Color Correcting Canon 7D Footage
Tuesday, January 26, 2010 at 12:35AM 
A frequent concern about shooting to a heavily-compressed digital format—something the DV Rebel often finds herself doing—is the degree to which the footage will be “color correctable.” Will the shots fall apart when subjected to software color grading? Or will you be able to work with the footage as fluidly as you tweak your raw stills in Lightroom?
It’s a valid concern. The movies that the current crop of HDSLRs shoot are highly compressed. This compression is perceptual, meaning that it takes advantage of visually similar colors and shapes, and represents those regions with less accuracy than the detailed and varied parts of the image. This makes perfect sense, but often in color grading one seeks to enhance color contrasts—to make a face pop off a similarly-colored background for example—and so you may well create high contrasts between colors that were once nearly identical, and as such were given short shrift by the camera’s compression.
You might have noticed a similar phenomenon in audio. An low-bit-rate MP3 that sounds decent enough can suddently sound awful after even a tiny amount of EQ. Another case of perceptual compression limiting your options.
While you will never find as much data and detail in your HDSLR video as you do in that same camera’s raw stills, the H.264 movies created by the Canon 7D, 5D and 1D Mark IV will withstand some massaging in post. Here are some tips (similar to those found in greater detail in The DV Rebel’s Guide) to help you get the best results.
- Shoot flat. If you read Flatten your 5D, you know that I am a proponent of setting up a “flat” Picture Style using the camera’s built-in controls. The same settings I specced out for the 5D Mark II apply to the 7D and 1D Mark IV as well, although with the 7D I’m less likely to use Highlight Tone Priority, as this setting can increase shadow noise, and the 7D is not as noise-free as the other Canon HDSLRs.
- Chose WB wisely. Use a white balance preset that gives you as nuetral an image as possible. Shooting with an incorrect white balance reduces your dynamic range, because you wind up with an image that’s prematurely blown-out in one color channel, dark and noisy in others.
- Expose to the right. Make the brightest image you can without clipping something important. A rule-of-thumb considered gospel by many photographers, but our reasoning is a bit different. Yes, we, like the stills guys, wish to avoid excess noise in the shadows, but that’s not our main concern. Remember that term perceptual compression. Dark areas of an image get less bits. If you underexpose, you’ll have to brighten the image in color correction, and you’ll reveal all kinds of nastiness the camera thought you’d never see.
- Do denoise. It doesn’t really matter what denoising software you use, but use it. When you carefully and subtly denoise your footage, you rebuild your pixels anew, which is especially nice when you follow the next tip:
- Work at high bit-depths. If you start with an 8-bit image and do a gentle de-noise, you’re blending pixels values together to create new colors. Although there’s no such thing as something for nothing, doing this at a higher bit-depth means those new colors have massivly more gradations than the original image. Your subsequent color work will hold up much better.
- Sharpen last. Your flat Picture Style removed the camera’s built-in sharpening. Add your own at the very last step. The amount you use will vary depending on the output medium, so test test test.
By folllowing these guidelines you can make good-looking shots even better with color correction. But what about a shot that isn’t so great to start with? Turns out there’s hope. Below is a 7D shot that I grabbed in an uncontrolled situation. In my haste, I underexposed, and used the “cloudy” white balance when I probably should have used tungsten. But with a little denoising, careful analysis of the colors in the image, and a Colorista Power Mask, I was able to rescue this shot.
Yes, you can color correct your HDSLR footage, and you should. Color correction can make a good shot great, and in a pinch, put an unusable shot back in the game.
Reader Comments (50)
Hurray for this article! There has been a lot of ballyhoo (as usual) about whether or not to use "Superflat" gamma curves and so on, an even more controversy over Eugenia Loli's recent post about her own custom Extraflat curve.
I brought this up with a cinematographer friend who still shoots film primarily, and he argued that the problem with artificial superflat curves in one of these cameras (that isnt shooting RAW etc) is that they aren't extending the dynamic range at all nor are they preserving details in any way - they are only doubling up on certain portions of the image.
Regardless of that argument, I have been following the guidelines in DV Rebel's Guide, your Red Giant video on color and Chris Andersen's stuff in the AFX books and that was with HDV. Now that I am working with the 7D (where I believe I have effectively 9 stops range as opposed to the HV20's 4 stops) I wondered just what all of that translated to in post.
This came right on time! I hope this comment spurs further discussion because I want to learn more about optimizing the video coming out of this baby. Sorry I cant even afford to rent a RED for one day (currently a package goes for about $500 here in Hwood) thus either I could rent ye olde RED for four days or buy a 7D!
>whether or not to use "Superflat" gamma curves and so on
Like Stu, I'm a proponent of the flat shoot look. It helps you color grade with much ease, let's you take one look and make it another one, in a much easier/nicer way than saturated/contrasty footage is able to. It scales better.
Last week, I shot a live performance in San Francisco, where there was very little light. Thankfully, the 5D MkII was able to work well in that low light compared to my SX200 IS 720/30p HD digicam that I also had with me as a backup. I had set the camera to completely flat, on the "Neutral" preset (contrast, saturation, sharpness, all the way down). When I got home that night and checked the footage, I still found it over-contrasty compared to unprocessed RED footage. I could lower the contrast, and still recover detail on Vegas.
But here's the catch. If I shoot as I did, and change the contrast in post:
1. I don't get all the possible detail as I would if I was shooting with superflat or extraflat picture styles.
2. Lowering the contrast down, revealed terrible h.264 artifacts in the dark areas.
Since I like the film look, which is traditionally low contrast and low saturation, it's much better to shoot using superflat/extraflat and then grade, rather than shoot contrasty, and then having to change that in post and end up having artifacts.
It's possible that when you shoot flat you get more sensor noise in dark areas, but honestly, between sensor noise and the alternative (h.264 artifacts in post after artificially lowering contrast), I much prefer sensor noise. I see it as the lesser of the two evils.
>is that they aren't extending the dynamic range at all nor are they preserving details in any way
I don't believe this personally. If the picture style is used by the camera at an early stage of the internal process, you do get something good out of it.
As I wrote on my blog post yesterday, if the Curves dialog on Picture Style Editor get updated to let us change its origin point, and have that new ability loaded into the camera, then, we should be able to get some additional shadow detail from the 5D. However, even as it is now, shooting flat, does a good-enough job.
Then again, you might not be after the film look, or wanting to color grade aggressively, so in that case, shooting in plain "neutral" color setting might just be acceptable for some people. Especially those who don't color grade at all. It all depends what you're after, I guess.
Eugenia, I most certainly am after the film look - whatever I shoot (mostly shorts) are intended for submission to festival screenings so quality is the standard.
Stu, I notice some serious aliasing/artifacting the second example image. Is this the product of underexposure? What were the settings for the second shot (ie. was it 1920x1080 / 23.97D / superflat curve of some sort?)
I bought a 7D after reading about it here and over at videocopilot.net. Coming from a photography background, it's good to get behind an SLR again. Shooting with the 7D is a lot like working with a film camera.
Thanks Stu for all of the information. I read the DV Rebel's Handbook twice over Xmas holiday. Good stuff.
Why can't I see the images?!?
But by de-noising and blending pixels together, you are essentially losing resolution, correct? Hence, why you probably say CAREFULLY and SUBTLY denoise.
Hi Stu, I have asked before in the 1d post... what are you using to denoise? Mac OSX compatible?
Max
Stu, as well as Max's question of what Denoise, I wonder what tools you are using for sharpening?
I've had simply amazing results with Neat Video's de-noise plugin. I emailed the developer and he gave me a FCP and AE bundle deal. Use it lightly and it really doesn't do too much softening.
~Dietrich
what about vfx on this footage? curious if you did any tests to see how h264 holds up...?
Beyond proper exposure and white balance, does Neoscene help? It is said that this software converts h264 "raw" from 7D to a more editing-friendly avi, with 10-bit depth and chroma sample 4:2:2. In theory color grading with this specifications is more effective, am I wrong?
I didn't test it, my 7D arrives next mouth (sadness). Thanks for the post, Stu.
KMS, please view the images at their full size (half HD) before evaluating the aliasing—they are scaled down by the browser, which could introduce additional artifacts. We all know that aliasing is an issue with HDSLRs, but the 2nd shot doesn't contain egregious artifacts to my eye. It definitely does suffer from the underexposure though.
I agree with your cinematographer friend to some degree. It's hard to know at which point one's "super flat" curve is getting the most out of the camera, when it's just de-contrasting what's already there. It is certainly possible to take a flat Picture Style so far that it becomes counterproductive to subsequent grading. I can't say that my settings are scientifically proven to ride this line perfectly, but they work for me and I can set them up in the field without the use of a laptop, so I like 'em.
Paul, it's true that any denoising must strike a balance between removing unwanted noise and reducing image detail. Some denoise techniques result in overly-clean, plasticy images. A little bit of noise is a very good thing for moving images.
Max, in this case I used Neat Video, but I am still testing various options.
Carl, I use a combination of After Effects's Sharpen and Unsharp Mask filters.
D, the biggest challenge with VFX on HDLSR footage is the rolling shutter. But I've had good luck doing planar-tracking based effects using Mocha AE (which ships with AE CS4). There are minor effects in NOCTURNE, for example. To get into 3D tracking, one would probably need to start by using The Foundry's Rolling Shutter plug-in.
Cris, simply converting the footage to a new codec won't help in and of itself, but if doing so means that your subsequent work will be done at a higher bit-depth (as is the case in most NLEs that match their sequence settings to the codec), then yes, it will help. Also, NeoScene may do some chroma up-sampling when converting, but I'm not sure about that. I use Magic Bullet Deartifactor (comes with Magic Bullet Frames, part of the Magic Bullet Suite) to up-sample the chroma in the uncompressed, 16-bpc world of After Effects.
Great Article Stu,
I use Neat Video extensively and have found it to be the best option with the most control out there, especially if you manually adjust and play with the sample numbers above 1. I also have the Magic Bullet De Noise that came with the Steady plugin but it is quite aggressive and obliterates a lot of detail.
Often I find that de-noising 7D footage gets rid of many macro-blocking artifacts but has a detrimental effect on large color areas like skies and or underwater footage which show segmented banding.
In that case the only real help is to ad a grain component or monochromatic noise to the stack in order to minimize it. Sometimes it looks better to have a less clean image but mileage may vary. So maybe adding subtle noise or grain after CC could be an added point in your article by occluding artifacts.
BTW I am using your flat settings from the 5D instead of curve profiles for all my footage and so far has worked extremely well. Thank you for that!
One other tip I have found is that in 720p 60 mode, the red color channel seems to be the most aliased of them all, using a 1 pixel red channel blur can save a slow mo shot from the delete bin since it minimizes the aliasing a lot.
Here's a question:
when using the white balance settings like cloudy, sunny, tungsten and the like. Do these lock the white balance? Or does it fluctuate between or during shots around a K value close to what would be suitable for the lighting?? I have a disease where i always hash the white balance and being able to use those preset settings would be awesome but i always thought they fluctuated.
They are locked. I'm a big fan of WB presets for exactly that reason.
So question for anyone with the experience: what's the difference between color grading Canon 7D footage converted to Uncompressed 4:2:2 10-bit and 7D footage converted to ProRes 4:2:2 HQ? Is it substantial?
I use Neoscene and find I can push the image hard in post. Neoscene does chroma upsample/smooth chroma noise and the result is a 10-bit 4:2:2 file which I work with Vegas in 32-bit with 2.2 gamma. Using Magic Bullet Looks, I'm very happy with how much I can CC even just using Neutral settings in camera.
I am getting weary of the greens/blues/ orange skin thing.. was watching some 1982 films.. and the color is refreshing.... smooth, bold, warm, and still colorful in the blues and greens... I am finding that grading for the Bruckheimer makes me feel like a wannabe colorist - which of course... I am.
Stu, have you tried the NewBlue FX Rolling Shutter plugin? It's in the Video Essentials 3 pack. I just bought it yesterday, but haven't had any heavily skewed footage to test it on yet. Apparently you need to select an edge, and it will work from that.
As for noise reduction, I'm a big fan of Neat Video also. I use it on pretty much everything these days. Even footage that already looks great can benefit, especially in the shadows. Below are some before and after shots from a clip I did last year where I used Neat Video. The fluorescent location lighting was horrible, so there's not much I could do to improve the colours. The low light performance of the HV20 at night with a 35mm adapter attached didn't help either.
1: Original Footage.
2: Colour correction + Levels.
3: Colour correction + Levels + Neat Video.
I would think that something that works on edges can only correct overall skew such as from a whip-pan. The Foundry's plug-in works by calculating the motion of every pixel in the frame.
Actually, it probably doesn't just work on one edge. The manual says to find a frame with visible skew (a vertical edge I guess) and to calibrate by adjusting the amount of correction until it's straight. Then it says to use that same setting for all clips from the same camera at the same frame rate.
I just tried it on some HV20 footage. I found a clip with a fast whip pan, but the skew on that was minimal, and the pan was so fast it was all blurred anyway. The best setting appeared to be 2. So then I tried the same setting on some other footage and there was a slight difference. The footage looked tighter. Hard to explain, but it does seem to work.
Glenn - the real test would be wobble. We shot a zero budget feature last summer with a bag full of HV20s and our car mount footage suffered much wobble. Do you have any of that you could try?
You are right on the money with the sharpening. I didn't believe it first but then I tried it on footage and it made it much better. However I am still puzzled over the 8bit-16bit flow. Isn't the raw footage already 8 bit how can you get 16 bits out of it. And I know AE has option even to work with 32bit but is there some option in Premiere. I could find any options to change bits in Premiere. I don't like edit in AE cause of the sound and it's slow preview compared to actual editing software. Could you elaborate on the 16bit work-flow. Thanks
David, there's a detailed explanation in The DV Rebel's Guide of why working at 16-bit is of great value, even when you start with 8-bit footage. In short, as soon as you move the colors around, you're better off having more room to do so. And operations like denoising and color noise reduction actually create new pixel values at the new bit-depth. You don't get something for nothing, but you give yourself more room to play with what you have.
For Vegas users if you like the teal look it can be achived with two plugins that come with Vegas:
color corrector:
Low: Angle=286.5 Magnitude=0.437
Mid: Angle=305.5 Magnitude=0.133
High: Angle=128.7 Magnitude=0.397
Saturation=0.944
Gain=1.102
Gamma=1.314
Offset=-14.2
Cookie Cutter:
Circle
Feather=1, Border=0, RepeatX=1, RepeatY=1, Size=0
Stephen, actually, I don't think I have anything that bad. If I do come across anything that bad though, I'll give it a try. My HV20 is attached to an SGPro 35mm adapter on rods most of the time, so that keeps it pretty stable.
Where does the choice of YCrCb or RGB fit into all this?
Do I convert from mpeg2 or mpeg4 to RGB image sequences or lossless codec ? DV Rebel Style and use a compositor like AE for the majority of work? 32bit Linear even for videos with only minor compositing of titles, blurs, etc and work with proxies for realtime playback.
Or do I rely on the NLE / Cineform codec, which appears to be YCrCb or RGB, to handle the conversions back and forth when applying effects, blurs etc and trust the NLE to do it at a higher bit depth.
Do Cineform users use RGB?
It's all rather confusing.
is anyone having issues shooting PAL 24P on the 1DMKIV? I find the image extremely jittery with very sharp multiple images when panning or having someone move a lot in shot.
just playing the native .mov files at this stage so any advise on what i am doing wrong or having done yet to get good clean 24P footage would be deeply appreciated!
~A~
This is a great article Stu. But one thing I have always wondered every time I see an "expose to the right" suggestion is how does one go about exposing to the right when shooting a low key lit scene or even chiaroscuro?
Doesn't seem like it would work as most the frame would be shadows and the histogram would surely show exposed to the left or?
Or are you supposed to shoot it bright/high key and get the low key look in color correction? Never really understood that but if exposing to the right gets the best results I would like to do that. I just don't know how when shooting low key lit scenes. Thanks in advance for your explanation.
Thanks James. Yes, in that case you'll never get a big pile of histo over on the right, but you can still place the brightest values in your image as fer to the right as possible without clipping. That's what would constitute "exposing to the right" in that case.
Thanks Stu. That makes sense. By the way, do you have any rule of thumb to expose skin using the histogram? What I normally do if shooting in a controlled situation is to turn all lights off but the key light and make sure my skin falls within the 3rd zone in the histogram which is in the middle. Do you think this is underexposing?
Hi,
I have posted this in the forums, but as I got some information from this post, I thought I would also post it here in case somebody came across this first...
I wondered if anybody could offer their thoughts on a workflow I have for my graduation film shot on the 7D. It's essentially going to get finished to HDCAM SR primarily, but I have been offered a test print to 35mm of a short section, so wanted to make sure that I haven't missed anything stupid. The company have said that they will apply a LUT to the rec709 corrected footage, so to not 'bake in' any LUT on our export for them. Anyway, this is how things are being done at present.
- It was obviously shot in h264 in the 7D, and all raw footage was transcoded to ProRes422HQ before any other work was done.
- The ProRes files were imported into FCP6 and synced to the external audio recorded on set.
- All editing was done inside FCP, and when complete a ProRes422HQ video will be exported (which I believe is 10bit 4:2:2)
- This will then be imported into AE CS4, and the footage colour space interpreted as sRGB (is this right, as I don't know whether to use the camera or ProRes as source for this setting)
- In a 16bit project space, with working space of REC-709, I will carry out noise reduction on the footage using Neat Video Pro, both to remove a little video noise, and to smooth out a little of the artifacts before grading.
- This will then be exported out of AE's 16bit project, into a 10bit DPX image sequence (I don't think these are 16bit, are they??)
- The DPX sequence will be graded in Apple Color, and exported as the DPX (linear, I believe?) and possibly as video format for putting to tape HDCAM SR (ProRes or other suggestions are welcome, if accompanied by a little reasoning or info)
I hope this is clear enough as a starting point, and I think it's clear what I'm unsure of. Either way here are some clearer questions if necessary:
- Is a DPX sequence 16bit or 10bit?
- Should I relink to the original h264 raw files before exporting, or stick with the ProRes422HQ transcodes used to edit with?
- What is the best format to end up for laying back to HDCAM SR tape? (video or image sequence, etc.)
- What settings should I use in the DPX export from AE for importing into Apple Color?
- Are there any 'gotcha' type things hiding in amongst this?
- Does anybody have any alternative recommendations or advice?
Thanks a lot in advance, and I haven't got time to proof this right now, so I apologise for typos and bad grammar.
Regards,
--Ian Pinder
Hey Guys,
I've scoured the internet for an answer to this question but have been unsuccessful:
Does Magic Bullet Looks play nice in a 32bit work-flow?
I'm following the Adobe Whitepaper HDTV work-flow to the tee, yet when I open looks in my 32bit project in AECS4, it appears I'm being double gamma-ed!
I'm getting a nasty crushed image that doesn't correspond to what I see in after effects.
What am I missing? Do I just need to stick to Colorista?
Regards,
Jake Jelicich
Hi Jake,
Looks does indeed play nice with a 32-bit workflow. All of its processing is 32-bit.
But it does expect video input by default. You'll need to get a little advanced to make Looks work with linear-light images.
First, in LooksBuilder, go into Preferences and set Input/Output Processing to Advanced. Click OK.
Now you should see four tabs at the top of your screen: Input, Look, Preview, and Output.
Click on Input. Turn sRGB OFF, and enter a gamma of 1.0.
Now go to Preview, de-select the chain icon (Gang), and turn sRGB ON.
You now have a setup in Looks that takes linear in, gives linear out, but previews with sRGB gamma.
Thanks Stu! Really appreciate you taking the time to answer, means a lot.
I've read that you use the Neat Video plug-in, which I purchased, to find out its only 16-bit. Do you have a workaround for this? Or do you just remain in a 16-bit work-flow?
Regards,
Jake Jelicich
It's generally OK for NR to be done at 16-bit, since you're applying it to footage that doesn't have overbrights. If your footage does have modest overbrights from a SMPTE scale conversion, just use Levels to bring all the values into 0–1 before denoising.
Hi,
I just read my above post on workflow, and realised it was a bit of an essay on here. I just wondered whether Stu or anybody could offer any thoughts on the questions at the bottom, as I'm kind of trying to get a handle on some of the aspects of this project. If I clarify anything regarding, I will attempt to do so on request.
Thanks very much for any help in advance.
Regards,
--Ian
Hi Stu
Thanks alot for this great post. I just bought and read the dv rebel guide and your chapter about color correct dv footage. As I'm using a canon 5d mk ii to shot my projects I have the following question:
In the guide you suggest to remove dv artifacts at the beginning ot the color correction process. As I'm using a 5d I'm wondering if its also required to start with removeing dv artifacts with your after effects script or do I dont need to do that? Isnt it enough to just denoise the footage with Neat Video or any other denoising software?
What can you tell me about dv artifacts on 5D and/or denoise?
I would be great if you could help me with that!
Marcel
The artifact reduction method I describe in The Guide works very well on 5D footage.
Hi,
Thanks very much for taking the time to respond. I did read through the guide with this in mind, but didn't see too many details on first glances. I'll go back and refresh my memory on this section. Thanks again.
--Ian
Hi Stu
You've written you use Magic Bullet Deartifactor to upsample chroma to a 16 bit project in After Effects. When I'm trying to do that I get a warning the Deartifactor-Plugin only works with 8 bpc?! Does it also work with 16 bpc or should I use you Rebel Deartifactor script instead? Which one works better with 16 bpc. Magic Bullet or your Rebel Script?
Thanks for an answer
Peter
Sounds like you have an older version of the plug-in Peter. Like, maybe really old?
Stu,
No, I have Magic Bullet Frames V1.0 installed which I downloaded from redgiant some days ago. I cant download any newer version there.
Interesting, thanks Peter, I'll look into it!
Stu,
you say expose to the right with your explanation. Phillip Bloom says expose to the left one stop with explanation of more detail recovery rather than overexposing and blowing out image detail. You guys hang out....zacuto great shootout 2010. Have you guys talked about that. Which is ultimate? Seem to lean towards your end since your specialty is in the final stages before releasing a film. His specialty is in shooting. Can you offer thoughts?
I think we're both saying the same thing—expose as far as you can to the right without blowing out important highlight detail.
Well, I went back to double check to make sure I was not mistaken and if you have Phillip Bloom's 'How to shoot 7d' from F-Stop academy at 20:20 into tutorial he says always try to underexpose to retain image detail rather than overexpose as you will lose image detail. All of this may not be a big deal but if I can control it and if it is a matter of using a shot or not in post production than I would rather know which way to go. Can you give Phillip a call and get back to us? Your the best.
I have struggled with a few clips, particularly ones with clouds (stormy, grey clouds). I convert to prores, color correct etc. In the package of choice (whether that is FCP or After Effects) results look good. If I render to animation codec, all looks good, but any other codec I find various degress of banding between different shades of grey. I have seen posts regarding picture style, but not sure if there is anything else that can be done to reduce this problem. It seems that whilst I have access to the highest quality version, things look great, but as soon as it's rendered out, these artifacts appear.
Interesting stuff, Stu. I imagine a lot of these priniciples translate to the 7D's kid brother, the T2i (my camera of choice). Do you have a specific workflow that you follow in post-prod. leading up to the sharpening? I always find sharpening the hardest part of the "massaging", as it were. I never can hit the sweet spot between the correct definition and over-doing it on the image noise. Neat Video sounds like a good idea. How processor-intensive is the pro version?
Thanks again for the great explanation.
hey ho
you may have already done this, but if you have not.....please consider an in depth dslr post production tutorial (i'd pay) what you described in your article "color correcting 7d footage" is very important. for instance, walk me through the after capture process of ingesting, noise reduction, color correction using what you believe are the best tools, best practices.
thanks
paul hirschberger