Uncompressed 7D footage LOOKS UNBELIEVABLE, so HOW DID THEY DO IT?

So what do you guys think, will I probably have to give more headroom when blocking the shots to take into account for the RED record sign, for later cropped.?

Or

I could just use the live wiew to capture the footage into final cut, that way I won't have to crop it al all, and actually get the entire view. Not to mention as a side Plus to that type of workflow, i wouldn't be limited by the 7D's 12min record time limit. HMMMMMMM....
 
Mention of the MX02 mini not doing 24fps and losing audio sync is bothersome. But these complaints could be coming from improper use or setup. I'd sure like to know that the mini works well. It's only 6x4x1.5"...very portable.

WOOOOOOOOAH, hold the brakes. I didn't catch that part, I'm diggin arounfd now to see if A) this is true, and B) if there is a solution, or another model to purchase which allows 24p capture.

To be continued.....:cry:
 
Some things to keep in mind:

1. Don't expect to be able to capture uncompressed to your laptop. Recording 1920x1080 4:2:2 24p (8 or 10 bit) requires >=5 drives in your RAID.
2. Don't expect to be able to capture to ProRes 1080p with your MXO2 Mini and MacBook Pro laptop. You need QuadCore.
3. You can capture to DVCProHD on-the-fly with MBP. Also, you should be able to capture to 720 ProRes. Not sure 7d can output 720, but MXO2 can downscale the input signal in hardware. Sure, you loose resolution and there might be no point in it. Also, you might try 1080 ProRes LT. On the PC you can capture to extremely high bitrate MPEG2.
4. MXO2 Mini not supporting 24p shouldn't be a problem at all. As far as I know 7d outputs 1080i so it must add 3:2 pulldown (my guess). You can remove the pulldown in Cinema Tools (if using Final Cut Studio).
5. There is a problem with the latest MXO2 Mini OSX driver. It causes kernel panic from time to time, so the solution is not very robust. Matrox is aware of this problem, so we can expect this issue to be addressed in the future updates.
 
If you read what the Swedish guys are doing on their blog, the signal coming out the 7D is 60i no matter what the recording frame rate, and the frames are randomly split between frames and fields. They had to write a script that compares frames to pull the 60i apart into complete frames, like sometimes half the frame is on one set of lines in one frame, and the other half of the frame is in a set of lines in the next frame. So you won't be able to do much with the signal even if you can get it recorded onto your computer unless you go through the rigmarol that they have. After all that the aliasing and moire is still there (you can see it in the full size hockey player jpeg).
 
do you by any chance have a Maxtor mini?

And if so what about 24fps. I'm not exactly familiar with the whole pulldown shuffle, how does it work, related to capturing uncompressed video from the Maxtor mini from 7D.
 
If you read what the Swedish guys are doing on their blog, the signal coming out the 7D is 60i no matter what the recording frame rate, and the frames are randomly split between frames and fields. They had to write a script that compares frames to pull the 60i apart into complete frames, like sometimes half the frame is on one set of lines in one frame, and the other half of the frame is in a set of lines in the next frame. So you won't be able to do much with the signal even if you can get it recorded onto your computer unless you go through the rigmarol that they have. After all that the aliasing and moire is still there (you can see it in the full size hockey player jpeg).

I actually did see that, and actually emailed them as well as posted on their comment section asking for a download to that script. I'm not exactly sure how that would work, not can I write scripts if I did know, but if they've done it, and are willing to share that info I'm willing to test it out.

Anyone have any idea how we'd begin creating this 60i frame rate algorithm?
:shocked:
 
the signal coming out the 7D is 60i no matter what the recording frame rate, and the frames are randomly split between frames and fields.

The frames are not split RANDOMLY between fields. There is a 3:2 pulldown scheme.


They had to write a script that compares frames to pull the 60i apart into complete frames, like sometimes half the frame is on one set of lines in one frame, and the other half of the frame is in a set of lines in the next frame. So you won't be able to do much with the signal even if you can get it recorded onto your computer unless you go through the rigmarol that they have.

You're wrong. You can remove the pulldown easily with Cinema Tools on Mac. I've done this many times when capturing my HMC151 HDMI output (which is 60i as well). You just have to specify manually which frame(s) do the first two fields of the footage correspond to. The rest of the process is automatic. You can either create the new file, getting rid of redundant fields to save disk space, or change the metadata, which is done in split second.

After all that the aliasing and moire is still there (you can see it in the full size hockey player jpeg).

True. And the rolling shutter :(
 
I'm only quoting what I read on their blog. THEY said it was random. Did they write their script for the fun of it??

I can't find the blog post where they talk about the script, I assume they have taken it down.
 
Last edited:
I'm only quoting what I read on their blog. THEY said it was random. Did they write their script for the fun of it??

Sure. You're right - they said that. That was unfortunate expression.

The frames are not put into the 60i stream randomly. They meant something else:

4 24p frames (ABCD) are put into 5 consecutive frames (10 fields) of the 60i stream as follows:
AA AB BC CC DD
To be able to extract the 24p (remove pulldown) you need to somehow determine whether the first frame of 60i stream is AA or AB or BC and so on (this is random). To do that you either have to tell the software (it's very easy to tell by watching the first 5 frames of the 60i footage) or let the script do it by comparing fields.

All I wanted to say is that you don't need a script to be able to remove the pulldown. You can do it by eye easily. But you have to do it for every single clip you record.
 
Well you seem to know more about what they meant than they do. Found it now, here's what they said, note the words THERE IS NO REPEATING PATTERN:

We wish the HDMI stream was as the framerate recorded. But according to what we recive it's not the case. We get the recorded framerate packed into a 60i container from the camera, then we unstuff the correct source out of the 60i stream. A quite odd process since there is no repeating pattern, the solution is a script that analyze each frame and extract what is needed to recreate the clean source out of duplicated frames and frame material spread out over field 2 on one frame and field 1 on the following and so on. If somebody have a solution to make the camera output HDMI at recorded frame rate that would be a huge helper. It might be somthing that the magic lantern people can sort out. But so far we have this work around solution with 60i coming from the camera. We do not loose any data but we need to render our clips trough our script before we can use it.
 
Well you seem to know more about what they meant than they do. Found it now, here's what they said, note the words THERE IS NO REPEATING PATTERN:

We wish the HDMI stream was as the framerate recorded. But according to what we recive it's not the case. We get the recorded framerate packed into a 60i container from the camera, then we unstuff the correct source out of the 60i stream. A quite odd process since there is no repeating pattern, the solution is a script that analyze each frame and extract what is needed to recreate the clean source out of duplicated frames and frame material spread out over field 2 on one frame and field 1 on the following and so on. If somebody have a solution to make the camera output HDMI at recorded frame rate that would be a huge helper. It might be somthing that the magic lantern people can sort out. But so far we have this work around solution with 60i coming from the camera. We do not loose any data but we need to render our clips trough our script before we can use it.

If this is the case, how will you explain the fact that HDMI monitor, connected to 7d's output displays the (moving) image just fine? You don't experience the drops, jumps etc. Just the usual 3:2 judder. If the frames were laid out in the 60i stream 'randomly' you would experience such artifacts. Does the monitor have built-in magic script?
So, I still claim I know what they mean ;) They said 'there's no repeating pattern'. And they're right - there's no repeating pattern from clip to clip. You need to determine the 'starting point'. But I'm sure there is repeating pattern every 5 frames of 60i footage. The only reason for adding pulldown and outputting 60i to make its HDMI compliant with more devices (plasmas etc - all but latest ones cannot handle 24p stream over HDMI). Randomizing the fields and making 7d not compliant with anything in the world would be simply silly. By the way - people have been successfully doing inverse telecine (pulldown removal) with 5d HDMI output without any 'magic script'. Don't expect 7d to be different.
 
Last edited:
Trez is right....all they are talking about is 3:2 pulldown. It's only the way they expressed it makes it seem like something out of the ordinary. It's not. Cineform, SOny Vegas, most NLE's should be able to fix that.
 
Fair enough.... from reading their blog they sound like they know enough about what they're doing to just say it's 3:2 pulldown.... but they didn't. I was just passing on the info.
 
Last edited:
...and a small thing, but Syndicate is a Swedish company, we are not Germans, have never been and hopefully will never be....;)
 
Well I was impressed so much by the SWEDISH clip of the racecar, that I hooked up my friend's Canon 7d to the Nanoflash and hit record. Well let's just saw that the size of the final image, once I cropped out that red record button came in 810 x 440. Which is right around an SD frame size. So I don't quite know how these guys did it. Plus it was interlace and full of drop outs.
 
Then you are doing something wrong. It should be around 1600X900. It is when you use the Black Magic Intensity card to capture.
 
Well I was impressed so much by the SWEDISH clip of the racecar, that I hooked up my friend's Canon 7d to the Nanoflash and hit record. Well let's just saw that the size of the final image, once I cropped out that red record button came in 810 x 440. Which is right around an SD frame size. So I don't quite know how these guys did it. Plus it was interlace and full of drop outs.


You need to hit the info button a couple of times. Then it will net you the higher resolution.
 
Back
Top