best Avid canon C500/C300 workflow
I am currently planning my post workflow for editing a long form documentary feature. This film has a lot of footage to be dealt with, mostly shot on a Canon C500, or C300 in 1080p 23.97 canon XF files.
These files link up well via AMA to my Avid, and in testing I am able to do light editing and color grading quite easily directly to these files. I have not yet done any consolidating or transcoding, because I want to make certain I am making the right choices before handling the large volume of footage. Ample storage has been budgeted for and it would be nice to not have to offline, if possible. I have a fast Avid mac based system.
My questions pertain to the benefits and any drawbacks to the consolidating/transcoding options.
1) I am hearing that working directly with the files without transcoding or consolidating is possible but not recommended because of performance issues- I am wondering exactly what those issues are?
2) If I were to transcode all of the footage for the project to DNxHD175X, would there be ANY quality loss?
3) What benefit is there to just consolidating the canon files into the avidMediaFiles folder? Is this also for performance? Does it maintain the best image quality?
I am hoping the color grading can be done in Avid so I am hopeful we can maintain the best quality throughout and endpoint quality is the biggest deciding factor in workflow.
Thanks in advance for any ideas or shared experience!
"BEST" workflow is relative. Because I work on two series with different workflows. One we work full resolution, simply AMA linking and consolidating (recreation style doc...scripted). The other we have TONS of footage (reality show) so we AMA link to then transcode to DV.
[Mary Haverstick] "1) I am hearing that working directly with the files without transcoding or consolidating is possible but not recommended because of performance issues- I am wondering exactly what those issues are?"
Sluggishness playback...sluggish system response in general.
[Mary Haverstick] "2) If I were to transcode all of the footage for the project to DNxHD175X, would there be ANY quality loss?"
There always is when you convert from one codec to another. BUT...no need to do that. CONSOLIDATE the media. It's already in a codec Avid works with, XDCAM 422...it's already MXF...but it's not quite in the Avid file structure, so that's why it isn't quite up to speed. It's best to AMA link to the footage on the archive drive, and then consolidate it to the MEDIA drive. That's basically just a COPY...zero quality loss. And it takes up less space than DNx175.
So, instead of copying the cards to the media drive on the desktop level...plug in the archive drive, AMA the footage, and consolidate.
[Mary Haverstick] "3) What benefit is there to just consolidating the canon files into the avidMediaFiles folder?"
More responsive system. Zero quality loss. Less space used than DNx175
[Mary Haverstick] "Is this also for performance? Does it maintain the best image quality?"
yes and yes
Little Frog Post
Read my blog, Little Frog in High Def
Shane's advice is sound, but some things to consider since you mentioned you mentioned image quality in color grading...
I believe your cameras codecs are 8 bit. Transcoding to Avid DNX 175X may be of value there, It won't improve the base image quality, but will give you more latitude in Color grading to avoid banding. However, If you are color grading within Avid, you can accomplish the same thing by just simply changing your render codec in the Media Creation settings to DNX 175x, and your renders will be 10 bit DNX175x.
One other thing to consider. Long GOP codecs might make your system sluggish, even if consolidated to Avid media. This is because it's more processor intensive to decode. If your system isn't up to it, you might want to consider transcoding to DNX 175x, if you have the drive space and your drives are fast enough.
John and Shane have pretty much covered all the different considerations. The issue with AMA linked to longGop and similar codecs is that it can work well for a while, then as you get more footage, timeline get longer or more complex, you want to do multicam, etc, the performance is just not there to what you expect it to be. Most production will take the transcode hit up front to eliminate that right off the bat.
The Avid storage calculator hasn't been updated in years, but still works for DNxHD workflows. This will give you an idea of how much storage to plan for given your total acquiaition.
This is all great info- I thank you for sharing- I am running some tests and comparisons before making my final decision- this has been MOST helpful!
I think I am going to run a test by creating a long sequence of gibberish (from consolidated) but see if the system starts to bog down. I can throw some color effects on and just see it it doesn't run or play well. I don't want to get wayu into an edit on consolidated media and find it isn't running well enough.
I am also interested in testing the image quality difference of 175X- if it is even noticeable. And thanks for the reminders on color grading!
Thank you Shane- this is great info and very thorough- much appreciated!
I wanted to follow up on my question and your great responses with a little testing phase before I make my final decision and have to live with it for the next few months. I am sharing this in case others reading this thread find themselves in a similar situation.
My first test was to take a scene and transcode it to 175X dnXhd and see if I could perceive ANY quality loss. I intercut shots from the 175X series against the same shots from directly AMA linked files. I know avid says you can't perceive any difference- and on most shots I didn't, but on a few shots where I had a speaker on a close up I felt I did perceive changes in the detail on the skin pores and eyelash details- ever so slight but just a little softness compared to the AMA file. These were viewed full frame HD and not altered in any way. Could I be imagining a tiny difference? I think it is there- and while I could not perceive anything on architectural shots I was able to pick the AMA file consistently on a close up. My subject in the CU was a person who moved a lot and that could have been the reason the compression/transcode showed up on these shots more than the architectural wide shots. As some mentioned, the color on the dnXhd175X showed no degradation.
Would this be enough to sway me not to transcode? Not necessarily-the difference was so slight-But since I did not film in 4k I am not in a mood to lose anything of what we DO have in detail here. I am now going to consolidate a fair amount of footage and create a very long complex gibberish sequence and see if my system chugs- Will post results on that soon!
Thanks for the follow up. Always appreciated.
I think you might be overly critical, but without seeing the actual files, it is difficult to tell. Are these 4K files you are linking to? If so, there can be differences in how the resize is affecting the overall quality as a link versus an actual transcode and you may be seeing some difference there. While Media Composer's resize algorithms are good, they are great and do not optimize based on scaling up or scaling down.
A test I usually do is create DNxHD out of Resolve using it's resizers and compare it to Media Composer for critical material. I tend to find Resolve consistently does a better job, even if fractionally better in some cases. As an example for another project I was doing, I referred to this difference in a blog I was writing on "portrait" editing where it did involve both a rotate and a resize and compared the output frames of Resolve and Media Composer.
If 4K sources and finishing at 1080, there will be a downrez scaling operation somewhere in your pipeline. If those shots are critical, then you need to decide where that is being done in your workflow pipeline. DNxHD 175x has been used to finish many a broadcast show, so it may be something that is unique to your footage, lighting, angle, etc. A quick test would be to transcode those clips to uncompressed and see if you see the same thing. If it's still there, then it is not the compression, but either your source material or the downrez (if greater than HD sources).
How did you critically judge this quality? On your computer display, or a broadcast monitor connected via Nitris/Mojo/AJA/BMD IO device? Full screen playback on a computer monitor isn't a true representation of the quality..
Little Frog Post
Read my blog, Little Frog in High Def
Good question- quality was viewed on full screen playback on the cinema screen as well as via Blackmagic design IO to a sony HD 1080 TV, consumer set. No studio level HD set avail here- but quality difference seen at cuts and confirmed by my assistant editor as well. Direct AMA linked footage is the tiniest bit better-problems seen in same areas of the frame- small differences in skin detail on close ups in motion- very consistent. I shouldn't say the word "problems" as it is acceptable- but I am just seeing the smallest amount of loss differential.
Did you try the uncompressed test just to see?
My decision was made fairly simply by the intervention of the editing Gods.
I was leaning toward consolidating and had tested a sequence that was long and complex and seemed to run well, from consolidated footage on my system. But then I did some simple house cleaning on my hard drives making additional space for any new footage- and even though I did nothing that should have interrupted my existing footage or hierarchies I had one of those buggy days where nothing runs well. In particular I had trouble accessing and linking some of my AMA footage- and that was stressful. I did relink it thru disk permissions, but I hate those buggy days.
Realizing how incredibly stressful that would be should I be 3-4 months into my feature edit- and knowing the AMA links can be more finicky than avid DNxHD I decided the slight, perhaps imagined and viewed on inferior monitors, possible tiny quality drop did not compare to performance headaches that could induce heart attacks near the end of the edit.
Transcoding over 30 hours of Canon footage took only one long day and I am up and running with my edit already. I immediately felt a sense of additional confidence knowing my important footage was now saved and accessible several different ways. I thank everyone of you experts who assisted my decision and I will drop an update on the project from time to time. I hope this thread helps a few who are planning their Canon workflows.