Luminous Landscape Forum
Equipment & Techniques => Motion & Video => Topic started by: l_d_allan on October 03, 2013, 06:27:53 am
-
from:
http://www.luminous-landscape.com/reviews/camcorders/black_magic_pocket_camera.shtml
> this camera is capable of producing stunning image quality. Quality that far exceeds that of any video DSLR,
I'm curious how much of an "equalizer" the Magic Lantern raw video capability (and other advanced video + focus + metering capabilities) provides for Canon DSLR's, especially the newer 5d3.
Or is this "apples and oranges"?
-
I saw some footage of Canon raw DSLR footage when the hack first appeared and it was markedly better than the BMCC in some comparison shots. The link to it would be in the Magic Lanterns forums somewhere.
-
hard to know what markedly better means. Depends on grading, capture codec etc
I have both a 5d and BMPCC and markedly prefer the output from the latter
-
raw aside, color quality from ML is 4:2:0, the bmpcc is 4:2:2
-
hard to know what markedly better means. Depends on grading, capture codec etc
I have both a 5d and BMPCC and markedly prefer the output from the latter
Is that raw from the Canon too?
The shot of the path up to the trees is in this comparison (http://nofilmschool.com/2013/05/blackmagic-cinema-camera-canon-5d-mark-iii-raw-video-test/) is where the BMCC wasn't as good as the Canon with the raw hack - which has been much improved since this test. And the article also mentions poorer high ISO performance compared to the Canon.
-
Is that raw from the Canon too?
The shot of the path up to the trees is in this comparison (http://nofilmschool.com/2013/05/blackmagic-cinema-camera-canon-5d-mark-iii-raw-video-test/) is where the BMCC wasn't as good as the Canon with the raw hack - which has been much improved since this test. And the article also mentions poorer high ISO performance compared to the Canon.
I know when you test these motion cameras, it's much more than just color information, or bit depth, even at times mbps/kbps.
When you look at 444 which is 100% color info, 422 50%, 420 25% you'd think at 420 you'd have a file that would be lifeless, but there is so much more to digital video than just the numbers.
What most of these tests don't show is cameras at their true iso.
I don't know about the black magic (other than the people I work with that have used them on set), or for that matter a hacked 5d2/3, but for the RED 1's, the base iso is 800. Actually the iso is always 800 and all your doing when dropping or raising the iso is pushing or pulling a curve.
Actually with the R1's, pulling it down to 100 is worse than pushing in regards to highlight clipping, (at least in my experience) and now we just set the camera at 800 and leave it, using filters, or light to get to the desired exposure.
It's funny, the Reds shoot 444, the gh3's we use shoot 4:2:0 but under the right conditions you can't tell them apart, even when moving the files in post. Then again, under other conditions the difference is huge.
I'm not an engineer, but I know there is a lot going on under the hood of these things that makes a huge difference, especially combined with different lighting, subject, ambient color, ambient bounce and the big killer, heat.
With the R1's I would leave them on, when shooting stills, or just to avoid the long boot up time and if you then go into a long take, you start building up noise.
The Gh3's will do the same, though not as pronounced and I've been told the original bm 2.5k cameras are also affected by heat.
So comparing these cameras is a difficult task, especially if you just go from the specs, because all of these cameras have a sweet spot. When your on, they are all marvelous, when your off, they can be a mess.
I think Digital Video is where digital still capture was about 10 years ago, with the quality and use all over the place and I assume in a few years with improved and more standard software, graphics cards and better camera design, they will all level off like still capture has.
Not having any inside information, I think the next step is a true combination camera, something like a Canon 1dc 4k and a Panasonic gh3, but both with better sound inputs, maybe even some dedicated coloring/processing suites.
IMO
BC
-
Alexa's native iso is 800 too, like the R1.
IMO, a part from Red (Arriraw is too expensive), raw video has still a long path to recover until it becomes
as matured and widely supported as still imagery, hopping they will avoid the proprietary mess, wich I think they obviously won't.
But guys, I don't get a point there: how can Canon hack claims raw capture if deliver 4:2:0 and BMPC 4:2:2 ?
No possibility of 4:4:4 ? Absurd. Better get a good Prores or DNx 444.
I've graded a lot of imagery, from a lot of cameras and all I can say and be definite, is test, test, test.
Every camera looks a little different, but lighting and subject, including ambient bounce have as much to do with the final look (and resolution) as anything the camera does.
Obviously a 1/3" chip, avchd file won't hold up under high iso and hard 13 stop lighting, but then again hard 13 stop lighting probably shouldn't be used anyway.
I just finished an edit with RED and GH3 footage mixed. I rarely compare imagery direct, but had one scene shot A cam with the R1 and we shot b camera with the gh3's and it's a fairly challenged look at the end of the day.
Bottom line is I can't see any difference, I mean any difference and the RED is 444 the GH3 420 and if there is a difference it's so minor it will never be noticed.
This doesn't hold true for all imagery, but then again, a 420 file is 1/4 the color information (on paper) of a 444 file and we should see a difference . . . we don't.
I know other people will disagree, but I have nothing against a baked in gh3 file vs. a 444 raw. The secret is to shoot the gh3 slightly flat, but not c-log, flat and colorless, but just slightly smoothed out, then grade to your hearts extent.
If anything I would like to see is the next round of cameras shoot a prorezz file, rather than an h264, even if it's a larger size and not for quality, because a high bitrate h264 file looks virtually identical to a prorezz file, but just do prorezz so we don't have to transcode.
IMO
BC
-
Bottom line is I can't see any difference, I mean any difference and the RED is 444 the GH3 420 and if there is a difference it's so minor it will never be noticed.
This doesn't hold true for all imagery, but then again, a 420 file is 1/4 the color information (on paper) of a 444 file and we should see a difference . . . we don't.
I don't know about the black magic (other than the people I work with that have used them on set), or for that matter a hacked 5d2/3, but for the RED 1's, the base iso is 800.
It's funny, the Reds shoot 444, the gh3's we use shoot 4:2:0 but under the right conditions you can't tell them apart, even when moving the files in post. Then again, under other conditions the difference is huge.
Just want to mention a couple of things.
It's not really right to describe a RAW camera as having colour encoded space like 4:4:4 or even 4:2:2. This is the wrong terminology. This is video encoding terminology.
RAW cameras like the RED and the BMCC and even the ML hacked 5D are just that RAW cameras. They can be Log or Linear encoded, and you can TRANSCODE 4:4:4 or 4:2:2 or even 4:2:0 files from their RAW files, but you don't refer to them by these descriptions.
That's because there's a lot of other things that affect the colour subsampling ratio.
It also gets confusing because Bayer sensors, which these RAW cameras generally have, have differing ratios of GREEN to BLUE to RED pixels so some would even argue that you can never have a 4:4:4 image without massive oversampling. (an 8K sensor like the SONy F65 for true 4K 4:4:4 encoding).
A lot of people ALSO make the MISTAKE of assuming that the RATIO in bayer sensor must have something to do with the RATIO's used to describe video encoding. For example, that because bayer sensors generally have 2 green photo sites for every red and blue photo site, this lines up nicely with 4:2:2. It's a coincidence. You need to look way beyond this gross simplification and understand it's a lot more complex than this.
Someone will likely jump on this and elaborate, but it's all kind of futile. RAW sensor data can't be described as video encoded space like 4:4:4 until it's transcoded. And even then you'll have endless arguments about what the original sensor size needs to be to derive that number.
I should also add, that even though can't see the difference, you will certainly see it once you start to grade some files. And here its' just like comparing a JPEG to a RAW. So if you can get a JPEG right out of camera then you're golden, but as soon as you want to push it around, you are quickly in trouble....
You also describe that the R1 has a native ISO of 800. It's actually 320, as is the R1 MX and the EPIC (which uses the same MX sensor)
RED have long advocated exposing at 800 ISO to protect the highlights, and that's what they call a "recommended" exposure. I personally find the RED has very little HEADROOM and rating at 800 is their way of trading some shadows for highlights.
jb
-
Good to see you here JB. I have a BMCC and am loving it.
Does anybody here have the BMCC and the Canon 5d3 with Raw? I heard the Canon raw has 14 stops of DR, and the BMCC has 13, I am curious if this is true.
-
John,
Ok your right, it is linear until encoded to 444 or whatever, but regardless, a 444 out of cinex doesn't look much different than a 420 out of a h264 gh3 of the same scene.
I've run em side by side enough to know that part.
Now the color response and look can be different, but in the end there is not a heck of a lot of difference in color visual depth or dr. Maybe the REDs have a stop more.
In regards to iso, yea, I've heard 320, I've heard 640 but every dp I know says lock it on 800 and don't pull it down to 200 or it clips. In fact with all digital video I find I have a lot better chance to recover shadows than pull in detail from highlights. I know from experience don't go past 1000 with the mx sensor, ever.
Still it depends on what, how you shoot. I do know that with any motion image how you set up in the front end can save you a week in the back. These things, raw or cooked are sensitive, have their own properties and getting a base white balance then filtering for color I find preferable to moving the knobs trying to get a look from the camera.
I now treat all dv, either raw or cooked like I'm shooting transparency film. I get my base, filter for the rest and don't try to produce a 100% different look in post. That's a fight and meant for people with a lot larger budgets than mine.
People always like to say, _______________was shot with a _________ s it must be good, but they kind of forget that when they are looking at big budget theatrical they are looking a huge money and time.
Gravity took 4 years to do and they invented technology as they went along. Not that it wasn't worth it because it's a good movie, but it's also something you can't do with any camera, working with two lights in the basement.
Today, someone brought over some underexposed 5d3 footage from a recording session shot at a billion iso, with a wb of daylight, shot tungsten. I dropped it into color and opened it up, moved some of the red out and it was pretty, though very noisy, but still pretty and that's from a cooked h264 file shot at the wrong wb and about 2 stops under.
Then again, everyone works different though nothing and I mean nothing improves the look of a digital motion file like time and money. With time and money in the pre, shoot and post production you can make anything look good if you work professionally.
Given that I don't think raw is a magic godsend any more than I think one camera is 100% better than the other.
IMO
BC
BTW: I love the shot on your site titled Reef Ireland. Very pretty.
Just want to mention a couple of things.
It's not really right to describe a RAW camera as having colour encoded space like 4:4:4 or even 4:2:2. This is the wrong terminology. This is video encoding terminology.
RAW cameras like the RED and the BMCC and even the ML hacked 5D are just that RAW cameras. They can be Log or Linear encoded, and you can TRANSCODE 4:4:4 or 4:2:2 or even 4:2:0 files from their RAW files, but you don't refer to them by these descriptions.
That's because there's a lot of other things that affect the colour subsampling ratio.
It also gets confusing because Bayer sensors, which these RAW cameras generally have, have differing ratios of GREEN to BLUE to RED pixels so some would even argue that you can never have a 4:4:4 image without massive oversampling. (an 8K sensor like the SONy F65 for true 4K 4:4:4 encoding).
A lot of people ALSO make the MISTAKE of assuming that the RATIO in bayer sensor must have something to do with the RATIO's used to describe video encoding. For example, that because bayer sensors generally have 2 green photo sites for every red and blue photo site, this lines up nicely with 4:2:2. It's a coincidence. You need to look way beyond this gross simplification and understand it's a lot more complex than this.
Someone will likely jump on this and elaborate, but it's all kind of futile. RAW sensor data can't be described as video encoded space like 4:4:4 until it's transcoded. And even then you'll have endless arguments about what the original sensor size needs to be to derive that number.
I should also add, that even though can't see the difference, you will certainly see it once you start to grade some files. And here its' just like comparing a JPEG to a RAW. So if you can get a JPEG right out of camera then you're golden, but as soon as you want to push it around, you are quickly in trouble....
You also describe that the R1 has a native ISO of 800. It's actually 320, as is the R1 MX and the EPIC (which uses the same MX sensor)
RED have long advocated exposing at 800 ISO to protect the highlights, and that's what they call a "recommended" exposure. I personally find the RED has very little HEADROOM and rating at 800 is their way of trading some shadows for highlights.
jb
-
I've graded a lot of imagery, from a lot of cameras and all I can say and be definite, is test, test, test.
Yup.
Every camera looks a little different, but lighting and subject, including ambient bounce have as much to do with the final look (and resolution) as anything the camera does.
Even the same camera can look very different at times. Sometimes matching two cameras in same lighting setup is probably easier than matching same camera with very different lighting conditions. Particularly if ISOs are very different. Though this is where a raw file is particularly useful as at least you can colour balance after the fact which can to help match shots.
I also think video is like shooting transparency, so is like stepping back in time as a result.
-
Yup.
Even the same camera can look very different at times. Sometimes matching two cameras in same lighting setup is probably easier than matching same camera with very different lighting conditions. Particularly if ISOs are very different. Though this is where a raw file is particularly useful as at least you can colour balance after the fact which can to help match shots.
I also think video is like shooting transparency, so is like stepping back in time as a result.
In an interview type scene we have set up two R1's with every setting identical and one Scarlet. Just working three angles when you go to cinex out of the can looks, there will be a difference, mostly just from the camera angles and color ambient bounce, maybe due to a different exposure setting. The R1's to the Scarlet never match exactly, regardless of RED gamma and RED color settings.
I just believe that dv is very, very receptive to ambient color and bounce or there is something else going on under the hood that I don't understand, but the looks are different.
In fact watch a news roundtable discussion where they run multiple cameras and you will also see a slightly different skin tone and color look from each angle and this is from studio situations where nothing moves.
IMO
BC
-
In an interview type scene we have set up two R1's with every setting identical and one Scarlet. Just working three angles when you go to cinex out of the can looks, there will be a difference, mostly just from the camera angles and color ambient bounce, maybe due to a different exposure setting. The R1's to the Scarlet never match exactly, regardless of RED gamma and RED color settings.
I just believe that dv is very, very receptive to ambient color and bounce or there is something else going on under the hood that I don't understand, but the looks are different.
Lenses can make a big difference too. F stops can be somewhat shall we say optimistic at times and is why t-stops are used on cine lenses. My f2.8 primes are much faster than my f2.8 zooms. Not to mention the colour variation between different lenses. Had issues matching shots from 2 different Canon bodies because of this once when doing a two camera setup.
-
John,
Ok your right, it is linear until encoded to 444 or whatever, but regardless, a 444 out of cinex doesn't look much different than a 420 out of a h264 gh3 of the same scene.
I was making a few points, but I do think it's worth making distinction here about what "looks" to the eye fantastic, though its easily fooled and then what happens when you try to grade it later.
This is when H.264 will fall flat on it's arse. It can look great OOC but once you push the grade around you will soon "SEE" a difference. And that's where RAW cameras do have an advantage....once you need to move into any kind of PP work, higher bit depth and less compression means better results. Just like RAW Vs JPEG....
In an interview type scene we have set up two R1's with every setting identical and one Scarlet. Just working three angles when you go to cinex out of the can looks, there will be a difference, mostly just from the camera angles and color ambient bounce, maybe due to a different exposure setting.
I shoot narrative TV drama and we regularly use more than one camera, and often 3 and as many as 6 on a single scene.
I'm in pre now for a drama series which in it's 5th season. We shot the first RED one, then RED ONE MX and then EPIC and now Alexa.
None of the cameras ever match with all the settings being the same. Every camera is different even in RAW and it just gets compounded by slight differences in lenses (even the same versions), ND filters and the biggest of all, different shooting angles.
Even at the same exposure, shooting from one angle light gets reflected off of a persons face totally differently to shooting at 90 degrees offset, so they will never ever match at "default". That's exactly why you have a colourist and the footage needs to be graded. To me, that's an even bigger leap than the simple small differences between individual cameras. And that's also why higher bit depth and less compression counts for a lot to me....
I have four Alexa bodies on the main unit of the show I'm now and two on the second unit. Out of 6 bodies, I'm lucky if two of them line up the same. There's even bigger differences when shooting EPIC.
jb
-
When you look at 444 which is 100% color info, 422 50%, 420 25% you'd think at 420 you'd have a file that would be lifeless, but there is so much more to digital video than just the numbers.
For most natural scenes, 4:2:2 and 4:2:0 throws away information that was hardly there in the first place.
For "reasonable" editing, 4:2:2 and 4:2:0 seems to matter little to most human viewers.
For imagery that will be lossy encoded for distribution, 4:2:0 is the norm.
For camera sensors that have a Bayer CFA at the same pixel count as that of the luma channel, "full"/true 4:4:4 information is not available.
...Which is not to say that 4:4:4 _never_ have merits. I understand that green-screening have benefits from 4:4:4. In that case, one is "encoding" specific and critical encoding into the color difference channels, and using nonlinear editing that serves to make chroma subsampling more visible. Colored, sharp text (end credits) on low-resolution DVD can have annoying 4:2:0 artifacts.
If you look at the e.g. 100:1 "throwing away of information" that lossy compression, 8bit/gamma, 4:2:0 etc does to a raw signal, one might be surprised that our digital images resemble the scene at all. Then compare the massive loss of information that any sensor/lens does to a scene. I guess the take-away point is that bits & bytes are poor indicators of perceived quality.
-h
-
For camera sensors that have a Bayer CFA at the same pixel count as that of the luma channel, "full"/true 4:4:4 information is not available.
.
As I mentioned earlier in this thread that however tempting it is to do, it's really incorrect to link video encoding terminology like "444" to the ratio of pixels in a bayer sensor and to talk of it never having "444" worth of information.
While the ratio in both examples seem to correlate and seems to indicate otherwise the truth is that it's much more complex than simply looking at the ratio of pixels.
Yes. Bayer sensors have a differing ratio of Colour pixel ratios but you never describe them using video encoding terminology. You're only perpetuating an incorrect understanding of how these numbers affect the end result.
JB
-
As I mentioned earlier in this thread that however tempting it is to do, it's really incorrect to link video encoding terminology like "444" to the ratio of pixels in a bayer sensor and to talk of it never having "444" worth of information.
I believe that you are wrong. Of course a Bayer CFA is not a "4:2:0" format, but there is some value to comparing the two.
While the ratio in both examples seem to correlate and seems to indicate otherwise the truth is that it's much more complex than simply looking at the ratio of pixels.
Yes. Bayer sensors have a differing ratio of Colour pixel ratios but you never describe them using video encoding terminology.
I describe sensors using sensor terminology and video standards using video standard terminology. Whenever I think that it makes sense to compare those two, I attempt to use appropriate terminology.
A Bayer sensor of 1920x1080 sensels will have approximately 2 million samples.
A 1920x1080 4:4:4 image frame will have approximately 6 million samples.
Intuition and information theory tells us that a 2 million sample sensor cannot generally feed a 6 million sample file with independent information; there will be some redundancy.
If you have a scene where all spatial information happens in the "blue" channel, the effective resolution of the Bayer CFA is 960x540. Clearly, this does unjustice to a 1920x1080 4:4:4 format where the Cb channel is sampled at 1920x1080.
Scenes tends to behave in certain ways that are beneficial to the Bayer CFA. Humans tends to perceive visuals in certain ways. That is where things get complex. The maths of standardised color 3x3 matrixing and spatial up/downsampling is not that complex, and there is no reason to make it into something complex.
You're only perpetuating an incorrect understanding of how these numbers affect the end result.
JB
I am sorry that you see things that way.
-h
-
None of the cameras ever match with all the settings being the same.
What I don't get is why, especially with raw cameras, one cannot shoot a gregtag, or other brand card and have the software build a profile for each camera - colour by numbers totally exact.
One click, perfect match*.
Sinar had this in stills in 2005.
S
*obviously angle change and the like may make the shots not match but that is a different story.
-
What I don't get is why, especially with raw cameras, one cannot shoot a gregtag, or other brand card and have the software build a profile for each camera - colour by numbers totally exact.
One click, perfect match*.
Sinar had this in stills in 2005.
You calibrate a monitor to match outputs, it seems daft to not match inputs particularly when you have to match shots from [expensive] multiple cameras.
-
I believe that you are wrong. Of course a Bayer CFA is not a "4:2:0" format, but there is some value to comparing the two.
So we're in agreement that using "chroma sub-sampling" terminology is not the correct language to use when describing a bayer sensor performance. You're wanting to "compare them" using the wrong terminology.
http://en.wikipedia.org/wiki/Chroma_subsampling
(above link has no mention at all of Bayer sensors)
http://en.wikipedia.org/wiki/Bayer_filter
(Above link has no mention of 4:4:4 or 4:2:2)
I'm not really disagreeing with the point you're actually trying to make about differences in the way chroma is "captured" but using chroma sub-sampling terminology is the wrong way to make your point, which is I believe, that there are less blue and red pixels relative to green in a given bayer sensor.
I presume that you know what those actual numbers mean in that ratio. 4:4:4 ? Or 4:2:2 ?
They refer to the way the video is "encoded". The first number is the brightness information and is also the "green" channel. The second two numbers are the "colour difference" signal. So they contain no brightness information, only colour information. You "re-create" the colour gamut by summing and subtracting these two signals from the Lumininace / green channel.
This video encoding is also known as YUV or YCbCr or component.
http://en.wikipedia.org/wiki/YUV
This is why it's described as "encoded" video. It's NOT the same as RGB. Apple only recently created an actual ProRes RBG codec option, called ProRes 4444 which offers the user the choice of RBG or traditional YUV encoding.
A bayer sensor doesn't work that way. At all. There is no "encoding" into this colourspace until the sensor data is transcoded to a "video" format. There is not separate brightness channel and colour difference channels. A sophisticated algorithm "creates" The colour information and although there are differences in the sampling by virtue of the fact that there are generally twice as many green pixels as any other, it's only co-incidence that it falls into the same ratio as 4:2:2 chroma sub-sampling.
I'm being pedantic about this because your next leap of logic is really a gross simplification and in my opinion gives a misleading idea about Bayer sensors having "half" the chroma resolution.
A Bayer sensor of 1920x1080 sensels will have approximately 2 million samples.
A 1920x1080 4:4:4 image frame will have approximately 6 million samples.
I see it differently.
At 1920 x 1080 there are a fixed number of pixels. And this is the point. You're throwing pixel resolution into a discussion about colour fidelity and making a direct leap to the ratio of RGB pixels and equating that to "colour" resolution.
The inference of the sensor only having "960x540" worth of blue pixels doesn't make sense because you never ever only have the blue channel, because nothing we shoot is ever so highly monochromatic. Even BLUE LED's would have a range of "blue" And for that matter, even the "blue" pixels have a huge overlap of sensitivity to the other colours. It's not so highly monochromatic. If it was, then we'd have a very odd looking image indeed.
It also discounts the "mathematics" of debayering. You only have to look at the visual differences between ACR, C1 and even Resolve to see that those algorithms can make a large difference to the way the image presents in post.
So while I would absolutely accept there *IS* a difference in the chroma resolution, you can't use video encoding terminology as it over simplifies what's actually going on and leads to these kinds of numbers being kicked around and it leads to silly conclusions about cameras not being good enough for chroma keying for example.
Nobody ever talks about bit depth and I think that has a bigger vector on captured colour gamut than the sample size of the bayer sensor. In this case the BMPCC is greater than 16 bit at the sensor, 16 bit linear internally before going to 12bit LOG when recording to DNG or 10 Bit ProRes (422 HQ). In fact they unpack as 16 bit files again in Resolve or ACR once you start to work from the DNG's.
Yes there is a case for oversampling with Bayer sensors, and that's exactly the thinking behind a camera like the Sony F65 having an 8K sensor for 4K raw files. Can you name for me another RAW camera that oversamples in this way to address this "problem" of reduced chroma resolution ? The Panavision Genesis (Sony) is the only other one I can think of, though it's not really a RAW camera.
RED oversample and Arri oversample somewhat but not enough to deliver 4K cinema files that are flawed by the logic you're presenting.
jb
-
What I don't get is why, especially with raw cameras, one cannot shoot a gregtag, or other brand card and have the software build a profile for each camera - colour by numbers totally exact.
In the end you build a LUT that means they match, but usually they aren't matching because they are shooting different things.....No LUT will solve that....
jb
-
You pay (in some way) to build a lut, I dont get why you would not want the processed simplified/automated. Some stuff still photgraphers do and have done for years to speed their workflow might just be useful in the motion world
Also read the second half of my post which covers multi angle.
-
Back to the original question. BMC vs 5drawhack.
Would I not be correct in thinking that all the raw hacks line skip - which will simply make the image more artifaced than the BMC.
I saw an example of some telephone lines that has turned into a nice flight of steps with a hacked 5d.
In terms of usability the BMC scores with SDi lead to go to your monitor, mini HDMI is basically unusable.
I guess if ML has a centre crop 1080 mode with no line skipping then the result could be very good but the crop would render most canon lenses too telephoto and the mirror box restrict other lens choices at the wide end
Raw on a bubget? Id pick a BMC over a hack eight days of the week.
S
-
Morgan.
A LUT is used to create a LOOK as well as neutralizing colour imbalance. Mostly we shoot LOG so some kind of "grade" has to be applied anyway.
In resolve a whole look / LUT can be applied nearly automatically with a click of the mouse across hours of rushes.
This way transcoded rushes for editorial will have a non destructive "look" applied. For example I create a day for night LUT for when I need to fake night when shooting day.
JB
-
Yes I get that. What I am saying is to me there are two steps - getting your camera(s) 'normalised' - making grey grey, or more complexly normalising it to a gregtag card (which is automatic in some stills softwares) and then applying an artistic look to the footage
If you add a single LUT to two slightly mismatched cameras they wont match - but if you normalise the two cams and then add the LUT they will match
So we would have layers/nodes/whatever you want to call them..
greybalance/exposure (camera tab)
normalise lut (makes the colours fit a gregtag card or other)
artistic (could be kodak warm, day for night, whatever)
Having the normalise node/layer helps matching (multiple) cameras and prepares a camera(s) look for the application of an artistic look - which could have been prepared using a third camera at a remote location ie in pre or close to the colourist
..or the look can come 'out of a box' - something that we probably all emotionally dislike, but to a shooter, maybe doing weddings with a large throughput and minimal post budget a very efficient system, one that can be presold to the client - but a boxed look wont work unless working from a normalised input.
I think in stills 'normalise-add look' is the standard workflow, and the normalisation usually a simple click grey, or more complex calibration for product photographers, is entirely autmated or one click unlike Davinci.. and yes it is all 'pastable' to entire sets of images not done clip by clip.
-
Yes I get that. What I am saying is to me there are two steps - getting your camera(s) 'normalised' - making grey grey, or more complexly normalising it to a gregtag card (which is automatic in some stills softwares) and then applying an artistic look to the footage
It's "normal" to have LUT's for each camera.
Narrative motion grading workflow is a more complex process.
Shoot on set....Right now I'm shooting ProRes Alexa, (4444) ProRes BMD and DNG, plus RED epic occasionally....
Offload and copy to three separate drives.
Transcode to editorial format (in this case AVID DNx). At this point each camera has it's own LUT and specific LOOKS, which are easily applied scene by scene or even shot by shot (depending on what's required). In pre I shoot a chart in a couple of lighting scenarios and we strike the day-to-day LUT's from that. Again, these aren't usually neutral, but usually have a bit of a look depending on what the show's going for. These are to give the editors kind of normal-ish looking shots to cut with.
Editorial Cuts.
We go back and CONFORM to the camera original files. We end up with a mixed timeline in Resolve where R3D's, ProRes and DNG's sit alongside each other. None of them matching due to being different cameras, different lenses and different photographic positions.
The colourist then "grades" them together to help smooth out the edits...
At that point you're ignoring the LUT's and LOOKS and starting again with the "raw" material. The original LUT's that are applied are an approximate grade for editorial. And in storytelling, as it is with stills, the CORRECT WB and profile is rarely the one that looks the best. Mostly any mismatch in the cameras disappears in the to creative balancing of the grade anyway. There's no need to be "fussy" in a way, because the colourist will start from scratch anyway. They usually have your LUT there if they want to switch it on, but LUT's are a bit of a cheat and to get the most they'll start from scratch on each shot generally.
A commercial hour of TV in Australia is 46 mins. We generally have 16-20 hours to grade 600-800 shots in a timeline. Some shots have dynamic grades. Most have a few nodes and are simple. Some have windows and multiple layers to help them along. That's considered pretty fast. I've spent 12 hours grading a 30 second commercial. A 90 min feature film might have 3-4 weeks.
Unlike stills, shots in motion that cut against each other in the edit have differing demands in terms of colour correction. They need to "match" in order to maintain the suspension of disbelief. In story telling terms we might also be intercutting between different looks in a flashback or a another place that's in a different timezone. They have to cut against each other and usually what's "right" doesn't look best.
Lighting also changes dynamically within a shot. In a single shot an actor walks from darkness to a window, lifts the blind to look out the past the dirty curtains which then illuminates the BLUE wall near him before he then moves over to the cabinet where they turn on a dodgy fluro light So where do you put the grey card ? Which light source do you balance to ?
I actually think WB is a fools errand. Like in the old days on film I shoot at two WB settings. 5600K or 3200K. Every now and then I will go outside of those numbers but I balance my lighting to the WB of the camera, and not the camera to the lighting. If you chase WB on every setup you'll go mad and you'd be there all day shooting frigging grey cards which in the end don't save you any time in post and only cost you time on set.
The only other time I vary this is to maybe take the green out of some very heavy ND's so I'll swing the Magenta / Cyan in the camera, but manually though adjusting by eye.
Otherwise I never touch the WB and all I ever hear from colourists is how consistent my shots are to grade. Chasing WB (like ETTR) just leads to inconstancies that take longer to balance out. Yeah you might get individually better shots chasing those techniques, but in the scheme of things it just makes the workflow more difficult.
jb
-
Cinematographers tend to think, insultingly, that photographers are people who take a snap of a sunset and then twonk with that single frame in post for hours.
Do you not think a professional photographer has to create consistent looks across a hotel brochure or fashion brand look book? Not a fleeting cut, but images that must sit on the page for the client and his wife to evaluate at their leisure..(after it is back from the printer with a whole new set of communication issues)
.. and get some sort of look on set for their clients too if shooting tethered?
Do you not think they have to do that to time and to budget?
Do you not think that doing raw commercially for a decade longer than 99% of cinematographers we might have learned something?
I can tell you our "unify+add look" colour workflow is faster, better, cheaper and more consistent because we have more capable software that automates and unifies some of the process - to not be interested in learning about that is purely blinkered.
Sounds to me like your colourist is basically working 'blind'. Costly and doubtless a route for miscommunication. Adequate for drama and art but useless for fashion or product where maroon is 'so 1995' and 'dark red is so 2013'
Can I do this? Well some of the people on this board can, and I do a reasonable job at my level and resource set.
The only wall we don't have to cross is tracking shots.
IMO :)
-
Morgan...
I'm not trying to get into some kind of pissing match...
I'm just explaining my workflow, just like you're explaining yours and we're way OT for this thread. I'm aware I'm talking to mostly photographers here.
jb
-
Yes I get that. What I am saying is to me there are two steps - getting your camera(s) 'normalised' - making grey grey, or more complexly normalising it to a gregtag card (which is automatic in some stills softwares) and then applying an artistic look to the footage
If you add a single LUT to two slightly mismatched cameras they wont match - but if you normalise the two cams and then add the LUT they will match
Middle grey for every sensor is different. Normalizing it to a 'standard' grey value is flying blind, because you're trying to get the sensor to normalize to something it doesn't know. And you don't know the exact middle grey for a sensor because the manufacturer hasn't shared his coefficients with you.
If you shoot white and expose bang on in spot mode, then you'll get the middle grey the camera was calibrated to. It looks different on every camera. It is interesting to see how IRE is used in the video industry to place skin tones, blacks and whites. With RAW and sRGB monitors, all that goes out the window.
I'm not sure you can get two disparate sensors to match in color. If you get the reds to match, the blues and greens will be off, etc. You can't get perfection in video. Too many problems - poor color space, bayer sensor, motion, changes and variations in lighting and color especially in daylight scenes, artificial lighting on locations (especially fluroscents), codecs, bit rates, chroma subsampling, all kinds of crazy sampling, and poor quality of display units (which is getting better).
The correct way, in my opinion, is the 3D LUT, which pulls all colors together for organic color changes. It is such a simple (to use) method, which anyone can create in a few hours (and years of experience behind him/her). The Waveform and Vectorscope keeps you in check.
Is this slow?
You pay (in some way) to build a lut, I dont get why you would not want the processed simplified/automated. Some stuff still photgraphers do and have done for years to speed their workflow might just be useful in the motion world
Well, in photography, there's almost always only one person doing everything. In filmmaking, there are many. I would assume the cinematographer and/or DIT will take ownership of all aspects of the image, color being paramount.
Sidney Lumet says in his book 'Making Movies' how he was horrified of labs developing the negatives in the wrong way, or a color timer making the movie look like it wasn't meant to look. I wonder how the Godfather would have looked if someone used a standard curve to lift the shadows...
Photographers get into using software and then become slaves to it. The problem for them is that there's nobody to take them out of their habits. In filmmaking, many individuals constantly battle your cherished beliefs, and compromises and discoveries constantly get made. I don't think I would ever want to lose that.
It even applies to colorists. Photographers don't have anybody looking over their shoulders, while colorists have to please many people, and work fast at the same time. There is a school of thought that LUTs are bad for colorists, since it restricts their art. But the DPs vision is final, and that vision in encapsulated in a LUT.
To answer the OP, the Pocket camera is miles better, simply because it's a better workflow. I think 99.9% of those who use it would be very happy with just Prores. I was blown away by it. All the issues have been fixed, and it simply delivers great video right out of the box with Prores in Film mode. If there is something I don't like, it's the noise in the shadows, even in raw (studying the DNG stills John had posted on twitter). I'm not sure ETTR is good enough to fix it, but I'll need to see more footage to know for sure.
Can't wait for the 4K version to come out.
-
Im talking about normalising to a gregtag card (or other more appropriate tool) - sure one can geek out about sensor differences, but two raw stills cams even with a simple click grey (missing in resolve) will get a lot closer than not doing that - instantly with no skill/time. Sinar Capture Shop does not grey balance it matches a colour card with a single click.
It is theoretically possible in video from raw to a crappy codec - just the latter may break
As for no one looking over the shoulder of photographers - well if you shoot professional shoots you can have all sorts of heavy people looking over your shoulder right down a long chain from onset tethered capture to final print delivery
And yes IMO the BMC is better than a hacked DSLR for a million reasons :)
-
Sareesh,
If you know a photographer that makes a living at this and doesn't have someone looking over their shoulder in creative treatment, pre production, production, post production and delivery, then I'd like to know who they work for because that's the gig we'd all like.
In my experience if there is money on the table, there is somebody across from you counting it out.
When I work with colorists, I try not to spend a lot of time in the booth as I think it's usually counterproductive if you know where you want a shot to go.
I take a series of stills from the footage, run them through either DiVinci and/or Photoshop, make prints, with an instruction sheet of where I want the scene to go and then spend about an hour going over the footage and maybe another hour in room, but after that I usually go. The colorists I work with can run virtually real time online and I can run a monitor in the studio while I work and stay in touch that way.
Morgan,
I'm not sure your ever going to find an easy system (raw or cooked) where you can exactly match every camera, especially if their different camera models. Even our Scarlet next to the two RED 1's with the same RED color and RED gamma, the settings as close as possible, don't exactly match, in fact it's scene dependent but sometimes they don't get very close to matching out of the camera and when click balanced still don't match up without a lot of hand work.
IMO
BC
-
Sareesh,
If you know a photographer that makes a living at this and doesn't have someone looking over their shoulder in creative treatment, pre production, production, post production and delivery, then I'd like to know who they work for because that's the gig we'd all like.
Maybe I should have stated it better. I was not referring to clients or customers actually, but to professionals who will challenge your technical fundamentals - directors, DPs and DITs usually clash when it comes to these things. A well-informed client might have a good eye for the winning image, but will never be able to tell you how to pull it off with a LUT or a software setting or a camera setting.
Even in post, editors challenge directors, colorists challenge directors (and DPs), and so on. Reality checks come on various levels. The cool thing about film collaboration is that you have people challenging you while and even before you're working, so you can 'fix' it. What you don't know technically, you can't fix.
Of course, all this is assuming we all know are technical craft. Stanley Kubrick or Kurosawa would have kept their DPs on their toes, and a Roger Deakins will keep any director on his or her toes. Wasn't trying to comment on photographers at all in general, just making a point on the importance of DPs and DITs.
When I work with colorists, I try not to spend a lot of time in the booth as I think it's usually counterproductive if you know where you want a shot to go.
True, unless you have a personality that can hold back. You can't be overbearing, because that defeats the purpose of collaboration. Everyone finds a system based on what resources they've got and who they have to put up with.
I'm pretty sure you would stick around if it was a new colorist, wouldn't you?
-
I'm pretty sure you would stick around if it was a new colorist, wouldn't you?
On some of this I was kidding you, but like most things in life there is perception and reality.
A DP and operator that moved in next to me said he'd "like to get into still photography because it's just one person's art and the money is better".
Maybe he's right, but basically all of the photographic arts are collaborative. Actually the very best photographs I've ever produced I can't show because everything was so challanged going in, that it took every thing I knew to save it, though it was still challanged due to subject and circumstance.
But would I hang around a new colorist? I dunno, maybe it depends. If I go in prepared and have shot so the imagery can go a certain direction then hopefully it's not a huge leap to get where I want it to be.
If I didn't I guess I'd stay there the whole session, but today's world is much different than before, because the very first thing we say today is what we want, punctuated very strongly and quickly by "how much?".
We all know that given today's post production reach a lot can be done on the backend as long as the budget and the schedule permits it.
But all of this, whether your a still photographer a second unit camera operator or Roger Dekins is all down to personality and who you work with. If I work with A grade crew and a open minded client the project will always sing, but if we work with people with agendas that don't match (understand not wrong agendas, just not the same agendas), then it will suffer.
Last night I saw Sandra Bullock on a show and she said "nobody goes into a project to do a bad movie" and she's right, nobody starts out to make something that won't work. The thing is client's subject, stories, have to be flexible, same with colorists, or anyone in the technical arts.
A day that calls for sun and it rains is still a rainy day and either requires coming back, or rewriting for rain. To ask anyone to make it a sunny day just won't work and that's where most things go sideways.
Unintended but well meaning anticipations without anyone talking reality never works.
So, I hope when I walk into the color suite my anticipations match not just the talent of the artist but the reality of what I'm handing over.
If I do that, then no I won't spend much time there.
IMO
BC
-
I'm not sure your ever going to find an easy system (raw or cooked) where you can exactly match every camera, especially if their different camera models. Even our Scarlet next to the two RED 1's with the same RED color and RED gamma, the settings as close as possible, don't exactly match, in fact it's scene dependent but sometimes they don't get very close to matching out of the camera and when click balanced still don't match up without a lot of hand work.
IMO
BC
You have to admit that a grey click gets the cameras in the ballpark - davinci does not even have that.
My old sinar (similar to a phase p25) would click to a gretag card malking all 24 colours number accurate - in fact I think it worked off the bigger chart too maybe 64 numbers colour accurate by number - one click .
Do that with two cameras and they "match" and are ready to add and artistic grade
Of course if one camera has plastic highlights or muddy darks then they won't actually match - but to me it seems the best start point..
I've basically retired that sinar but it still get the occasional outing in the studio for product and really I'm completely lost without that one click start point when photographing stuff that needs accurate colour. I seem to one of very few who rember how well that sinar workflow worked :)
-
On some of this I was kidding you, but like most things in life there is perception and reality.
Totally with you. Reality bites.
In the end, regardless of the kind of work we end up doing, I feel we're all making videos because it gives us, if only for a few days, the power to play god.
-
Cinematographers tend to think, insultingly, that photographers are people who take a snap of a sunset and then twonk with that single frame in post for hours.
Do you not think a professional photographer has to create consistent looks across a hotel brochure or fashion brand look book? Not a fleeting cut, but images that must sit on the page for the client and his wife to evaluate at their leisure..(after it is back from the printer with a whole new set of communication issues)
.. and get some sort of look on set for their clients too if shooting tethered?
Do you not think they have to do that to time and to budget?
Do you not think that doing raw commercially for a decade longer than 99% of cinematographers we might have learned something?
Cinematographers are also highly skilled at things right in camera as they do not have the luxury of fixing in post, so maybe stills photographers may have something to learn from them. Plus I'm not sure why you are having a go at John, he's simply posted useful stuff about how he works and not how you should work.
I can tell you our "unify+add look" colour workflow is faster, better, cheaper and more consistent because we have more capable software that automates and unifies some of the process - to not be interested in learning about that is purely blinkered.
Sounds to me like your colourist is basically working 'blind'. Costly and doubtless a route for miscommunication. Adequate for drama and art but useless for fashion or product where maroon is 'so 1995' and 'dark red is so 2013'
Can I do this? Well some of the people on this board can, and I do a reasonable job at my level and resource set.
The only wall we don't have to cross is tracking shots.
IMO :)
I agree that stills post processing tools are very good and having to process what is effectively jpegs is frustrating when one has come from a stills RAW workflow background, but the workflow in professional film productions is not something that got worked out a few weeks ago. It's been around as have most of the issues to be overcome since the colour film was first used as seeing and as seeing a badly matched shot is very rarely seen that indicates film/tv/ad makers do actually know what they are doing. And are not exactly working blind.
Though having said that a raw film workflow would certainly make life much easier.
-
Lighting also changes dynamically within a shot. In a single shot an actor walks from darkness to a window, lifts the blind to look out the past the dirty curtains which then illuminates the BLUE wall near him before he then moves over to the cabinet where they turn on a dodgy fluro light So where do you put the grey card ? Which light source do you balance to ?
I actually think WB is a fools errand. Like in the old days on film I shoot at two WB settings. 5600K or 3200K. Every now and then I will go outside of those numbers but I balance my lighting to the WB of the camera, and not the camera to the lighting. If you chase WB on every setup you'll go mad and you'd be there all day shooting frigging grey cards which in the end don't save you any time in post and only cost you time on set.
The only other time I vary this is to maybe take the green out of some very heavy ND's so I'll swing the Magenta / Cyan in the camera, but manually though adjusting by eye.
Otherwise I never touch the WB and all I ever hear from colourists is how consistent my shots are to grade. Chasing WB (like ETTR) just leads to inconstancies that take longer to balance out. Yeah you might get individually better shots chasing those techniques, but in the scheme of things it just makes the workflow more difficult.
I recall when first starting working on film one cameraman I worked with did exactly what you did, when everyone else I'd worked with was using a white/grey card to get a perfectly neutral white balance. Now there are two issues with WBing the heck out of a scene.
1. As you said where do you WB as it can will vary across a shot. A lot.
2. More importantly you lose the room tone of the location and scenes won't look right/good.
-
You have to admit that a grey click gets the cameras in the ballpark - davinci does not even have that.
My old sinar (similar to a phase p25) would click to a gretag card malking all 24 colours number accurate - in fact I think it worked off the bigger chart too maybe 64 numbers colour accurate by number - one click .
Do that with two cameras and they "match" and are ready to add and artistic grade
There's an action/script you can run in Photoshop to work out an ACR calibration for any camera. I used it in early iterations of Lightroom as ACR struggled to differentiate oranges and reds in Canons. There are scripts for various colour charts to be found here (http://www.rags-int-inc.com/PhotoTechStuff/ColorCalibration/). So if you want to use PS or LR to grade shots then this may be of use.
-
Jjj
I acknowledge that movie people (paticularly john, as a shhoter,who is a hero shooter to me!) get a nice result - often with vans of gear and troops of people.
I seek a good result without those resources. My imaging journey started with an fm2 and a roll of tmax followed by deep post by me in the darkroom milking the most from a negative. That simplicity stopped for a decade 1998 to 2005 as I worked with shitty digital stills cams slowly I employed trucks and troops to make my stuff nice. The d3 and raw in 2009 brought back the simple shoot joy I had in my early 20s. The d3 and raw processing really enabled me to strip back my stills shoots to my early style.
I now seek simplicity on set (followed bu deep post) with shooting motion and witness cine people suggesting that the trucks and troops is "how it must be done" - I feel that with deeper exploration of the post of raw shooting movie cameras is does not need to be like that (if course it still can be) - I will take on any cinematographer who sticks to those guns however much I like them and their work. I just do not beleive that most cinamatogtaphrts are that smooth with raw post which we have been doing in the main for a decade longer than them. Stills and motion have merged we have a lot to learn from video people from sound to..,well everything, and pretty much the only thing we confidently bring to the party is good skills with raw!
Thanks for the scripts but I don't want to post process motion in Lr or ps - I'd like to use davinci which is awesome apart from a missing "start point" - I cannot see how wanting a start point is in any way controversial .. And davinci is developed though feedback from users of which I guess John (as a bm "ambassador") is an influential such user.. I'm just trying to get davinci made as good as it could be through public debate.. :)
-
.. A start point doesn't have to be used hard.. it can be used intelligently to create "room tone" for example I love leaving my tungsten s yellow the cam on 5.5k and just adding a little 5.5k kicker to paint a picture.
-
Morgan,
You can white click in cinex if you shoot with RED, though I find it rarely works for continuity. I agree with John that it is preferable to shoot a series of scenes with a locked in color temperature, usually 3200 for tungsten and practical tungsten, 5800 for daylight but that depends on the camera and scene.
But if you want to click on a grey or white card Apple's 3 way color corrector has a dropper for shadows, midtones and white balance and good auto settings for all three ranges.
Proress 422 HQ holds up quite well in converting and we use the 3 way color corrector for basing out the footage prior to the edit.
I also agree with John that ambience color bounce effects a scene.
I just edited a scene where the subjects walk from a daylight room into an area that is mostly tungsten and I balanced out two clips, the overlayed the adjusted tungsten clip over the adjusted daylight clip and then changed the opacity in a ramp style of transition so the correction from daylight blue to tungsten orange was subtle, but everybody has a different way of working.
One software I've just began to use is RED Giants colorista II. I find it's excellent for finishing out once the edit is locked, but soon as we move to 4k editing, I'll probably have to change that.
Not to go off topic, but I assume that DiVinci 10 is the beginning of making a color suite a full featured NLE, so some of this is probably mute in a few years. I think DiVinci will be the bridge between editors that work in FCP 7 and need a faster 64 bit editorial suite, though are resistant to fcp X. (I think I fall in this category).
IMO
BC
-
I know there are a million subtleties to each scene or job - from clinical accuracy to wild "art" and a million tricks like grey balance maybe outdoors or close to a key light but then shoot in a different location but overall I think an accurate start point is important. And the smaller your budget the more you hit odd lights and have to work with them. The bigger your budget the more you can just do what you want with lights..
As for suite - I'm a huge fan of resolve in every aspect but this
The power windows tracking layer opacity make it shine above all other suites IMO
-
I know there are a million subtleties to each scene or job - from clinical accuracy to wild "art" and a million tricks like grey balance maybe outdoors or close to a key light but then shoot in a different location but overall I think an accurate start point is important. And the smaller your budget the more you hit odd lights and have to work with them. The bigger your budget the more you can just do what you want with lights..
Or you could string together four Arri L7-Cs and have 'raw' light as well. Paint the scene in real time. Cost to you? $10,000, a van and a gym membership....maybe the occasional visit to the chiropractor as well.
-
I made a simple comment suggesting a raw software should have a grey balance button and now I'm sposed to be investing $10g in lights?
-
I made a simple comment suggesting a raw software should have a grey balance button and now I'm sposed to be investing $10g in lights?
Just kidding!
-
I made a simple comment suggesting a raw software should have a grey balance button and now I'm sposed to be investing $10g in lights?
Come on Morgan don't be cheap.
You should look at 30 grand in lights, because you never know what will happen and we always need backups.
IMO
BC
-
Thanks for the scripts but I don't want to post process motion in Lr or ps - I'd like to use davinci which is awesome apart from a missing "start point"
A film maker colleague recently tried LR instead of DaVinci and was amazed at how much easier it was to use in comparison.
I cannot see how wanting a start point is in any way controversial
It isn't. Being rude to a helpful poster is however well...rude.
-
Lr and c1 are certainly simpler to get a start point and single look with a file but without any motion tracking I cannot see how they are really any use at all for processing motion. I'm firmly of the belief that resolve is by far the best affordable motion colouring suite.
It is never my intention to be rude especially to those whose work I deeply respect. I also however expect some respect for the stills industry where people like coots are clearly better and more experienced with colour than most of Hollywood put together. I d like to promote mutual learning and exchange between all those passionate about the creation if good images.,
-
So we're in agreement that using "chroma sub-sampling" terminology is not the correct language to use when describing a bayer sensor performance. You're wanting to "compare them" using the wrong terminology.
Well, I never used "chroma sub-sampling" as a terminology to describe a bayer sensor. Seems like you are argueing against a straw-man.
I'm not really disagreeing with the point you're actually trying to make about differences in the way chroma is "captured" but using chroma sub-sampling terminology is the wrong way to make your point, which is I believe, that there are less blue and red pixels relative to green in a given bayer sensor.
The fact is that our camera sensors are (usually) Bayer, and our video storage format is (usually) YCbCr with or without chroma subsampling. Comparing those two _is_ relevant, no matter how this hurts your philosophical view.
People will ask themselves things like:
*Will I loose anything by converting a (Bayer originated) native 1080p stream to 4:2:2 or 4:2:0?
*Will I loose anything significant?
*My target/editing occurs at 1080p 4:2:2, and color transitions are utmost important to me. What sensor resolution would be a (necessary) component in delivering this?
Answering those questions can be hard, and will usually include some "ifs" and "depending". Some idealized ballpark measures can still be handy, though.
I'm being pedantic
Frankly, it seems to me that you have just learned something new and wants to boast about it.
about this because your next leap of logic is really a gross simplification and in my opinion gives a misleading idea about Bayer sensors having "half" the chroma resolution.
I see it differently.
At 1920 x 1080 there are a fixed number of pixels. And this is the point. You're throwing pixel resolution into a discussion about colour fidelity and making a direct leap to the ratio of RGB pixels and equating that to "colour" resolution.
The inference of the sensor only having "960x540" worth of blue pixels doesn't make sense because you never ever only have the blue channel, because nothing we shoot is ever so highly monochromatic. Even BLUE LED's would have a range of "blue" And for that matter, even the "blue" pixels have a huge overlap of sensitivity to the other colours. It's not so highly monochromatic. If it was, then we'd have a very odd looking image indeed.
Talking about single-colorchannel variation is an extreme case that can enlighten us on the more general behaviour.
So while I would absolutely accept there *IS* a difference in the chroma resolution, you can't use video encoding terminology as it over simplifies what's actually going on and leads to these kinds of numbers being kicked around and it leads to silly conclusions about cameras not being good enough for chroma keying for example.
I never claimed that any cameras were not good enough for chroma keying. In fact, I believe that all of my claims were fairly accurate, while you have been putting words in my mouth throughout this discussion.
Yes there is a case for oversampling with Bayer sensors, and that's exactly the thinking behind a camera like the Sony F65 having an 8K sensor for 4K raw files. Can you name for me another RAW camera that oversamples in this way to address this "problem" of reduced chroma resolution ?
If anything, I would dare to claim that this "problem of reduced chroma resolution" seems to be overrated*) by many videophiles and enthusiasts. If there are few cameras that oversample (specifically) to overcome this, it could be a reflection that it really does not matter that much for most material. There are other good, technical reasons to oversample that would be outside the scope of our discussion.
I have some knowledge about the technology and theory behind video. I have never touched any of the cameras that you mention, and if you gave one to me I would know nothing about how to produce interesting videos with it.
-h
*)Like stated in my first post: I acknowledge that there are or might be situations where 4:2:0 is a real problem, just as there are situations where the Bayer CFA is a real problem. It is not apparent to me that this occurs very often, and when it occurs, it might be that jumping 1080p 4:2:2->4k 4:2:2 is a better option than jumping 1080p 4:2:2 -> 1080p 4:4:4.
-
Well, I never used "chroma sub-sampling" as a terminology to describe a bayer sensor.
I think you do though when you make statements like this...
For camera sensors that have a Bayer CFA at the same pixel count as that of the luma channel, "full"/true 4:4:4 information is not available.
The fact is that our camera sensors are (usually) Bayer, and our video storage format is (usually) YCbCr with or without chroma subsampling.
My point actually is that YCbCr IS chroma subsampling full stop, even when it's 4:4:4. Even if the ratio is 1:1:1 because it's no longer RAW sensor data. It's "encoded" video. In fact it's VIDEO. The bayer data isn't video until it's gets transcoded into VIDEO, and it's at that point one chooses the choma subsampling used, even if it's 4:4:4.
It's no longer RGB.
It's no longer Raw Bayer sensor.
The terminology is important.
Choma-Subsampled video = Encoded video
All of the variants, be they 4:4:4 4:2:2, 4:2:0, 4:1:1, 3:1:1 or 22:11:11 are all ways of describing encoded video.
None of those terms are ever used to describe raw bayer data.
People will ask themselves things like:
*Will I loose anything by converting a (Bayer originated) native 1080p stream to 4:2:2 or 4:2:0?
*Will I loose anything significant?
*My target/editing occurs at 1080p 4:2:2, and color transitions are utmost important to me. What sensor resolution would be a (necessary) component in delivering this?
Which are all fine to ask. But to use "encoded video" terminology to compare to a bayer sensor's RBG pixel ratio is just misleading and a simplification.
I never claimed that any cameras were not good enough for chroma keying.
...Which is not to say that 4:4:4 _never_ have merits. I understand that green-screening have benefits from 4:4:4.
And again, this is my point. Using 4:4:4 terminology and associating it with "benefiting" chroma keying means that 4:2:2, the next qualitative metric down from 4:4:4 is somehow inferior.
Of course it is in Keying 4:4:4 will be an advantage, but you can't quantify bayer sensor data as only having "4:2:2" worth of chroma resolution, the next step down from 4:4:4.
If there are few cameras that oversample (specifically) to overcome this, it could be a reflection that it really does not matter that much for most material.
I agree. The cameras I mentioned are at the very top price bracket for any camera, and in the end it doesn't matter very much at all.
But, there is a large difference to me between 4:2:2 and 4:4:4 video in terms of end result of colour information.
That is my very point at taking you up on the use of this terminology.
There is very little difference to me though in cameras that have oversampled bayer sensors so that I can then get what your own mathematics demand as being of "true" resolution.
4:2:2 Vs 4:4:4 = a big difference in end result. It really does have "half" the chroma resolution.
8K sensor sampled for 4K --> 4:4:4 video Vs 4K camera sampled to 4:4:4 video makes much much much less of a difference in terms of chroma resolution. It's certainly not half which is what your comparison is saying.
By your "comparison" there is half the chroma resolution on the 4K sensor for 4:4:4 video compared to using an 8K sensor but the end result doesn't actually end up that way.
It is not apparent to me that this occurs very often, and when it occurs, it might be that jumping 1080p 4:2:2->4k 4:2:2 is a better option than jumping 1080p 4:2:2 -> 1080p 4:4:4.
Agreed, in general assuming all other metrics are equal. There are very few cameras that can actually do this though and bit depth, compression codec factor more heavily most of the time.
jb
-
Lr and c1 are certainly simpler to get a start point and single look with a file but without any motion tracking I cannot see how they are really any use at all for processing motion. I'm firmly of the belief that resolve is by far the best affordable motion colouring suite.
One doesn't always need motion tracking. Just like one doesn't always need a magic wand from Photoshop to select an area to change with LR's different way of working. Sometimes LR is better than DaVinci, just like sometimes a Canon is a better tool than a Hasselblad. Horses for course and all that. :)
It is never my intention to be rude especially to those whose work I deeply respect. I also however expect some respect for the stills industry where people like coots are clearly better and more experienced with colour than most of Hollywood put together. I d like to promote mutual learning and exchange between all those passionate about the creation if good images.
I certainly agree the raw workflow is superior to a non-raw workflow and find it frustrating to effectively go back in time when grading video footage. But I would hesitate to claim photographers are necessarily better at colouring/grading than Hollywood colourists because photographers have better [in some respects] tools than them. Heck, they've been at it longer than us. :)
It seemed like you were having a go at John because of arguments you may had had elsewhere with other people, I didn't notice anything where he put down photographers. He simply explained a workflow used by a crew of professional film makers.
-
I think you do though when you make statements like this...
What I was trying to explain was that using 6 megasamples to store 2 megasamples will always be a redundant way of storing things*). Equivalently, if you want a 6 megasample file to contain true (unpredictable) information, then you need a sensor of at least 6 megasamples (realistically more). Thus, when you store your 1080p Bayer image as 1080p 4:4:4, you are not getting the information that this format (ideally) can convey. Which can be a bad thing or might not matter.
I never wrote that a 1080p Bayer sensor inherently _is_ 4:2:0, or that it maps 1:1 to the information of a 4:2:0 stream.
*)some caveats that I will add at a later time if the interest is really there
My point actually is that YCbCr IS chroma subsampling full stop, even when it's 4:4:4. Even if the ratio is 1:1:1 because it's no longer RAW sensor data.
That may be your usage of chroma subsampling, but it does not correspond with how the rest of the world use that word.
Would you say that a 16-bit developed still-image *.tiff file is "chroma subsampled because it is no longer RAW"? I don't think anyone would do, except possibly a few people who mix up Bayer CFA with CbCr sampling...
It's "encoded" video. In fact it's VIDEO. The bayer data isn't video until it's gets transcoded into VIDEO, and it's at that point one chooses the choma subsampling used, even if it's 4:4:4.
Are you the kind of guy who protest loudly anytime anyone talk about a digital image, claiming that "it is only bits & bytes, it is not an image until it is printed"?
I never claimed that any cameras were not good enough for chroma keying.
I understand that green-screening have benefits from 4:4:4.
And again, this is my point. Using 4:4:4 terminology and associating it with "benefiting" chroma keying means that 4:2:2, the next qualitative metric down from 4:4:4 is somehow inferior.
Both of your quotes from me seems to be accurate and true.
If I was to design a chroma keying algorithm and had the choice between 4:4:4 input, or the same input converted to 4:2:2, I would prefer 4:4:4. For some kinds of sources it might not matter much, but it would never hurt (except bandwidth and processing cost).
That is my very point at taking you up on the use of this terminology.
I am sorry, but I think that you are taking up the wrong guy, for the wrong reasons with little effect other than wasted bandwidth.
-h
-
I am sorry, but I think that you are taking up the wrong guy, for the wrong reasons with little effect other than wasted bandwidth.
Yup certainly 'taking up the wrong guy'. hjulinessen don't like facts that conflict with his slightly strange world view, so best avoided.
-
One doesn't always need motion tracking. Just like one doesn't always need a magic wand from Photoshop to select an area to change with LR's different way of working. Sometimes LR is better than DaVinci, just like sometimes a Canon is a better tool than a Hasselblad. Horses for course and all that. :)
I certainly agree the raw workflow is superior to a non-raw workflow and find it frustrating to effectively go back in time when grading video footage. But I would hesitate to claim photographers are necessarily better at colouring/grading than Hollywood colourists because photographers have better [in some respects] tools than them. Heck, they've been at it longer than us. :)
It seemed like you were having a go at John because of arguments you may had had elsewhere with other people, I didn't notice anything where he put down photographers. He simply explained a workflow used by a crew of professional film makers.
Now two total threads :) If find nearly every secondary needs motion tracking, also I like the big scopes, nodes and many other bits of Resolve.. but motion tracking is the big one vs a stills software. After a lot of experiments Resolve is now my choice to deliver colour on time and on budget.
A workflow by professional film makers. Yes everyone is welcome to use their system especially those who make it work wonderfully. But that does not make it right or the correct solution for all. Especially when one considers the implications even of a 'monitor' that is sensitive enough to dial out a filter green cast 'on set' - basically that is a tethered monitor with AC supply or at least a multi k onboard monitor - (you cant see shit on a $1k SmallHD in terms of colour - but coming from raw stills I only see on set colour as a guide and the smallHD is good enough for that). Or 'lighting to a certain temperature' that is wonderful if you have a van of HMI, but not so good for solo efforts.
As for the experience of grading - well most corporate video looks horrid, most hollywood is graded by professional specialists, neither path really puts the DP in the driving seat in that way that smaller stills houses work which for me for example means I control the look from lighting and exposure choice to delivered file on 90% of my jobs for 15years. Also not all work is dramatic, imaging work may go from copying art works and product to any level of dramatic contrivance. The suggestions I make a more towards doing the first end of the scale on time and on budget.
Fundamentally I think we all need/want to become faster/more flexible better or cheaper.. to get more from our budget and I think adding those tools to Resolve would aid that, is the right thing, is the 'normal' thing in raw grading, and of course any user can choose to leave them untouched.
-
As for the experience of grading - well most corporate video looks horrid, most hollywood is graded by professional specialists, .......snip......I make a more towards doing the first end of the scale on time and on budget.
Fundamentally I think we all need/want to become faster/more flexible better or cheaper.. to get more from our budget and I think adding those tools to Resolve would aid that, is the right thing, is the 'normal' thing in raw grading, and of course any user can choose to leave them untouched.
Morgan on some of this I agree, but I think where some of this falls down is there is a constant comparison of "Hollywoood" production, vs what most of us do.
If Hollywood wants an effected, blue tone look, with great deep skintones, they build a set or change a practical location to match, wardrobe for the look and color, test various stocks of digital settings, test through the colorists and then go forward.
I see it all the time in our work. A client wants a cool desaturated look, though the location has brown walls and yellow trim, the wardrobe is bright primary and that will never color as they anticipate.
On a scene like that you can color, mask, matte, select and effect all you want but a brown scene becomes very global when brought down to deep un saturated blue. You can get close to the look, but it's never going to have the exact refinement or look of a "hollywood" project that is planned for the look way before the cameras start rolling.
If you've seen the movie "Rush" which is shot by a very good dp Don Mandel and in Hollywood terms, shot on a "low budget", $38 million (which didn't allow for any over the line expenses), to achieve the look the DP wanted he had teams of colorists working on site and didn't move off a shot until he was sure he got the look he wanted.
The Movie Gravity took 4 let me repeat this Four years to complete, so in my view any of us comparing our work to budgets that range in the 100's of millions is not an apples to apples comparison.
Few of us have that luxury and the common thought is shoot what you got, then we'll try to fix it later, which means post work is double the time of the actual pre production and on set time.
I do agree that Di Vinci is good, (though I don't think it's that great) maybe 10 will be more intuitive, though if we had a lightroom interface that worked with layers, had some tracking and keying and used sliders instead of those silly wheels, our (and even professional colorists) would find it a godsend.
I personally like coloring on the timeline, because I can play and replay and see that my continuous look is true to the story and though suites like Colorista II are not exactly an easy slam dunk, I find once we have a locked edit and are ready for finish, staying on the time line gives me a more cohesive look, even if round tripping is linear, because round tripping rarely gives you the sound, the soundtrack, the effects and the titles that rest on the timeline.
In video, or motion, or digital cinema (take your pick of the terms) we still have one foot in the new world and one foot in the past and it's time to step up and find a way to go forward faster and easier.
It's funny, I have a friend and supplier that's a top drawer effects artist. At his fingertip he has used Flint, Fire, Flame, Smoke, every expensive effects suite ever made and when we talk about his projects and what he shows on his reel, I ask what was that done in and he routinely says uh . . . photoshop . . . uh After Effects . . . uh photoshop again . . .
But watching recent films and expensive television production, I see budget and time (same thing) effecting the look more and more. I just saw a Hollywood movie last night and the reverse shots of the actors had such a different coloration and look, it seemed like they were shot on different sets with different crews and I assume the dp and director wanted a better match, but I bet budget stepped in and said ok, that's good enough, lock it.
IMO
BC
-
After Effects gets used a lot for things like the HUD in Iron Man (http://provideocoalition.com/adobe/story/upping-the-hud-ante-with-adobe-after-effects) or green screen work like in Meet The Fockers (http://rebelsguide.com/forum/viewtopic.php?t=503) - although AE isn't mentioned in this post, Hype specifically mentions using it it lots of features elsewhere then links to this example of his work.
-
IMO
BC
Well if I was looking for examples of colour on a 'budget' (ie not building the whole thing from scratch with an art team) I would point to work such as your stills before I would point to many 'videographers' work :)
-
That may be your usage of chroma subsampling, but it does not correspond with how the rest of the world use that word.
Would you say that a 16-bit developed still-image *.tiff file is "chroma subsampled because it is no longer RAW"?
To obtain a Y'CbCr stream you must first sub-sample, always. What you sub-sample are split into Luma and Chrominance information, which forms part of the core specifications of a stream that can be properly labelled Y'CbCr.
John never mentioned a TIFF file, he specifically said "My point actually is that YCbCr IS chroma subsampling full stop..." In this respect he is correct. Scientifically you could sample for anything, even bacteria on the sensors...but in the video world, Y'CbCr is always in reference to sub-sampling of chrominance.
Using 4:4:4 terminology and associating it with "benefiting" chroma keying means that 4:2:2, the next qualitative metric down from 4:4:4 is somehow inferior.
It is, because you're sampling at half the frequency. Once you've digitized a sampled feed there is no going back. Sampling is always destructive, by definition.
However, the beauty of the solution is that you are managing to throw away 1/3rd of the data while retaining 100% perceptual color information. An algorithm that tries to extrapolate color information from a 4:2:2 signal will always struggle because it cannot be expected to know the conditions of the sampling (trade secrets, and even mistakes).
I have extensively tested 4:2:2 against 4:2:0 and RGB on many top-class keyers (Primatte, Ultramatte, IBK, alpha, etc.) and I must say that 4:2:2 is great for most tough keys. What I've found is that the 'impossible' keys also struggle at 4:4:4, and you must use multiple methods to pull something out. The results are similar if you had used 4:2:2.
If I was to design a chroma keying algorithm and had the choice between 4:4:4 input, or the same input converted to 4:2:2, I would prefer 4:4:4. For some kinds of sources it might not matter much, but it would never hurt (except bandwidth and processing cost).
In fact, you should prefer neither. When working with data it is always preferable to have the most 'virgin' feed possible. In the case of Bayer sensors, that feed is raw+a supreme understanding of the specific CFA+sensor+lens used. However, no camera manufacturer will divulge many of the details necessary for a third-party algorithm (not to say anything about patent restrictions and licensing). Therefore, if I were to design an algorithm, I would prefer RGB data any time. Without RGB data, you're always a slave to the manufacturer's proprietary RAW processor (Resolve, Redcine-X, Arri, Sony viewer, and so on). Not that that's a bad thing!
Any subsampling, even 4:4:4, is not preferable to RGB, which is preferable to RAW for unscientific reasons!
It's funny, I have a friend and supplier that's a top drawer effects artist. At his fingertip he has used Flint, Fire, Flame, Smoke, every expensive effects suite ever made and when we talk about his projects and what he shows on his reel, I ask what was that done in and he routinely says uh . . . photoshop . . . uh After Effects . . . uh photoshop again . . .
It can edit, it can grade (as good as Resolve and Speedgrade) and it can pull of VFX (as well as Nuke or Shake) like any other app on the planet. Where the Autodesk suite had its advantage was in working with 3D models and 3D space. AE now has that too. It has a tracker, superb masking (better than power windows), rotoscoping and keying controls. There's nothing that AE can't do...maybe except audio. In quality, nothing is better if not worse. And, its encoding engine is probably the finest I've seen for mastering.
To stay with the topic in this thread, where AE lags behind Resolve is that Resolve is a better RAW processor for Blackmagic cameras. Also, Resolve 10 beta works and feels much better than Resolve 9 (I only use the Lite version). In the next one week, Adobe is expected to release their CinemaDNG update to CC, so things might change.
-
To obtain a Y'CbCr stream you must first sub-sample, always. What you sub-sample are split into Luma and Chrominance information, which forms part of the core specifications of a stream that can be properly labelled Y'CbCr.
"Chroma subsampling is the practice of encoding images by implementing less resolution for chroma information than for luma information, taking advantage of the human visual system's lower acuity for color differences than for luminance."
http://en.wikipedia.org/wiki/Chroma_subsampling
The conversion of some R'G'B' format to some Y'CbCr format consists of:
1. Apply pre-shifting
2. Apply a 3x3 linear (invertible) transform
3. Apply post-shift
(http://www.poynton.com/notes/colour_and_gamma/YCbCr_RGBprime.gif)
4. Subsample chroma to something like 4:2:2, 4:2:0, or leave it at 4:4:4
(http://content.answcdn.com/main/content/img/CDE/YCBCR422.GIF)
5. Quantize and clip
steps 1-3 are invertible (lossless), steps 4-5 are not. In an actual design, the actual processing may deviate some, and the precision may be reduced, but the general black-box behaviour should be aligned with this description.
In fact, you should prefer neither.
That is a nonsensical answer. Given the choice of ice cream and cookies you prefer chocolate? No meaningful discussion can be based on that logic.
-k
-
steps 1-3 are invertible (lossless), steps 4-5 are not. In an actual design, the actual processing may deviate some, and the precision may be reduced, but the general black-box behaviour should be aligned with this description.That is a nonsensical answer. Given the choice of ice cream and cookies you prefer chocolate?
You are confusing digitally obtained data as lossless. 0 = 1, so to revert 1 = 0. To obtain this data you must first sample, and then sub-sample. There are many analog sampling stages before you even begin to see data.
Digitization is lossy by definition.
-
"Chroma subsampling is the practice of encoding images ...."
The conversion of some R'G'B' format to some Y'CbCr format.....
The point being that you're converting.
So why use terminology from what you're converting to when you're talking about what you're converting from ?
jb
-
You are confusing digitally obtained data as lossless. 0 = 1, so to revert 1 = 0. To obtain this data you must first sample, and then sub-sample. There are many analog sampling stages before you even begin to see data.
Digitization is lossy by definition.
You are confusing subsampling with sampling. Every digital signal is sampled (and thus carries some error compared to some analog source). That is irrelevant to this discussion imho.
The act of converting (digital) R'G'B' to (digital) Y'CbCr is not digitization. No ADC is involved. No subsampling is involved unless you choose to (spatially) subsample the chroma channels to e.g. 4:2:2 or 4:2:0.
There is (usually) some loss in converting R'G'B' to Y'CbCr 4:4:4, but this loss is not what an engineer would call "subsampling", it is quantization and clipping. I routinely do such conversions in MATLAB that is lossless (due to using floating-point numbers).
For further background, I recommend Charles Poyntons book "Digital video and HDTV: Algorithms and interfaces", p.p 90, "Chroma subsampling".
-h
-
You are confusing subsampling with sampling. Every digital signal is sampled (and thus carries some error compared to some analog source). That is irrelevant to this discussion imho.
The act of converting (digital) R'G'B' to (digital) Y'CbCr is not digitization. No ADC is involved. No subsampling is involved unless you choose to (spatially) subsample the chroma channels to e.g. 4:2:2 or 4:2:0.
Well I guess the camera manufacturers told you when they do what, in what sequence, and how.
There is (usually) some loss in converting R'G'B' to Y'CbCr 4:4:4, but this loss is not what an engineer would call "subsampling", it is quantization and clipping. I routinely do such conversions in MATLAB that is lossless (due to using floating-point numbers).
What is the definition and process of quantization?
Anyway, this argument is pointless. So I'm going to stop.
-
Well I guess the camera manufacturers told you when they do what, in what sequence, and how.
No, but the corresponding ITU documents as well as the reference litterature gives a pretty good description of _what_ should be done. If you want to claim that some manufacturer does it very differently, I think it is reasonable to expect you to support that claim.
What is the definition and process of quantization?
http://en.wikipedia.org/wiki/Quantization_(signal_processing)
My attempt (at describing basic quantisation):
For some real number R (e.g. '4/3'), represent it by the nearest integer (i.e. '1').
Anyway, this argument is pointless. So I'm going to stop.
Yes, I have found this branch of the discussion to contribute little to the discussion. I guess I have a hard time not responding whenever I see some claims that are (in my view) totally wrong. I shall work on my "let it go"-skills.
-h