1
May 26, 2013, 03:03:40 AM
Show Posts
This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.
2
EOS Bodies - For Video / Re: ALL-I or IPB?
« on: May 21, 2013, 11:07:52 PM »
Yes I think a lot of the lore about ALL-I being better had to do with NLEs or post workflows where the rendering of IPB frames wasn't very good. On FCPX you can't tell the difference, but of course FCPX is an all-new state-of-the-art rendering engine (even though many like to think it's still in its launch state feature set).
ALL-I vs. IPB is sort of old news though with clean HDMI and Magic Lantern RAW available. I predict ML will develop further powers (using better internal codecs, higher resolutions, if not higher framerates) for the 5d3 that will even make clean HDMI a nonstarter. Right now in practical use I would use the Ninja 2 with the 1.2.1 firmware, but note, the screen on the Ninja 2 takes about a half second to update the image, which can affect focus pulling a bit. This lag is NOT present for the C100's clean HDMI. The C100 is really the camera I would recommend for video unless you must have stills in the same body.
ALL-I vs. IPB is sort of old news though with clean HDMI and Magic Lantern RAW available. I predict ML will develop further powers (using better internal codecs, higher resolutions, if not higher framerates) for the 5d3 that will even make clean HDMI a nonstarter. Right now in practical use I would use the Ninja 2 with the 1.2.1 firmware, but note, the screen on the Ninja 2 takes about a half second to update the image, which can affect focus pulling a bit. This lag is NOT present for the C100's clean HDMI. The C100 is really the camera I would recommend for video unless you must have stills in the same body.
3
EOS Bodies - For Video / Re: ALL-I or IPB?
« on: May 21, 2013, 12:59:28 AM »
The 1.2.1 clean HDMI out is 8 bit uncompressed 4:2:2 at 1080i60. If you are shooting 24p the Ninja 2 will do the 3:2 pull down for you.
4
EOS Bodies - For Video / Re: ALL-I or IPB?
« on: May 19, 2013, 11:05:23 PM »
There is no perceptible difference between the two encodings that I can detect. Not on moving or still or simple or complex images. The only difference I can find is ALL-I wastes a lot of card space.
What you want to do for broadcast use is get an Atomos Ninja 2 and the new Canon 1.2.1 Clean HDMI firmware for the 5D3. The Ninja 2 will record the HDMI output direct to ProRes in either 1080i or 1080p formats, up to 60i or 30p frame rates. You will have the best quality that way and the files will come out of the Ninja ready to submit and be cut.
http://www.atomos.com
What you want to do for broadcast use is get an Atomos Ninja 2 and the new Canon 1.2.1 Clean HDMI firmware for the 5D3. The Ninja 2 will record the HDMI output direct to ProRes in either 1080i or 1080p formats, up to 60i or 30p frame rates. You will have the best quality that way and the files will come out of the Ninja ready to submit and be cut.
http://www.atomos.com
5
EOS Bodies - For Video / Re: Magic Lantern... CONTINUOUS raw recording @ 24fps on 5D3
« on: May 19, 2013, 11:02:21 PM »6
EOS Bodies / Re: 5D Mark III with Continuous RAW Video Recording
« on: May 18, 2013, 01:49:12 AM »Look at how awful the false color artifacts (rainbow colors) are on the BMCC sensor vs. the 5D3 RAW rendering the pebbles on the path up the middle:
[size=78%]http://nofilmschool.com/wp-content/uploads/2013/05/cinema5D_5D_Mark_iii_BMCC.jpg[/size]
Hmm is this a general issue with Black Magic sensors? Or is it a contrived example?
It tends to be a problem with sensors that work at or near native resolution. The BMCC sensor is only 2.5 megapixels and as a result has no downsampling approach to dealing with moire. At that low resolution, they decided not to use an OLPF because it would lose them their sharpness. With a bayer pattern sensor the grid of photosites tends to cause false color artifacts, which are hard to suppress...but Canon and Sony have learned how to do so quite well. Not so Blackmagic.
Blackmagic has a camera in development that shoots 4K video and is Super 35 sized. However the inexpensive sensor they chose for that is not particularly great at dynamic range. The core ergonomic problems of the BMCC design have also not been addressed in that camera. They have a pocket camera coming out that competes with the GH3 but has the same problems as their current sensor.
What Blackmagic is extremely good at is internet marketing. Not cameras.
7
EOS Bodies - For Video / Re: So, has anyone recorded externally hdmi from 5d3 firmware updates?
« on: May 17, 2013, 03:06:40 AM »
The Ninja 2 is the way to go if you want to use the Clean HDMI out of the 5D3. And there's a lot of good reasons you would want to even if you are interested in the ML RAW hack. For one, you get hours of shooting time, and for two, you get a decent external monitor. You can also get internal proxy recording on the 5D3's own cards simultaneously. And you have the stability of factory firmware rather than a hack.
The quality improvement is a lot better than Andrew (who has a legendary axe to grind with Canon) insists. It's real uncompressed 422 at 8 bit, which can handle the ~11 stops of the 5D3 video when using Cinestyle. It is a bit noisy, you will want to run Neat Video to clean it up at higher ISO's. It's not as good a picture as the C100 recorded to the Ninja, but it's just as easy to work the Ninja on the 5D3. The C100 is a much better video camera in general, but it will never have a full frame sensor or the RAW hack (you will need the C500 for official RAW video on a current Canon).
Right now it may be best to wait a couple months if you can before buying into either option. The cost of the Ninja 2 and a couple laptop drives (you don't need SSDs, rotational ones are fine, unless you want to move around a lot in which case SSDs won't have a gyroscopic effect) is about similar to a stack of the fastest CF cards for the RAW option. The RAW hack right now has a recording time limit of 49 seconds and only 15 minutes or so can fit on a 64GB card. The Ninja can record for hours at ProRes 422 HQ quality (not quite RAW but awesome codec, most of the people using RAW transcode to it anyway, though they can do a primary grade beforehand in something like ACR) without having to change drives or batteries. Also Atomos just came out with the Samurai Blade that has a much better screen (the Ninja 2 screen isn't so pretty, just usable) so a Ninja Blade may be coming out soon.
So I'm waiting to see what happens. The RAW is going to be the best absolute IQ, but at a fair amount of hassle. I may rent a Ninja 2 if a shoot comes up where I need it in the meantime. I don't recommend the Blackmagic Hyperdeck, it's cheaper but you get what you pay for. It's bigger, it needs some external battery solution, it doesn't do the 3:2 pulldown meaning you have to deinterlace manually in post (which sucks), and it has no monitor, and requires a $100 add-on plate just to mount it to something. Blackmagic sure knows how to work the internets for maximum hype, but their products (other than Resolve, which they bought from another company) are frankly half-baked cwap.
The quality improvement is a lot better than Andrew (who has a legendary axe to grind with Canon) insists. It's real uncompressed 422 at 8 bit, which can handle the ~11 stops of the 5D3 video when using Cinestyle. It is a bit noisy, you will want to run Neat Video to clean it up at higher ISO's. It's not as good a picture as the C100 recorded to the Ninja, but it's just as easy to work the Ninja on the 5D3. The C100 is a much better video camera in general, but it will never have a full frame sensor or the RAW hack (you will need the C500 for official RAW video on a current Canon).
Right now it may be best to wait a couple months if you can before buying into either option. The cost of the Ninja 2 and a couple laptop drives (you don't need SSDs, rotational ones are fine, unless you want to move around a lot in which case SSDs won't have a gyroscopic effect) is about similar to a stack of the fastest CF cards for the RAW option. The RAW hack right now has a recording time limit of 49 seconds and only 15 minutes or so can fit on a 64GB card. The Ninja can record for hours at ProRes 422 HQ quality (not quite RAW but awesome codec, most of the people using RAW transcode to it anyway, though they can do a primary grade beforehand in something like ACR) without having to change drives or batteries. Also Atomos just came out with the Samurai Blade that has a much better screen (the Ninja 2 screen isn't so pretty, just usable) so a Ninja Blade may be coming out soon.
So I'm waiting to see what happens. The RAW is going to be the best absolute IQ, but at a fair amount of hassle. I may rent a Ninja 2 if a shoot comes up where I need it in the meantime. I don't recommend the Blackmagic Hyperdeck, it's cheaper but you get what you pay for. It's bigger, it needs some external battery solution, it doesn't do the 3:2 pulldown meaning you have to deinterlace manually in post (which sucks), and it has no monitor, and requires a $100 add-on plate just to mount it to something. Blackmagic sure knows how to work the internets for maximum hype, but their products (other than Resolve, which they bought from another company) are frankly half-baked cwap.
8
EOS Bodies - For Video / Re: Need help with video lighting!!!
« on: May 17, 2013, 02:52:17 AM »
LED grids are also murder on the talent's eyes. Tungsten carries a lot of heat of course and 1K's can easily make people sweat. Plus you can easily overload a circuit breaker with Tungsten, which wastes most of its energy in the IR (heat) band, and the bulbs burn out fairly quickly so you need spares.
There is no ideal video light but tungsten is definitely the best looking and easiest to grade in post. For lighting up a green screen use fluorescents, the green spike helps in that case.
There is no ideal video light but tungsten is definitely the best looking and easiest to grade in post. For lighting up a green screen use fluorescents, the green spike helps in that case.
9
EOS Bodies - For Video / Re: Need help with video lighting!!!
« on: May 15, 2013, 03:38:27 AM »
Given your budget and needs I suggest the cheap Chinese LED panels that are battery powered. Get a couple of the $30 160 LED ones and one of the e.g. fotodiox 312A or 508A LED ones as a key light. Don't get the color changing ones, you will need the full power of all the LEDs at one temperature, and can use gels to change color just as well if not better. Along with some cheap stands and sandbags to place them.
They would be my very last choice to use if I had options but it sounds you have none and if I had to pack only one set for everything I suppose it would be them as I can use them anywhere.
For something that actually looks and works professionally, the cheapest you can do is a Lowel kit. The three Pro Light kit would be what I'd use if I had AC power and needed something small cheap and portable. But it will still set you back a lot more than you are willing to waste on junk.
They would be my very last choice to use if I had options but it sounds you have none and if I had to pack only one set for everything I suppose it would be them as I can use them anywhere.
For something that actually looks and works professionally, the cheapest you can do is a Lowel kit. The three Pro Light kit would be what I'd use if I had AC power and needed something small cheap and portable. But it will still set you back a lot more than you are willing to waste on junk.
10
EOS Bodies / Re: 5D Mark III with Continuous RAW Video Recording
« on: May 14, 2013, 08:29:10 PM »This might be a dumb question, but will shooting continuous RAW kill off your shutter? From what I've seen so far, the camera captures raw images and you must compile them in post, so do all these raw files add to your cameras actuations? Would really appreciate some insight on this, thanks!
It will be no different than any other live view mode: mirror is locked up, as I think the shutter is too.
Everything the sensor is doing is 100% identical here...it's just the cripple codec that is being bypassed. Running the codec is actually more processor intensive than skipping it! So the only critical thing here performance wise is all the continuous write activity to the CF card. We will find out what the fast and reliable CF cards are soon this way...
11
EOS Bodies / Re: 5D Mark III with Continuous RAW Video Recording
« on: May 14, 2013, 07:47:03 PM »
Look at how awful the false color artifacts (rainbow colors) are on the BMCC sensor vs. the 5D3 RAW rendering the pebbles on the path up the middle:
http://nofilmschool.com/wp-content/uploads/2013/05/cinema5D_5D_Mark_iii_BMCC.jpg
Game set and match for Canon.
This all comes out clearly on test chart shots but for some reason we are NOT to have test chart shots. Period!
http://nofilmschool.com/wp-content/uploads/2013/05/cinema5D_5D_Mark_iii_BMCC.jpg
Game set and match for Canon.
This all comes out clearly on test chart shots but for some reason we are NOT to have test chart shots. Period!
12
EOS Bodies / Re: 5D Mark III with Continuous RAW Video Recording
« on: May 13, 2013, 09:03:05 PM »The A/B test to do is the RAW version vs. the HDMI out using Cinestyle, recorded to the Ninja 2 in 220Mbps ProRes HQ and graded in post with the proper LUT. Because anyone who cares about getting the most IQ out of the camera will be using that setup (or something very similar) rather than internal...of course the RAW is going to kill the miserable internal codec.
And BTW, all my testing has repeatedly shown ALL-I and IPB are 100% identical IQ on the 5D3 internal. Haven't seen anything credible to refute that...I think it's just Mbps marketing to counter the GH2 hack.
That is alot of misinfo.
1. The internal codec isn't what does much damage. Point at a static scene and record with Ninja 2 at ProRes HQ and record internally 1.2.1 and there is NOT much difference you can see at all. A little but it is all very subtle to be honest. It's vastly smaller compared to the difference between what this ML recorded stuff looks like compared to normally internally recorded video.
2. For static scenes or ones with just bits moving around in the frame all-i and IPB are pretty much the same and all-i is just a horrible waste of space, perhaps worse if anything. If you pan around or the entire scene is changing frame to frame then ipb totally falls apart and all-i holds up much better (as does say pro res on ninja 2).
Please. I'm not the source of misinfo, I post to keep people honest. I have done quite scientific tests with the 5d3 and Ninja 2 (and the C100) to know very well what I am talking about, compared to the legions of people who have no idea how to properly ask these questions much less answer them. And dear Andrew is too animated in his disdain for Canon to claim any credibility on the matter if that's your authority.
For the hundredth time, video desperately needs an independent lab to publish fully controlled test charts and the like for us to work off objectively. A bunch of well-meaning enthusiasts and blowhards does not suffice. I will test the ML implementations when I get a fast card to do so and I will report back. While not a common net.idiot I am not a substitute for a lab either.
13
EOS Bodies / Re: 5D Mark III with Continuous RAW Video Recording
« on: May 13, 2013, 01:39:06 PM »14-bit RAW on Canon 5D Mark III vs. factory default - Night Image Quality & Dynamic Range on Vimeo
The A/B test to do is the RAW version vs. the HDMI out using Cinestyle, recorded to the Ninja 2 in 220Mbps ProRes HQ and graded in post with the proper LUT. Because anyone who cares about getting the most IQ out of the camera will be using that setup (or something very similar) rather than internal...of course the RAW is going to kill the miserable internal codec.
And BTW, all my testing has repeatedly shown ALL-I and IPB are 100% identical IQ on the 5D3 internal. Haven't seen anything credible to refute that...I think it's just Mbps marketing to counter the GH2 hack.
14
EOS Bodies / Re: 5D Mark III with Continuous RAW Video Recording
« on: May 13, 2013, 12:36:38 PM »I wonder why the Canon legal team would look the other way on this hack since it's on the 5D3, but throw down the hammer whenever hacking the 1D X is mentioned?
Well first of all the 1DC is just Canon's hack on the 1DX. I don't buy that there is some required heatsink, that's window dressing. They don't want 3rd parties competing for sales of hacks at that price point ($5,000+).
The second thing is the 1DX has a proper downsampler for its video rather than the 5D3's pixel binning. Which is necessary given the sensor dimensions (unless they wanted to go back to the hideous 1st gen line skipping). And so a hacked 1DX recording less lossy compression (with more actually distinguishable pixels, and more luma/color levels, in more situations) would give the C series a run for its money while this 5D3 hack is just going to please the ambitious kids that think they're getting a bargain. For those kids $3000 for a 5D3 is a lot of money and about all Canon can expect out of them. But people who will buy a 1DX are professionals that will prefer going C-series and Canon wants to usher them that way (for their own benefit I may add) rather than have them confused by kids and their hacks.
15
EOS Bodies / Re: 5D Mark III with Continuous RAW Video Recording
« on: May 13, 2013, 11:55:13 AM »and might be better than the C300 and definitely better than the BlackMagic Cinema Camera.
The BMCC still has more than a stop better DR (than the 5D3), and as of now, more resolution for continuous shooting. And while the c300 is a much easier camera to use, BMCC's best image blows the c300 out of the water.
Oh please not more BMCC Koolaid. The C300 image (especially recorded externally) is one of the cleanest 1080p images in the entire business with virtually no moire or false color artifacts and lovely color science. The BMCC fanboys (some of whom are or were apparently secretly in the employ of Blackmagic Design) keep repeating nonsense but repeating something ad nauseum does not make it true. IQ test: hand a better DP (who isn't taking money or gear from either company, and who didn't learn what they know from reading online forums) a C300 and a BMCC and see which one s/he hands to the intern to shoot BTS.
Yes even Andrew (who I think, to his credit, is too random a figure for anyone to consider buying off) has declared the BMCC dead with this new hack. It's just about as annoying to use as RAW on a BMCC and the 5d3's a much better sensor and camera. I think even the official HDMI out from the 5D3 recorded into a Ninja 2 is a superior image to BMCC RAW.
RAW is really not all that as so many people insist. If the only choice is between RAW and JPEG, ok, we all choose RAW. But ProRes 422 HQ 10 bit with a log gamma is every bit as good as RAW in practice, and vastly more practical. JPEG (and the video version MJPEG) is very lossy like H.264 is, and 4:2:0 AVCHD is very constricting in post. But the system we have working today with the 5D3, recording the entire pixel-binned sensor in 8 bit 422 with Cinestyle onto a Ninja 2 in ProRes HQ is quite a nice image. (Though not nearly as nice, lowlight or otherwise, as the C series image externally recorded.)
This RAW hack will give only an incremental improvement over that, in one of two ways. You will get 12 bit color but still it will be subsampled to 4:2:2 (not exactly RAW in my book, which I define as "a lossless record of all sensor information"). And your dynamic range will be stored in full fidelity 14 bit as opposed to being mapped to a gamma...gaining you about 1 stop in practice and a bit less propensity for banding than using the Cinestyle/Ninja approach. Worth constantly swapping costly CF cards and then having huge post hassles for? Maybe in an extreme HDR run-n-gun situation where it would be more hassle to use gels/lights/reflectors/butterflies to control DR.
The other thing against this hack is, if they are using the full sensor, they must be downsampling (binning?) to 1080p or similar resolutions. Which will give at least some of the 5D3 soft video look. Either that or they are cropping (and I understand at least in some settings they are) to the native resolution of the sensor. This cropping to 1:1 pixel is going to be terribly noisy and will require NR in post. You can also get this cropping/digital zoom trick with the current HDMI out, by zooming in 5x or 10x in focus assist and recording the zoomed image (which comes out 4:3 but you can crop in post) and it is noisy too.
Overall I think this is a good development only because it silences a lot of idiocy coming from the Blackmagic fanboys. I don't know how their paid operatives will spin this, other than "We don't force you to use hacks! We're on your side!" Well the entire Blackmagic camera line is a series of hacks so that doesn't change much. If ML is able to make this really pleasant in practice (perhaps with a hi-def codec...I will have to see how MJPEG looks) then this will replace the HDMI out/Ninja option for people not needing long shooting times. But I don't see any scenario where this hacked 5D3 is a superior image to a C300 or C100 + Ninja 2 under any circumstances. That's silly talk.