14bit RAW DNG silent pics! (silent.mo)

[1%]

In CS6 I just import all the files and then create comp... maybe I should only click 1 as it dupes the frames so the sequence repeats, kinda weird but I haven't had to click a single raw window yet.

[PhotogMike]

I noticed a weird thing where as I'm in the ML menu the Date/Time from the Canon menu flashes on my screen. Running 4.29 Nightly Build.

http://hiddenmoonproductions.files.wordpress.com/2013/04/photo.png

Wondering if anyone else has noticed this.

[1%]

This was supposedly the bug that 6D firmware fixed... I'm guessing thats on 5d3 and still not fixed?

[PhotogMike]

Thanks. Sorry. Using a 5D II and I hadn't noticed it until testing the silent pic functions so I didn't know if it was new or not.

[woodybrando]

ok figured it out. Just click import in after effects, then highlight the sequence of dng's and pick import as "Footage" and click "Camera Raw Sequence" box. And it'll let you choose raw settings for the sequence. Also, I'm not seeing the corrupted data anymore that was on the bottom and left side. The whole image is clean now.
Also, getting 49 frames when I choose to shoot raw only and shoot 720p. But it only fills half the frame with a squished image.
This would be great for 5 second films.

Jayson

[1%]

Other possibility is that time dialog is being used to make the scroll wheel work on 5d2. Seriously need a better fix for this... flashing menu is hell annoying on 6D/EOSM

[Audionut]

Ufraw uses the black level computed by ML. If it's wrong, you can't correct it with white balance.

Exif info is not implemented, just some dummy values.

I assume Adobe products expect the information to be in the exif.  Hit auto WB and it works it out.  It's no big deal, anyone using DNG's should be PPing anyway.

[Yoshiyuki Blade]

Recorded about 10s of random footage. Applied WB, converted to 8-bit tiff in photoshop's DNG converter, (slightly) stretched and cropped for 1080p and encoded it losslessly. No denoising or sharpening applied.

The 8-bit 4:2:0 is 279 MB while 4:4:4 is 508 MB lol. As expected, the visual difference isn't huge unless I find some bright vivid reds. I'll edit this post later with a link of both clips for you guys to compare.
Here are the lossless encodes for comparison:

Edit: (Links removed due to excessive dropbox traffic causing suspension. )

Raw is definitely way overkill for any kind of practical recording, but will there be a possibility for a compromise? The improvement in resolution alone makes it worthwhile, but what happens when the image goes from the raw buffer to YUV 4:2:2 that kills the detail so much? You'll see that even bringing a raw image down to 4:2:0 won't affect the image quality that badly. Is there some kind of filtering applied to increase compressibility at the cost of detail?

Just got my 5d2 to get 51 frames when taking out the raw+jpeg still function.

Wow, same here. I had it on raw and was getting 36 frames, but now I get 51 too when I set it to JPEG. Doesn't seem to affect quality in any way. I wonder what happened?

[menoc]

Wow . . . wait a minute, Would I be too presumptuous to assume that if this works on the 5D2, this may very well work on the 50D!? Remember, the advantage of the 50D is that it takes CF cards - including the 1000x and has less overhead without audio . . .

Please say it will, Please say it will, Please say it will . . .

[ted ramasola]

Reporting my progress with the builds.
Using 5D mkii with a transcend 133x 8gig card, I use this small card for tests since my 32gig cards are used in projects with the stable 2.3.
Builds:
April 28, I could only record around 8-9frames with corrupted bottom of frame and black bar on side, also exposure seem to vary per frame.
April 29, I could already record 11-12 frames and the frames are now clean, exposures stable throughout.
Now, just uploaded,
April 30 build. I was now able to record 102 frames, DNG!

I bet those using the newer mkIII with fast cf cards are getter longer results especially using 720P which the mkII cannot do.

Go for it guys! If we can get at least 15-30 seconds of raw this would be awesome already!

[hjfilmspeed]

Canon just released the firmware (i know it was leaked) But all of a sudden I dont feel the need to update ha ha

[ted ramasola]

I have edited my previous post that my test using the april 30 build is not 36 frames but 102 frames dng!

-edit sorry its only 51. Half of 102. I missed the break in the take since the first tests I did was locked down on tripod.

[a1ex]

I don't believe it, post them

[ted ramasola]

I'm editing together a resolution chart test to show if it is an improvement versus h264 and also comparing it with a mosaic filter installed.

[Pelican]

btw haeki, a "400x" card should be about 60MB/s but that 400x rating is probably for read speeds, the 7D supports UDMA 7 speeds which would imply it's faster then UDMA 6, which puts its potential max write speed somewhere between 133MB/s and 167MB/s (that's in 1000x territory).  The 7D buffer also supports up to 25 17.9MP 14bit RAW images at about 20MBs per file...  I think your bottleneck is your card's write speed.

The Transcend 400x 32GB card (which I have) can read more than 100 MB/s (specification says 90 MB/s) and write around 50 MB/s (spec: 60 MB/s) in my card reader. In the 7D the write speed of this Transcend card is only 32 MB/s. The fastest card I've ever tried in my 7D is the Lexar Professional 1000x and I measured 68.2 MB/s write speed, so the 7D is very very far from the speed that you mentioned.
http://pel.hu/down/Lexar1000x7D.bmp
http://pel.hu/down/Transcend400x7D.bmp

[Pelican]

I noticed a weird thing where as I'm in the ML menu the Date/Time from the Canon menu flashes on my screen. Running 4.29 Nightly Build.

http://hiddenmoonproductions.files.wordpress.com/2013/04/photo.png

Wondering if anyone else has noticed this.

Yes, it is a known bug. It happens every minute if you put the Date/time menu item to your custom menu.

[P337]

The Transcend 400x 32GB card (which I have) can read more than 100 MB/s (specification says 90 MB/s) and write around 50 MB/s (spec: 60 MB/s) in my card reader. In the 7D the write speed of this Transcend card is only 32 MB/s. The fastest card I've ever tried in my 7D is the Lexar Professional 1000x and I measured 68.2 MB/s write speed, so the 7D is very very far from the speed that you mentioned.
http://pel.hu/down/Lexar1000x7D.bmp
http://pel.hu/down/Transcend400x7D.bmp

Thanks Pelican, that's a little eye opening. 

UDMA 6 spec is speeds up to 133MB/s while UDMA 7 is 167MB/s so I would expect that a Card or Camera that advertises UDMA 7 is surpassing the limits of UDMA 6, but marketing might not see it that way. 

Were you using the newer 7D firmware (1.2.5 I think) that supports faster buffer and UDMA 7 speeds with that 1000x card?  The bottle-neck could also be on the card's max write speed; the fastest benchmarks I could find the Lexar 1000x card is 65.2MB/s for sequential write and a 135MB/s read speed, which I guess technically makes it a UDMA 7.

[P337]

I'm considering a 1000x card for 14bit DNG, but how many frames are you 1000x card owners getting?

Because A1ex seems to be using a 266x card and mentioned that it took about 30 seconds to clear the buffer of about 30 frames on the 5D3.  Each DNG is about 5.09MBs right?  So 30 would be about 153MBs and if it took 31 seconds to write all that to card that's only like 4.9MB/s and I think Alex's 266x card benchmarked at 20MB/s write so doesn't this point to the frame buffer as the bottle-neck?  If that's right then it wouldn't matter how fast the card is, since the camera is limiting the write speed to 4.9MB/s.

That would mean the only answer to get a sustained 24fps is compression before the "frame buffer" (or increasing the frame buffer speeds) and if the buffer really is limiting speeds to 4.9MB/s it would require a 25:1 compression ratio to fit 24 frames through it.  (5.09*24)/4.9=24.9 

If using Jpeg compression that is very close to a medium level JPEG (23:1 ratio), I don't know much about DNGs Lossy or Lossless compressions but can it do 25:1?  I've only ever seen 3:1 DNG lossy compressions.

http://blogs.adobe.com/lightroomjournal/2012/10/dng-1-4-specification-notes.html

Update: With the camera's photo mode set to jpeg only people are getting 53 frames and about 36 seconds to clear which is about 7.5MB/s and would require a 17:1 compression, similar to a "High Quality" 15:1 JPEG compression.  Ted Ramasola just got 53 frames with a 133x card,  I'm pretty sure the cards are not the bottle-neck here, anything that writes faster than an x50 speed (7.5MB/s) seems to have no benefit for these 14bit DNGs. (unless someone can find a way to drastically increase the in camera's buffer speeds)

[ted ramasola]

I don't believe it, post them

oops after checking my takes its consistent only at 51 frames NOT 102. I missed the break in the takes since the shot was locked down.

51 frames using a 133x cf card on 5d mk2.

[P337]

oops after checking my takes its consistent only at 51 frames NOT 102. I missed the break in the takes since the shot was locked down.

51 frames using a 133x cf card on 5d mk2.

Mind if I ask how long does it take to write those 51 frames to your card?  And what is your card's write speed benchmarking at in magic lantern?

[ted ramasola]

Mind if I ask how long does it take to write those 51 frames to your card?  And what is your card's write speed benchmarking at in magic lantern?

Between 31-36 seconds. Have not tried the benchmarking routine yet.

[ted ramasola]

For those who are interested at how it fares on a reschart compared to native h264 .mov and compared to using a mosaic filter and without.

[a1ex]

Can you also upload some DNGs and some frames extracted from the video?

[noisyboy]

For those who are interested at how it fares on a reschart compared to native h264 .mov and compared to using a mosaic filter and without.

Well that answers whether or not this hack helps with moire! Cheers

[ted ramasola]

Can you also upload some DNGs and some frames extracted from the video?

How can I post dng's here?
I can zip them and email it to you.