.7 Mins 

Update:
I just check the DNG's from the 64 GB full card record @ 1880x854 24p, 22,571 frames
Frame over ride on, small hack on , Photo small jpeg , ETTR = Off,
and i had the last 3 Images are corrupt, pink garbage frames
So i call that success ! with 22,568 Good frames 15.67 minutes on lexar 1000x 64GB card
If only i can get 1856x1003 to do 64 GBs (max is 42GB)

Do I have this process theory right or wrong, Because of the line skipping process from the DiGiC chip we have been shorted 266 continuous pixel for every frame 40 from the 1920 width and 226 from the height. I'm sure I might have something wrong, I was just wondering what was your opinion on 1880x854 for the Mark II vs 1920x1080 for the Mark II?

I kind of thought that would be the answer, so that's it. I think once ML run thorough test to every function we might have a good understanding to this Miracle Firmware Called Magic Lantern RAW. I would like to thank the Magic Lantern Team for this Cinematic Break Through, I'm sure you might have more in store with your BETA and I'm on pins and needles.

I don't know if this is the same question from gnarr but In the same fashion of the native H.264 from canon, is there anyway way ML firmware could allow ProRes 4444 or ProRes 422?

In the same fashion of the native H.264 from canon, is there anyway way ML firmware could allow ProRes 4444 or ProRes 422?

By the way when I say shooting h264 is easy I just mean you deal with the clipping coz you have no choice so you just accept it

Have you try filming something with the Digital Dolly? It's pretty awesome but I was wondering what would people use it for, would the use it I place of tilting the tripod?

Okay, so I know a lot of people are new to shooting raw footage now that we have been gifted with the power by everyone's favourite Canon DSLR "hack" so I thought it might be a good time to start a discussion about actually shooting with it 

There is a great thread going regarding post processing raw but what about the shoot itself. What are the best practices  when shooting raw and what tools do people use for getting the best exposure? Shooting h.264 is easy. You expose the subject and light to match (if you are lucky enough to be able to control your lighting) but what about if you want to gain maximum dynamic range from a dynamic scene?

I personally am a photographer who (of course) always shoots raw but even after years of experience doing this I still find myself exposing my raw video incorrectly.

Also, it would be really nice to start seeing some real world examples of exposed scenes instead of pushing the raw data to the extremes just for the sake of showing that it CAN do it.

Please feel free to share your techniques and experience with the rest of us

*ps. If this is already being discussed in another thread, feel free to move me along

I've been thinking about filming 2:1 format but is 2.35:1 or the better choice for Film presentation.

where did the audio recording go off to?
also i really wish you guys could bring back the resolution selection of when this first came out. I was shooting anamorphic and a resolution between 4:3 and 1:1 would be great 4:3 gives me 2.66:1 and 11 gives me 16:9 on the old version you didnt have to select an aspect ratio, you could just select width and height. Great work though, cant wait to see how far this will go! Been using magic lantern ever since i bought my first canon (t2i) back in 2010.

ML has change the Game, where from professional videographer to Cinematographers if you have a limited budget but good technique you can achieve very film industry professional end results. You were saying you could get anamorphic 2.66:1 out of 4:3 and anamorphic 16:9 out of 1:1 frame recording, but aren't you cropping out image anyway as ML does now in format selection? I do wish they could put description icon or writing to explain different in aspect ratio for example 2.35:1 or 2.39.1 Cinescope setting and 16:9 or 1.85:1 HDTV setting ect. Please correct me where I maybe make an error in characterization recent anamorphic 2.35:1 or 2.39.1 may be called super 35.

At first few frames the EXPECTED frames would say CONTINUOUS OK but after several thousand frames it would change to a much lower number. You have to see if,

1. It would stop when card capacity is reached.
or
2. It would change and stop before reaching card capacity.

I was misled during a couple of my tests when I immediately saw, CONTINUOUS OK then i stopped recording, when on my subsequent tests it would significantly lower.

I have ran 1872 x 1012 several times and it never would fill up my card.

This is with GD=ON, preview=Auto, since in production scenario you need them to be ON.

GD=off, preview=hacked makes no difference in 27 and 29 build.

Hello Ted,
Hope everything is good for you, I was ordering the method behind Card Warm Up function and what's your theory of when and why us the 3x crop when for closer subject its a distance challenge. But I love it no matter what, I will use it for far wide shots and maybe micro subjects.

You want the latest build, your build doesn't have Small Hacks on the build, this helps alot, and also the new Card Warm up seems to be nice, i found writing 128mb to the card on boot up to be best so far.

What do you mean the New Card Warm Up?

At first few frames the EXPECTED frames would say CONTINUOUS OK but after several thousand frames it would change to a much lower number. You have to see if,

1. It would stop when card capacity is reached.
or
2. It would change and stop before reaching card capacity.

I was misled during a couple of my tests when I immediately saw, CONTINUOUS OK then i stopped recording, when on my subsequent tests it would significantly lower.

I have ran 1872 x 1012 several times and it never would fill up my card.

This is with GD=ON, preview=Auto, since in production scenario you need them to be ON.

GD=off, preview=hacked makes no difference in 27 and 29 build.

I stand corrected, thanks

Can you confirm that with your card? With a 3sgig 1000x At that resolution with that build its not continuous. I only averaged between 2400 - 4600 frames, Same with jun 29 build.

For now its 1872 x 936, 1872 x 850, 1856 x 928, 1728 x 972, 1472 x 1104 and lower that can give you continuous recording.

I was able to get 1872x1012 continuous on my Lexar 32GB 1000x with June 27 Build, I'm going to install the June 29th build right now.

I love the isolation of the frame ratio in Canon Preview option, what is this a disliked preview? Most poster favor ML Grayscale and HaCKeD which I feel all are evenly necessary but why is the HaCKeD option better for frame increase?

Better yet, they should just include it within the code and ON at default and should not even be an option to be ON or OFF.

Hope you guys are having a good weekend.
I need answers to some question for me, I know 1872 is the 1/3 line skipping of the Full Sensor Res so is that the Max setting for Raw on 5D Mark II? Also what Aspect Ratio do you use the most and can you give a quick personal reason for your choosing? And finally Ted when you do your very instinctive test I alway wonder are you using the Max setting?

Using? Like in real project? I would not recommend one at this point since development by the ML team is moving so fast.

For testing, I recommend you follow the current builds and draw your own conclusions based on the CF cards you are using.

I was referring to which build has the most stable frame increase, but I get your point all builds are from a growing improvement. Going to test out the June 27 build.

Thats for 5D mkii direct from A1ex himself

How are they receiving 1920x1088 or 1920x1080, are those reads from crop option?

im currently using 5d2-fac94f686d23-pink-debug which A1ex posted earlier, so far so good best one so far, not all features there but i added the ETTR module my self and seems a'ok, nice n stable.

These post you guys are referring to are for the 5D Mark III?

Comparing Vanilla (5d2-fac94f686d23-pink-debug ) and a.d. pink fix jun 25 (d07dab0ba6c2)

Ran tests 2 times consecutive for each resolution except in 1856 x 928, ran it 4 times



Resolution     --  A1ex             --  a.d. jun 25 pink fix
                      Vanilla jun 25
1X

1872 x 1054  --  390/528       --  395/501
1872 x 1012  --  840/709       --  882/822
1872 x 936    --  3667/2009   --  2389/3359
1872 x 850    --  continuous   --  continuous
1856 x 1044  --  510/436       --  612/573
1856 x 928    --  3989/5177   --  8675/2445
                            /3321/6452      /9101/9927

1728 x 972    --  continuous   --  continuous
1472 x 1104  --  continuous   --  continuous
   
3X Crop mode
   
2144 x 1076  --  188/235       --  234/237
2144 x 1072  --   221/218      --  211/241
2144 x 974    --   487/376      --  353/373
1920 x 1076  --   416/386      --  402/411
1856 x 1044  --   736/748      --  677/794
1728 x  972   --  continuous   --  continuous

Results are close except for that in 1856 x 928

Ted thanks for your extensive information, after your testing so far which build are you using?

Thanks guy for use my workflow  to upscale.
Tiff can carry any embedded profiles that you can embed , or use Camera raw.

The Biggest think you must do is maintain the dpi, (240) Very important ! this is
where the image makes it or brakes it, I have been Working with PS & AE  If you try to upscale in AE
it will look Bad, because AE dose not see the image as Photo image but as Frame image @ 72 dpi.

Plus i have play around with "Film Color Space" in Photoshop & After Effects.
i have been embedding the PanaLog  "Panavision Genesis Tungsten Log"
in the Tiff image , i think that would be great for 2K or 4K "B" roll

Panavision Genesis Tungsten Log (Embedded LUT)or ICC for Photoshop

YouTube is the only ones that stream True 2K & 4K in the original Quality setting tab.

Reddeercity thanks for getting the conversation started with the upscale possibility of Magic Lantern gift of RAW video, but to think of the 5D Mark II ML RAW firmware ability as only use for B-Roll kind of sucks. So let's begin with why not use Magic Lantern's gift of RAW as main film production stock? Why not use the the Rawness DNG to upscale at least to 2K, the reports of the Data from the DNG holds up? Why not use the Canon 5D Mark II for motion picture projects, I ask these questions keeping in mind that the fast race car scene in the beginning of Ironman 2 was shot with the Canon 5D Mark II with no fear of Rolling Shutters, the full 6th season finale of House was shot Exclusively with the 5D Mark II (it was a visually broad canvas presentation), the feature film ACT OF VALOR was shot exclusively with the 5D Mark II for 12 million dollars then grossing 81 million dollars and many more worthy Canon 5D Mark II professional projects. All those example of professional film work was accomplish with the Canon limitation firmware, now even if the limitation of Magic Lantern Firmware in RAW giving the Mark III 1920x1080 continuous RAW and the Mark II only shooting 1880x850 continuous RAW but Magic Lantern firmware setting option is very professional. Final question, Isn't the Magic Lantern 5D Mark II firmware settings and RAW capabilities industry standard?

dariSSight#1 as far as i can tell, the Tiff file contains the full color information and original resolution. The DNG is transfered to Tiff with 16Bit color depth and so the full 14Bit should transfer in the process. No Banding etc. The PS Image Processor gives you the option to process the DNG to different sizes: 1880x854 (1,6MP) 2048x930 (1,9MP+) or 3072x1395 (4,3MP +) with 240 PPI the resulting 3K Tiff files really show much more detail in comparison to the original DNGs. I don't know if this resolution is just achieved through upscaling the 1880x854 or if the 3K data is backed in in the DNG files, but it really looks nice. It is a shame youtube just compresses everything to 2K in 8Bit 420.

When it comes to image quality and pure resolution and detail the Red cameras are the industy standard, but the images coming out of the 5D2 and 3 are nothing short of amazing. The Red cameras are designed to deliver the best possible image in the 2K+ range and there is a noticeable difference when you compare the upscaled 3K with an 3K R3d file, but it is not as dominant as one would think. I would say you can intercut footage and it would not be a major problem

WoW, For you to state the intercut possibility the 5D Mark II threat is real.

dariSSight#1 as far as i can tell, the Tiff file contains the full color information and original resolution. The DNG is transfered to Tiff with 16Bit color depth and so the full 14Bit should transfer in the process. No Banding etc. The PS Image Processor gives you the option to process the DNG to different sizes: 1880x854 (1,6MP) 2048x930 (1,9MP+) or 3072x1395 (4,3MP +) with 240 PPI the resulting 3K Tiff files really show much more detail in comparison to the original DNGs. I don't know if this resolution is just achieved through upscaling the 1880x854 or if the 3K data is backed in in the DNG files, but it really looks nice. It is a shame youtube just compresses everything to 2K in 8Bit 420.

When it comes to image quality and pure resolution and detail the Red cameras are the industy standard, but the images coming out of the 5D2 and 3 are nothing short of amazing. The Red cameras are designed to deliver the best possible image in the 2K+ range and there is a noticeable difference when you compare the upscaled 3K with an 3K R3d file, but it is not as dominant as one would think. I would say you can intercut footage and it would not be a major problem

WoW, For you to state the intercut possibility the 5D Mark II threat is real.

I can't reproduce the pink frame issue at all.

Straight from my card: http://a1ex.magiclantern.fm/bleeding-edge/5D2/5d2-fac94f686d23-pink-debug.zip

Write down the black levels (they appear in the top area on the screen, when starting and when stopping recording) and post them, along with the result (whether the recorded frames were pink or not).

I will test out the new .a.d new June 25 282abfa87fef build when I get home, but Alex where do you post you builds I have your pinkdebug build?

Just a simple upscaling test with the Photoshop Image Processor. DNG (1880x854) to TIFF (3072x1307). No Post-Sharpening etc.
It is not exactly RED-3K, but you can work with it.

Thanks to Reddeercity for providing the infos on the workflow 

1880x854 32GB Transcend 1000x
06/13/2013 Build

NewHourFilm I've been looking at your video, how does the Tiff files hold up when you blow them up to test for pixel depth? Can you go into a little more about the RED-3K comparison, I'm always geeked to know opinions on how far my 5D Mark II can take me?