Advice on 60D Workflow from Start to Finish

[Daniel B. Kim]

Hi,

I'm a longtime reader, but new poster on this forum and I'm also relatively new to the HDSLR scene and have been scouring the web trying to sharpen and improve my footage for a couple years now. Right now, on my 60D, I primarily use the Sigma 30mm f1/4 and the Samyang 85mm along with the Magic Lantern firmware.

The trouble I seem to be having is that I just can't get my footage to look as clean and sharp as I want to, and rather than dropping a few hundred to a thousand dollars on workshops to see where my problem could lie, I thought I could give the active readers of this forum a chance first. Please feel free to dissect my workflow and how I do things, I'm self-taught and really have been doing things from what I've read on the internet.

WORKFLOW:

- Shoot on my Canon 60D onto SanDisk UHS-1 95MB/s SD cards with Magic Lantern Firmware and Technicolor Cinestyle.
- Import footage via SD slot on my 27in iMac.
- Use 5DtoRGB to convert files to ProRes4444
- Bring converted footage into FCP7 timeline.
- Export using QuickTime conversion to H.264 container and highest settings or MPEG-4.

From experience, I think many of my problems occur in trying to capture the right settings while shooting and also during post-production. I am getting better at getting the right exposure levels, especially with the help of Magic Lantern, but it seems I still struggle with something, because when I bring the footage in, sometimes there's a greenish tint to it, and many other things happen to the image. As far as grading is concerned, when I grade I really just adjust the color correction to what looks the most accurate color-wise to the naked eye. I'm sure the professionals aren't doing this, especially because how crappy my image can become after I'm done with them. Any advice on how to improve my eye for color grading and grading more accurately?

When I watch footage shot by Phillip Bloom, or anyone else with more experience than I have really, I really can't understand how they can fit such a sharp image within the space constraints of Facebook or Youtube or Vimeo. Anyone know their secret to such a beautifully sharp image? I understand that they probably have equipment that dwarfs mine, but I've seen high bitrate 60D footage that is comparable. Could someone let me know?

I'm currently thinking about making the switch over to the Panasonic GH2 because of the footage I've seen with the hack, but I would rather stick with the 60D and figure out what I'm doing incorrectly with my workflow and fix what I need to fix before I make such an investment.

I understand that there are some purists out there who will reprimand me for attempting all of this on my own and attempting something out of my comfort zone, but I think that's the best way to learn. I'm on here because I'm frustrated from failing for years and would like any insight or help as to what I could be doing wrong and advice on what I could do moving forward. Sorry for the long post, hope to hear some helpful responses! Thank you!

[Murray Christian]

I don't think you're doing anything wrong, per se. Sometimes output settings can murk things up a touch. I seem to recall you can get different results, sometimes, exporting to compressed from your NLE versus exporting to intermediate and then letting Compressor or something else do the job. But I'm not sure.

The impression of sharpness can be a little bit subjective, particularly on the web. Much depends on how the footage was shot in the first pace. Using A grade lenses, exposing nicely and shooting to the camera's strengths are pretty important. Adding just the right amount of contrast back in post and that can make all the difference to how it looks, especially after youtube et al compression is done with it.
Those lenses are decent enough, so you should be ok there. Maybe if you had some examples of stuff you aren't happy with?

[Brian David Melnyk]

do you custom white balance when shooting and apply sharpening in post?

[Daniel B. Kim]

Murray and Brian, thank you both so much for your replies. Murray, you're right, I will definitely find something and post it up soon. I actually haven't tried exporting into Intermediate, and by Intermediate I'm assuming you mean the Apple Intermediate Codec, or are you referring to something else? Would the results vary drastically?

Brian, as far as white balance is concerned, I use the auto white balance with Magic Lantern, sometimes it's just quicker because it still takes me awhile to really get accurate white balance. I don't sharpen in post, I'm actually not really sure how. How would I go about doing that?

Thanks again for the replies, I can let out a slight sigh of relief now.

[Murray Christian]

By intermediate I mean any high bitrate lossless (may be redundant) codec (I would have thought ProRes was one such but wikipedia says it's actually lossy. Hmm. I haven't been on Apple for a while. Someone wiser will know if it matters much). I use the Ut codec, mostly 'cause it's free but most packages and systems come with something these days.
This way you've got a sort of master render you can check for quality and compress it various ways for delivery when you need to without having to fire up FCP and have it re-render everything each time and risk glitches etc.

I think Brian is talking about adding a digital sharpen filter to the clips. I used to think that was a bit taboo and a sign of poor photography (and I know a lot about poor photography). But I've come around. It's something about the way these cameras downsample internally, they can lose the contrast on some fine detail, particularly in certain subjects and movement I think (can't recall off hand). Used judiciously, a little touch of the sharpen filter can really help some shots. Enough so it'll keep those highlights and edges better as you downscale and recompress for the web etc. But you don't want to over do it.
It'll be under Effects->Video Filters

[Brian David Melnyk]

as i understand it, the image falls apart with DSLRs if extensively graded.
so the closer you get to color balance in-camera with a custom WB (as well as proper exposure), the less manipulation in post, and the better the image. i find the auto WB especially horrible if any artificial light is present... Daylight is ok, but always better to custom WB.

not sure about your delivery, but i experimented with the Technicolor profile and found that it required a lot more work in post, and for my needs, the old 'sharpness all the way down/contrast -4/saturation -2' setting was less work intensive and got me closer to the image i wanted in the first place.

i think the reason for not sharpening in-camera is to minimize the ill effects of moire etc. so the image is a little soft and needs sharpening to bring back the details. also, you have more control with a filter than in-camera, and i would imagine that the quality of sharpening is better (someone more informed could probably explain why... algorithms?) not sure which NLE you use, but i find the sharpening filter in FCP works really well, though it took me a lot of experimenting to get it right. but then you can save your best settings (ie: as '5D sharp') and apply to clips as a great starting point.

also, for a 'sharp' image, i am finding using a smallHD DP6 monitor for focusing is soooo much better!!!! nothing like coming back to the studio and finding out how many shots were unusable due to focusing issues to justify spending the cash...

my workflow always involves a 3 way CC (not just for color balance, but to make the image really 'pop') and sharpening as the last steps (and sometimes Neat noise reduction, but rendering is looooong). the difference between the 'flat' image and the final image is generally a noticeable improvement.

actually, i put the filters in the order of first sharpening, then CC, then maybe also a Broadcast safe. Not sure if there is a difference in IQ if the order changed to say CC after sharpening? anybody know if there is a 'proper' order for best quality, or does it make a difference?

[Jordan Hooper]

Quicktime can play the .mov files straight off the card. How does your footage look? If it doesn't look good at this stage then your problem is how you use the camera. Your footage at this point should be excellent.

If your footage looks good straight off the card, at what point in your workflow does it loose quality?

[Sareesh Sudhakaran]

Here's an alternative workflow. Try it out with a sample clip:

1. Instead of converting to Prores, edit native H.264.
2. Export to H.264 at higher than 44 Mbps.

[Nigel Barker]

You are over-complicating matters.

Forget about Cinestyle. It's noisy & difficult to get the exposure correct as the image on the LCD looks so pallid. Focusing is more difficult too.

5DtoRGB is a waste of time (literally) on Windows or on recent versions of OS X as it's only really useful for working around some gamma problems with older versions of QuickTime.

The easiest workflow is to use Premiere Pro & edit natively then for Vimeo export to H.264 MPEG4 at 5-10Mbps & add a sharpen filter.

The GH2 does natively have higher resolution but is like a tiny little plastic toy in your hands after a Canon DSLR (I own a GH2 & 5DII & 600d/T3i so can compare). The image doesn't look as nice as that from the Canon DSLRs as it has much more of a "video" look. The softness of the Canon image is actually an advantage as it looks more aesthetically appealing too.

[Daniel B. Kim]

@Murray: I'll definitely try Compressing lossless footage, sounds like it could work even a little bit. I'll definitely try the Sharpen tool also, definitely seen it within FCP but didn't want to touch, especially because I didn't know how it would adversely affect my footage. I'm working on figuring out how to display some footage of mine, but will definitely post some stuff tonight.

@Brian: Does that mean you don't use After Effects at all to do your grading? I wanted to try importing my footage into After Effects to do grading, but for convenience sake, I decided against it. I'm assuming Noise Reduction and everything else you mentioned can me found within FCP as well? I looked up the DP6 Monitor and it looks stellar! SDI out with great native resolution, but it's a bit out of my price range, is it worth the investment? Are there other monitors that do a great job for half the price? What is a Broadcast Safe?

@Jordan: I'll work on posting a few clips I shot recently, most of them I am not particularly happy with. Directly from the card, it doesn't look too too bad, noting like the Image Quality of other videos I am trying to emulate unfortunately. I would say the point where it lose the most quality would probably be after I export the video.

@Sareesh: Wouldn't editing natively take forever because of the render times? Are you implying I should use a different NLE system? How do I have that specific control over export bitrates? All the custom features I know is when I select 'Export' and choose which format and container to export to, there's a toggle bar that I just put to "Best" for Best Quality. Is there another way to customize the bitrates more granularly?

@Nigel: I did notice a surprising amount of noise for all the hooplah I was hearing about this specific picture style. I figured the pale picture was a result of the picture style doing its job of flattening out the image, no? I didn't mind it, gave it kind of a vintage feel. I guess my question for you would be the same as Sareesh's, in how I can export and set settings to specific bitrates. Is that possible in Adobe Premier? In FCP, I just put the setting to "Best" for "Best Quality". Have you tried your GH2 alongside the high bitrate hack? The footage I've seen from that hack are absolutely mind-blowing, especially at the GH2's current price point.

Thanks everyone for your responses. It seems like I can get a little bit more specific help once I post some videos here, I'll work on that tonight. But for now, maybe I can post this YouTube video of something I did awhile back when I had even less of an idea of what I was doing, but still came across the same problems. I know there is natural loss with YouTube compression, but I still feel that the image quality could be better. Let me know!

Perspectives on the World Christian Movement Class Promo - YouTube

[Daniel B. Kim]

Apologies for double posting, but my goal is to make my footage look something like this.

Canon 60D Magic Lantern & Technicolor Test Two - YouTube
Technicolor Cinestyle - Canon 60D - YouTube
Anyone know how they got such high quality videos into sizes that can fit on YouTube? My footage isn't as clean or high res as theirs and I still have alot of trouble shrinking my footage down to get it on YouTube.

[Sareesh Sudhakaran]

Originally Posted by Daniel B. Kim

@Sareesh: Wouldn't editing natively take forever because of the render times? Are you implying I should use a different NLE system? How do I have that specific control over export bitrates? All the custom features I know is when I select 'Export' and choose which format and container to export to, there's a toggle bar that I just put to "Best" for Best Quality. Is there another way to customize the bitrates more granularly?

On modern systems editing native is no longer an issue.

Regarding export - You could export to an uncompressed image sequence. Also, if you have AE installed, I highly recommend finishing from AE. Just port your project over (using Automatic Duck or whatever) and export from there. I have tested this workflow both on FCP and Premiere (I use the latter) and it is as robust as it gets for interframe codecs.

Assuming delivery for the web, here's my workflow:
1. Import native footage into Premiere and edit native - project settings correspond to footage, and NOT delivery format.
2. Once edit is locked, port the project to AE (No transcoding)
3. Finish in AE and export an uncompressed 8-bit image sequence in Rec. 709, all other settings correspond to the delivery format. This is the image master.
4. Mix audio and finish in Audition/Soundbooth/Premiere or other. Export as uncompressed WAV, each channel separate corresponding to delivery requirement. This is the audio master.
5. Use encoder (I use Adobe's via AE/Premiere) to combine image and audio. Apply Sharpening if necessary and export to delivery format. For best quality, you might have the option to pay for a professional account on some websites, and they might accept a higher quality master. They might also offer a higher bit rate for paying members.

Here's another tip: I don't recommend going higher than 1280x720 for the web. My tests have shown me that there is no perceptible difference between 1080 and 720 on average computer monitors showing highly compressed content. The 'space' you save can be better utilized in providing a higher bit rate and/or better frame rate for a perceptible resolution advantage. However, if you are also showing on HDTV monitors, then I recommend sticking to 1920x1080.

Unfortunately, I can't provide any FCP specific help, I'm afraid. Hope this helps.

[Nigel Barker]

Originally Posted by Daniel B. Kim

Apologies for double posting, but my goal is to make my footage look something like this.

Canon 60D Magic Lantern & Technicolor Test Two - YouTube

Technicolor Cinestyle - Canon 60D - YouTube

Anyone know how they got such high quality videos into sizes that can fit on YouTube? My footage isn't as clean or high res as theirs and I still have alot of trouble shrinking my footage down to get it on YouTube.

Those clips are in focus (generally) & correctly exposed in bright sunshine in one clip & decent lights in the other. Many recommend underexposing & then bringing it up in post whereas the correct way is to expose correctly in the first place. Until you learn how to judge exposure take a still photo & look at the histogram. As I recall the instruction with Cinestyle is to turn the in-camera sharpening & contrast right down.

Before mucking about with the complexity of Cinestyle, 5DtoRGB. Magic Lantern etc you really should just learn to shoot correctly exposed video with a standard picture style e.g. Neutral. Use Log & Transfer in FCP & convert to ProRes LT then edit that. Look up the recommendations on YouTube or Vimeo as to bit rate for upload but in general for 1280x720 5Mbps is ample. Unless your viewers have super fast broadband they won't be able to view 1920x1080 without stuttering.

[Brian David Melnyk]

Originally Posted by Daniel B. Kim

@Brian: Does that mean you don't use After Effects at all to do your grading? I wanted to try importing my footage into After Effects to do grading, but for convenience sake, I decided against it. I'm assuming Noise Reduction and everything else you mentioned can me found within FCP as well? I looked up the DP6 Monitor and it looks stellar! SDI out with great native resolution, but it's a bit out of my price range, is it worth the investment? Are there other monitors that do a great job for half the price? What is a Broadcast Safe?

Perspectives on the World Christian Movement Class Promo - YouTube

i don't use AE. I have experimented with Apple's Color, which is incredibly powerful, but for my purposes and workflow, using the Three Way CC in FCP is perfectly adequate, especially considering time/ease of use and the fact that my CC is not so extensive.

Neat noise reduction is a 3rd party plug-in, and not so expensive. it is also worth every penny, as it can make unusable footage usable, and good footage great if used judiciously. the Broadcast safe included with FCP clamps down the luma/chroma within a certain range safe for broadcast, and i think is subtle if you have done the work to grade the footage close to the range in the first place. i use it just in case my projects may be broadcast.

the DP6 is in my opinion very worth it! i tried a view finder magnifier and it was not a great solution for me. it didn't help my focus as much as i would have liked, and it made framing a painful experience. i also like to hold the camera low and high and at all sorts of angles, and with a viewfinder this does not work. the dp6 can swivel and tilt, has a hood (very important under the Malian sun), has peaking and false color, and other very helpful features, as well as compensating for the 5d mkii's unfortunate drop in resolution when in record- something other monitors do not do, or do not do automatically. in-focus, well exposed footage quickly justifies the cost!

if your problems are especially noticeable after export, i would look at using Compressor. it is a learning curve, but a very important step to take. there are also some great presets you can download and install that help greatly (ie- for you tube, ntsc to pal, best quality for DVD etc.)- search this site for posts on Compressor by Robert Lane for starters.

i use 5DtoRGB>prores>FCP>Compressor and it seems to work for me. i have read that 5DtoRGB does a superior job transcoding to prores, but there may be differing opinions.

[Zsolt Gordos]

I do the following in post:

throw all the clips at MPEG Streamclip to export them to ProRes (choose whichever you like to work, but 4444 is being told to be good for coloring)
Squared 5 - MPEG Streamclip video converter for Mac OS X free

edit the Prores files in FCP then depending on the final output needed export into compressor or use Elgato Turbo.264 HD (the one with the USB hardware dongle for faster processing) This tiny thing can compress your video to youtube formats with a number of setting choices very fast and with excellent results.

?Tropical Garden Trailer? 7 - Flowers - YouTube

I would never edit in other than prores for saving time. Well, need some hard drives due to the size.