Messages

Isn't the camera converting to 8bit during h.264 encoding?  Maybe you could just grab it there.  Anyway, I believe this is missing link.  Made a 2 minute youtube vid and I ended up with hundreds of GB's of footage in various formats along my workflow production line.  10 bit is better for a lot of reasons.

Anyway - cheers man! What resolution did you shoot at do you mind me asking?

1.5K @ 2.67:1 That was with my card getting 33MB/s.

With new card I can get 1.5K @ 2.35, which is nice.   Only a 1.25 crop, still very large.  The issue is that the moire is effectively being magnified by 1.25. Not good.

I wonder how much the mosaic antimoire filter helps... I might have to buy one of those.  It's just the way the h.264 algorithm is set up.  I think it has a denoise algorithm built in and that both reduces detail, thus reducing card speed requirements, and reduces moire.  Try putting the "dust and scratches" filter in premiere on.  Definite reduction of detail but also reduction in moire.

Plus I had no idea what I was doing with some of the clips.  I shot hand held because I was traveling and didn't have a tripod and I seriously didn't think I would do anything with the footage, then I popped it into ACR and was blown away.  I got a little carried away with the colors, but maybe you won't believe me when I tell you that the saturation is on average @ -30.  I had to turn it down.  Really a stunning day, very lucky.  Too bad about the ridiculous moire though.

Does anyone know how to embed to autoplay in HD?  Might just not be possible

First one to post a video!  Yeah, if you guys like it I'll make a tutorial for how my work flow works.  The trick is to batch process as much as possible because the render and sequencing times are enormously large. 

I sort of learned as I went how to get the best look.  One thing to consider is that moire and aliasing are often invisible until you sequence it in full HD.  Moire brush can work well, but often slight differences in the algorithm will show in between the frames.  My favorite is actually to use the de-noise algorithm, maybe a little bit and keep the details up and defringe under the lens corrections tab, just max out the sliders. 

After that I upscale using After Effects, I believe it's bicubic.  At this point it's at pretty much the same level interms of sharpness as H.264, maybe a little higher.  However, the color details are massively higher.  Youtube fucks it ofcourse, obviously none of the macroblocking is in there on my computer. 

26 hours of rendering later... http://www.youtube.com/watch?v=cnlooAejpZQ&feature=youtu.be 

Where are we posting these again?

yes i understand this.  The issue I have had is when you turn off raw and gop/flush the sound stays off.  Thus prompting noobs like me to scour the ML menu for a way to turn it back on.

I was freaking out when I had an important shoot and audio wasn't working, took me a while to trouble shoot.  I think for the release candidate audio should default to on when you turn off and on the camera.

Yea, shouldn't have sold 50D... CF cameras are boss and have more memory too Who would have thought.. now you can probably sell your 50D for $1k

Just luck of the draw.  50D is pretty absurd, also I think the price is going to level out after a few weeks.  Low light is same level as 5d mk ii.  But 6D is ridiculous, with raw it's pretty much a night vision camera.

I think when the 7D gets raw that it will be the real winner.

Lol - like the thread title  Catchy!

Yeah - something I was gonna post in the 6D - Magic Lantern and raw Shooting General Discussion about the lack of 6D raw videos. There are tonnes of vids showing off the other cameras but none of our beloved 6D's. I'd like to start a thread called "Raw Vids Shot on 6D" under Share Your Vids but haven't got anything to post myself 'til probably tomorrow night so seems a bit silly  In fact I'll start it and then link to it from here and the 6D general thread. 

I'm too much of a perfectionist to post just any video.  Hopefully all the work I've done will show how great the 6D is.

I think the resolution is fine.  I upscale the images, then lightly sharpen.  Then downscale in premiere, I have no complaints.  I shot some stuff @ 12800 3 ev under exposed and it's pretty impressive, plus no banding!! 

I honestly believe that for most shots, as long as it's raw and you upscale correctly, you could easily upscale to 3K and it would just look a teeny-tiny bit soft up close.

Only issue I have is that it's not full frame.  How does the 720p 24fps override work?

Yeah, well said. Even if I had 5D3 now I couldn't afford 1000x cards, 128GB costs as much as a nice lens 
We need to make some good comparison video with the res we have now vs 1080p. (I think we could use LV_rec module to get couple of frames in real 1080p raw on 6D as a benchmark) There are plenty of comparisons videos but always ungraded, wrongly exposed h264 with no sharpness/cinestyle with no lut applied vs nice raw.

I have some examples I'm working on now.  The difference in sharpness is stunning especially with good upscaling and sharpening.  Night and day.

With a 16:9 aspect ratio, you can record 24 fps at 720x1280 continues without frame skipping. With exfat formatted SanDisk pro card, it keeps recording till I decide to stop

Others are trying other aspects. Basically at the bottom of the screen ML will tell you the bit rate needed. Any size resolution that keeps the bit rate under 40mb seems to record continuously on the pro card

I just got that patriot 32GB, It's getting 39.5MB/s formatting is fat32, does exfat help?  It really seems like 40 is the speed limit.  Sandisk is still the best bet I think.

Anyone having this problem with frame inversion?  It looks like it drops a frame temporarily then writes it one frame late.  It doesn't ruin footage, but it's an issue.  Might just be an earlier version though.  I'm going through the 200GB I shot earlier this week.

Could you reveal the secret and tell us what card it is? I will keep my fingers crossed again for this 50...

I don't want anyone to go out and buy it then find out it sucks.  It's a patriot card.  http://www.patriotmemory.com/product/detail.jsp?prodline=4&catid=88&prodgroupid=238&id=1275

As I understand it, they all read at 90MB/s but only the 32 writes at 50, the rest are 35.  Makes my mouth water when they talk about recording 4K HD video.

Wow, I print photos for a lot of people as a hobby, 19x11's, I had access to the print lab at my university and I took full advantage :} People go bananas but I get my results only using ACR.  Now that I can use ACR for video, it's making me a little crazy.

Wanted to thank you guys.  This is an excellent app.  Everything worked perfectly and I enjoy this workflow a lot.

Also - I know we have discussed this a lot but choosing S1 jpeg def increases writes speed. EVERY time I choose it the speed jumps straight to 39+ then levels out at about 38.9-39.1 (never had these speeds before).

When I choose S3 for example - levels at about 38-39 but takes a while to get there. dunno if it's simply switching between the modes or the modes themselves but there's a noticeable difference and a solid pattern to prove that s1 wins.

Can other people try this so we can see if there is a difference? Which is better for you? S1 or S3? (S1 Higher quality by the way).

I see no difference between any of the S's. about 33.1 on a card rated for 35 right off the bat for everything.  I finally found a card that has a claimed write speed of exactly 50MB/s, comes in tomorrow, hopefully it works.  I wonder if there is a way to hack the camera to not record images at all?

The moire thing is pretty much clean actually with ACR anti moire brush and it upscales really nicely, but now the biggest thing I'm noticing is aliasing.  Where are we posting workflow?  I think I have a pretty good setup, can't wait to finish editing.  When we can get sustained burst of 1720x720 that's when this will become a boss feature.

Ooo one more thing, any chance on a config file for raw settings?  I'd really like it if it would pop up everytime with my settings. 

Thanks Devs!  Great job on this.  Never thought raw would be more reliable than h.264.  My camera has not crashed, at all.

The algos are all in the encoder part of the digic, I'd have to find where to manually set current QP, I don't know if you can just do it on the fly... that location is "init" or start qp and is probably only read once.

well if you do find it this is what the algo could look like:

if (currBitRate > maxBitRate){
...currQP = currQP +1 //try to get below max bitrate
...//alternatively: currQP = currQP + (currBitRate - maxBitRate)/K  (just to really cut it down if it needs to)
}

else if (currBitRate < maxBitRate & currQP > initQP){
...currQP = currQP - 1 //try to get to target qp
}

else{
...//it's all good (or maybe there could be a statement for a min bitRate if you wanted and it would bump the QP for that)
}

I'm not trying to be condescending or whatever (if this looks crazy just call me a moron and move on), I only understand the concept barely myself, and I have no idea how QBasic or whatever the firmware language is works.  I just think that this is a good way to do it because generally, the camera gradually comes into a complex scene, when you pan or focus pull into it, it usually doesn't just happen and if the current bitrate was signifigantly higher than the max then it would need to really cut the QP fast, so there could be a scaling factor for that.  It might be a good idea to start the currentQP signifigantly higher, like around 40 higher, than the target and then work it's way down, it would take not more than 2 seconds at 24p.  I don't know, just a thought.  I can't do any tests right now because I lent all of my cards and 600D out to a friend, I'll do more tests when I get those back.

Yea, you're recording fixed QP so it will drop frames when it can't save them at QP whatever. Thats why I increased the max qp in the configs and don't do fixed QP in general.

Right, I just recorded two files, one minute long on RC at bitrate = 60 Mb/s.  Calculated out that's exactly 450 MB for one minute.  I looked at the files and sure enough, exactly 468MB and 471MB.  What would be nice is if there was an RC where you set the target bitrate at something like 30Mb/s but have a max of something like 120Mb/s.  I thought that's what VBR was doing because of init QP, I assumed it just aimed for the init QP then altered if it went over the max bitrate.  I can imagine that kind of algorithm is ridiculously hard to implement on the fly though, and probably takes a few passes to get right, which is why arri, red, and BMC guys just say eff it and record the raw.

It's nice to be able to fall back on RC though if you just can't get the VBR to a nice number without it blowing up.

I don't know what improvements have been made in this area, but I want to do an update on the bitrates.  IPB is not safe to use with ANY bit rate modification.  It will drop frames when put under pressure, it happens very rarely, but I was photographing a tree on a windy day and everytime the wind blow and started rustling the leaves the read out would tickle 80%.  I looked back at the footage and sure enough, it had dropped frames. 

On All-i with the same settings it never worked for more than a second.  So I turned it up a tiny bit to qp8 and it was pretty good, same sized footage though, which is a large ~1GB/min. 

http://www.youtube.com/watch?v=wcY_WotjwLE

That's basically what they are doing with the newest sensors for the Black magic.  You can adjust dynamic range in camera believe it or not.  The trade off is noise and less accurate color representation I think.  They're getting crazy DR of 20 stops now, but the latest BMC is tuned for 12 out of the box.

There's a company in the Netherlands called Cmosis that is devoted to producing only sensors.  They make all the sensors for Black Magic and Leica. They're making full frames now too, won't be long until we'll start seeing full frame cameras with crazy high DR in pocket cameras.  It's pretty much a game changer.  It would be cool if magic lantern could somehow team up and get one of those in a 6D.  Would be pretty ridiculous.

Dude that was a GREAT post! Thank you

I appreciate it, I'm glad you found it useful.

The $6 or whatever it is for VisionColor is totally worth it and it's also now supported by FilmConvert too! Yippee

Wow, this is impressive.  Thanks I'm going to get this.

Based on some recent tests I can tell you that increased bit rate has a marked effect on the observed image quality.  Best way to test this is with an image with a few flat gradients and a good amount of detail particularly in the shadows.  You will see heavy macro blocking in the shadows and banding in the gradients.  This is reduced continuously as you use higher bitrates, even the difference between 100Mb/s and 200Mb/s is going to show a difference in the banding and macro blocking.

The settings that I use for my camera right now for actual shoots where I need 100% reliability is IPB: VBR: init qp = 6 (any lower and its a risk of it blowing up) max bitrate = 10.  This gives good results and has been 100% reliable thus far.

The biggest test on the encoder is to aim a something continuously moving with massive detail.  The encoder will have lots of trouble.  In low depth of field situations where one will be aiming at faces for example and other things that don't move erratically, one can easily push to qp=1 max bit rate =20 and still maintain reliability. 

In other news I've found the best picture style.  http://cinescopophilia.com/download-the-vw-158-canon-dslr-picture-styles-picture-profiles/

The CrookedPathFlat_3.0_HilightHelp.pf2 maintains excellent saturation and detail in the highlights.  It's really amazing.  I was shooting out my window on a sunny day and I had usable detail in the room and still had a blue sky outside.

From what I'm reading the IPB method could be used to encode extremely high quality video, the problem is that the encoder memory is limited to the memory of all the frames in between the I frame and the P frame.  Basically, it has to encode it all in a chunk like that because the B frames are based on the I frame (what comes before) and the P frame (what comes after).  If you want higher bitrates the number of B frames needs to be limited so that the buffer isn't overfilled.

Ideally a simple algorithm could be implemented such that 2.0x bit rate would make IPB at half the number of B frames as the normal compressor aims for, 3.0x would be a third, while .5x could have twice the number of B frames, etc. 

I really liked the 600d method of setting bitrates, but the IPB compression in this camera is also a huge advantage

From that article that I recommended earlier
"For high bitrates, it's usually preferable to use a GOP without any B-pictures"

How would this be done?

Anyone interested in whats going on with IPB and GOP and all that, take a look at this: http://dvd-hq.info/data_compression_3.php

Very comprehensive explanation for how video encoding works.

I have a question about the difference between rate control and vbr.  How does rate control differ?  CBR sounds like it already is rate control no?

Hi guys, back from a long shoot.  I have great news, using adobe speed-grade (I finally broke down and got the adobe cloud deal, I'm a student so this is awesome) you can really bring out the footage.  It looks really good and because it works so well I would recommend shooting underexposed by 1.5 stops, unless there are no blown highlights.  By doing this you save the highlight data and speedgrade can bring back the underexposed mids, then you just leave the shadows crushed for a cinematic look.  For the highlights you simply lower the gain and gamma, and slightly increase the blur for nice highlight bokeh that has a smooth roll off.

I recommend setting the camera to 120 Mbit per second with initial qp at 2 or 3 on vbr.  The result is that it is pretty reliable and very color gradable.  I've gotten footage that looks a lot like the BMC.  The BMC has 13.5 stops while the cannon has 11.3, so by under exposing by 1.5 stops you can bring the highlights just about to where the BMC is. 

**Do not attempt this if you want to colorgrade inside you non-linear editor, use a dedicated colorgrading program, it makes a big difference.

For internet streaming I recommend setting the camera to 30-40 Mb/s, init qp at 4 or 5. it saves a lot of space and will look identical after compression on the internet.

And for presentations with projectors shooting 800x600 I recommend setting at 20 Mb/s init qp = 8.

Hope this helps a bit.

What? I am using the new Yongnuo flash triggers. I can see front/rear curtain in the camera menu. I am even using a vivitar ETTL flash and the trigger works with it. With this trigger you can sync past 1/180th also using HSS(if the flash support it).

Interesting, I have a wireless flash trigger and when I try to pull up the menu it tells me: "flash not connected or not recognized" I will update my .bin to the latest and see if that fixes it.

yeah, I can push the shutter speed to whatever I want, but when I push it past 1/180, the shutter curtain starts to cut part of the flash off.  I'll play around with it and see what works.

Thanks guys.

This latest .bin is really, really good.  I almost wish that it would default to vbr flushrate = 4, qp = 1, bitrate = 8 because it's very reliable at that rate and produces good footage. 

I have one feature request.  Canon in it's infinite wisdom removed the flash on rear curtain functionality from it's menus.  As a result I can't take great motion trail shots with a rear curtain flash if I'm using a non-canon flash.  I'm using a yongnuo speedlight, which is really excellent for the money.  The thing doesn't sync past 1/180 shutter, which is fine, but I would really like it to take rear curtain shots on 1/4 second shutter, which has produced fantastic results in the past.


Also, Is there a way to do shutter sync through the remote trigger port in movie mode?