Conforming in Resolve: Inaccurate EDL Recognition

[Desmond Ford]

Trying to conform an EDL from a two camera shoot from FCP7. The timecode was synced correctly on set, no effects and everything was brought down to a single track in FCP7, so I wasn't expecting any issues... Then I see that Resolve is recognizing the wrong camera albeit with correct timecode in at least 20% of the clips.

This is easily verified from the EDL of which I've attached a screenshot. As you can see in the highlighted screenshot attached, according to the EDL, Clip #09 is B009_C004 with the start timecode 12:21:33:14.

However, Resolve recognizes it as A009_C004 with the exact same start timecode! It's the exact same shot in the exact same moment in time... but from the A Camera! The same editor also gave me an XML and it wouldn't conform either. Big surprise.

I went to the settings and went down the line changing everything in the conform section, testing, trying again, testing more, and this is the best I could get. Anyone have any ideas?


Yet another conform issue with Resolve... it's getting to be that simple projects like these take longer to conform than to actually grade.

d

 

[Marc Wielage]

Resolve will at least let you compare each shot and flag the ones that are questionable. I always make it a point to load in a reference movie and lock that together with the newly-conformed timeline and check it cut-to-cut to make sure everything is OK. As far as I'm concerned, this is a necessary part of the process.

I've seen Baselight screw up conforms sometimes with this as well, so this problem is not confined solely to Resolve. Note that Resolve is much better now at this than it was 2 or 3 years ago.

BTW, I still prefer using DPX files or ProRes 422 HQ files in Resolve rather than R3D, just to reduce overhead and let me use more nodes when necessary, rather than drag system performance down by real-time debayering. That's my preference, and I concede that when processors are fast enough someday, this won't matter.

 

[Desmond Ford]

Resolve will at least let you compare each shot and flag the ones that are questionable. I always make it a point to load in a reference movie and lock that together with the newly-conformed timeline and check it cut-to-cut to make sure everything is OK. As far as I'm concerned, this is a necessary part of the process.

True, but when you have approximately 20% of a job out of whack, that can easily add up to a couple hours just conforming... it often has been that a conform takes longer than the actual grade and render.

If it really is a necessary part of the process, then I need to reconsider my rates because this is just happening too often. I always estimate at least one hour for a conform (I do mostly commercial work), but it would seem that I really need to factor in two hours.

I've seen Baselight screw up conforms sometimes with this as well, so this problem is not confined solely to Resolve. Note that Resolve is much better now at this than it was 2 or 3 years ago.

I thought Baselight was supposed to have the best conform options of them all? From what I've heard, its statistical conform is very accurate.

BTW, I still prefer using DPX files or ProRes 422 HQ files in Resolve rather than R3D, just to reduce overhead and let me use more nodes when necessary, rather than drag system performance down by real-time debayering. That's my preference, and I concede that when processors are fast enough someday, this won't matter.

I really don't have this problem at all. My system is really powerful and I don't really notice any lag unless I'm in 5K full Res, which I usually take down to half res premium and then it runs perfectly. I can also consistently run 4K in realtime at full res.

What are you system specs?

Thanks for the response, Marc.

 

[M Most]

Your EDL has tape names that are truncated to 8 places and mismatched to the original R3D naming convention, and clip names that have a .mov extension. All of these things are reasons why I prefer Avid, but that's another conversation.

One thing you can try, depending on how many source clips you're dealing with here, is to edit the EDL to have an underscore in the tape name between the reel and the clip (i.e., make A001C002 into A001_C002). This will allow you to use the format string in Resolve's conform options to match the tape name without much difficulty (try "*/%R_*_*", that should give you what I'm referring to). That should "resolve" some of the conflicts that you might be seeing by making the source selection less ambiguous. It sounds like the way you have it set up now, Resolve is only using time code as its match criteria because it doesn't see a tape name or source file name match. This will clear that up.

 

[Desmond Ford]

Your EDL has tape names that are truncated to 8 places and mismatched to the original R3D naming convention, and clip names that have a .mov extension. All of these things are reasons why I prefer Avid, but that's another conversation.

Yes, I noticed that the tape names were missing the underscore, but if that is what is causing the non-conform, then how is it that 80% of the files do conform? All files are missing the underscore.

Also, all the files have .mov extensions because it was edited in ProRes 422HQ in FCP7. I thought that this was a standard workflow.

One thing you can try, depending on how many source clips you're dealing with here, is to edit the EDL to have an underscore in the tape name between the reel and the clip (i.e., make A001C002 into A001_C002). This will allow you to use the format string in Resolve's conform options to match the tape name without much difficulty (try "*/%R_*_*", that should give you what I'm referring to). That should "resolve" some of the conflicts that you might be seeing by making the source selection less ambiguous.

That's a really great idea, but unfortunately I have almost 60 source clips that did not conform. To go through and individually edit them would probably take just as long as manually conforming them in the timeline... and it'd probably be an even bigger headache.

It sounds like the way you have it set up now, Resolve is only using time code as its match criteria because it doesn't see a tape name or source file name match. This will clear that up.

I've already tried this, but to no avail, by changing the conform options in the settings, such as reel name, source file match, clip names, etc... still no luck.


Thanks for you response, M Most.

 

[M Most]

That's a really great idea, but unfortunately I have almost 60 source clips that did not conform. To go through and individually edit them would probably take just as long as manually conforming them in the timeline... and it'd probably be an even bigger headache......I've already tried this, but to no avail, by changing the conform options in the settings, such as reel name, source file match, clip names, etc... still no luck.

Honestly, editing just the reel name in an EDL is going to take less than an hour even if you do it by hand on every line in the EDL. Another way to do it if you're script savvy is to write a script to take every line in the EDL file that starts with a number (0-9) and go 9 or 10 places in (whichever it it in your particular EDL format), insert an underscore, and write out the line again. Then again, if you can get the conform completed in less than an hour, I don't know why you're seeing it as such a problem.

It sounds like you're shooting in the dark and trying a lot of different things without really understanding what each of them do. I'm trying to help you by giving you a method that will work, but it's up to you as to whether you want to try it or not. What I'm telling you is to go into the conform options, and deselect everything except "assist using reel names from the" and highlight "source clip file pathname" and enter the string I gave you above. If most of the show is properly conformed, you can do this on just the shots that aren't. As long as you edit the tape names in the EDL in the way I suggested, this will absolutely work provided it can find the clips in question. Conforms are not magic and they're not complex. They depend on tape name and time code, and that's all they depend on. Anything else is just substituting some other string for the tape name. If the tape name and time code match, the relink will happen. If they don't, it won't. It really is that simple.

 

[Frank Cueto]

Heck, just using the find and replace function in TextEdit on a Mac should suffice. Tell it to replace the "C00" with a "_C00".

Then Dave that as a separate edl. Use that edl to bring the bad stuff but keep te good stuff from the first edl. 20 minutes tops.

 

[Frank Cueto]

Dave=use

Damnyouautocorrect!

 

[M Most]

Heck, just using the find and replace function in TextEdit on a Mac should suffice. Tell it to replace the "C00" with a "_C00".

Geez, Frankie, I must have been delirious. That IS a lot simpler. What was I thinking? Oh yeah, right, it's a holiday weekend........

 

[jake blackstone]

People always ask why bother to use EDL, when one can use XML. Well, this would be one of the reasons. It's much easier to troubleshoot conform, that gone awry with EDL, than with XML.
I used the find and replace trick in the past, when Resolve had no "ignore the extension" function to replace .mov with .r3d.

 

[Mikael Lubtchansky]

Wrote a free tool a while ago called Red16EDL for FCP edl... It will change the truncated FCP reel names to full 16 characters clip names using the comments from the edl... Haven't used it myself in a while but it should still work fine... See the link in my signature...

 

[Desmond Ford]

Honestly, editing just the reel name in an EDL is going to take less than an hour even if you do it by hand on every line in the EDL. Another way to do it if you're script savvy is to write a script to take every line in the EDL file that starts with a number (0-9) and go 9 or 10 places in (whichever it it in your particular EDL format), insert an underscore, and write out the line again. Then again, if you can get the conform completed in less than an hour, I don't know why you're seeing it as such a problem.

It sounds like you're shooting in the dark and trying a lot of different things without really understanding what each of them do. I'm trying to help you by giving you a method that will work, but it's up to you as to whether you want to try it or not. What I'm telling you is to go into the conform options, and deselect everything except "assist using reel names from the" and highlight "source clip file pathname" and enter the string I gave you above. If most of the show is properly conformed, you can do this on just the shots that aren't. As long as you edit the tape names in the EDL in the way I suggested, this will absolutely work provided it can find the clips in question. Conforms are not magic and they're not complex. They depend on tape name and time code, and that's all they depend on. Anything else is just substituting some other string for the tape name. If the tape name and time code match, the relink will happen. If they don't, it won't. It really is that simple.

Thanks for your replies. I hope that my responses have not come off poorly. I really appreciate your help as you have some really great and helpful ideas (not to mention your contributions on other threads). I will definitely try them out tomorrow... but unfortunately I don't know any scripting.

I guess I see the conforming in Resolve as a major issue because spending one hour to conform a project that will take me 45 minutes to grade and render just doesn't make sense to me, it's so frustrating...

Manually conforming can be so tedious and with so many similar-looking clips, it can get very confusing and I get lost...

What's more, I've spoken to several other Resolve colorists (a few veterans too) who share the same frustrations as myself in respect to conforming in Resolve, and from what I've gathered, the general consensus seems to be that this is just a fact of life for Resolve users.


So I went ahead and tried to conform the exact same EDL in Smoke for Mac... and I got a 100% perfect conform! Thus, I have to ask, if Smac can give me a perfect conform, then why can't Resolve with the exact same EDL? Surely this must be a Resolve Conform Issue...

Thanks for your help though. I'm going to continue trying.


dezzy

 

[jake blackstone]

Since Resolve had been ported to Mac, BM made great progress in improving the conforming workflow. In the past, I would frequently use Scratch just to conform R3D and export DPX for grading, because Resolve was so poor in comparison. Now using Resolve, when used with properly prepared EDL, XML or AAF it is a breeze to conform. Recently I conformed a whole feature with mixed R3D and Prores transcodes of 5D in less, than an hour, with zero errors. The majority of time was used to verify the accuracy of the conform. The actual conform was, at most, 10 minutes. I do wish, that Resolve would employ something similar to Scratch or Smoke conform window, where it reports missing or found material with necessary metadata. That would make troubleshooting conforming issues much easier. But as it stands right now, I personally find Resolve's conforming tools pretty much in line with the rest of the industry software and in some respects, like roundtripping to FCP or FCPX is even better than most.

 

[Desmond Ford]

Since Resolve had been ported to Mac, BM made great progress in improving the conforming workflow. In the past, I would frequently use Scratch just to conform R3D and export DPX for grading, because Resolve was so poor in comparison. Now using Resolve, when used with properly prepared EDL, XML or AAF it is a breeze to conform. Recently I conformed a whole feature with mixed R3D and Prores transcodes of 5D in less, than an hour, with zero errors. The majority of time was used to verify the accuracy of the conform. The actual conform was, at most, 10 minutes. I do wish, that Resolve would employ something similar to Scratch or Smoke conform window, where it reports missing or found material with necessary metadata. That would make troubleshooting conforming issues much easier. But as it stands right now, I personally find Resolve's conforming tools pretty much in line with the rest of the industry software and in some respects, like roundtripping to FCP or FCPX is even better than most.

Thanks for the input, Jake.

I've considered doing the conform in Smoke and just grading from DPX, but I always thought that I'd be cheating myself... maybe not... guess I'll have to rethink this one.

Yes, I also hear from a lot of people who seem to have great results conforming in Resolve. So jealous. I have to wonder what is wrong with my workflow...

As far as properly prepared XMLs go, I always have the editors bring everything down to one track, remove all effects, transitions, speed ramps, and graphics... Is there something else that I'm missing?

Also, is there a particular XML version that plays better with Resolve? I ask since I usually get Version 4 (because that is what the Flame artists prefer), but I've also had version 5. I'll have to start keeping track to see which ones I have more trouble with.



dezzy

 

[jake blackstone]

Thanks for the input, Jake.

I've considered doing the conform in Smoke and just grading from DPX, but I always thought that I'd be cheating myself... maybe not... guess I'll have to rethink this one.

Yes, I also hear from a lot of people who seem to have great results conforming in Resolve. So jealous. I have to wonder what is wrong with my workflow...

As far as properly prepared XMLs go, I always have the editors bring everything down to one track, remove all effects, transitions, speed ramps, and graphics... Is there something else that I'm missing?

Also, is there a particular XML version that plays better with Resolve? I ask since I usually get Version 4 (because that is what the Flame artists prefer), but I've also had version 5. I'll have to start keeping track to see which ones I have more trouble with.



dezzy

Make sure you have the proper reel names sorted out. Actually, FCP effects work quite fine with XML for Resolve. Resolve just ignores them during the conform, but when it creates the new XML for the roundtrip all effects are preserved and they get automatically applied, when graded project is re-assembled in FCP. Cuts, dissolves and dip to/from black work fine too.
When working with EDL, I do exactly, as you described: "bring everything down to one track, remove all effects, transitions, speed ramps, and graphics". Although, cuts and dissolves are allowed.
Personally, i just don't trust editors to properly prepare the EDL or XML. Not every editor is well versed in preparing the project for the conform. I had many instances, when editors just didn't understand the importance of the reel names. So, when possible, I'd ask for the FCP project in addition to EDL, XML and the offline copy, so, if needed, I can open the project and troubleshoot it myself, rather, than asking the editor to do it.
Personally, I just would use DPX for grading and get on with my life. Finish the job and then try to figure out why the present workflow doesn't work for you.

 

[Desmond Ford]

One thing you can try, depending on how many source clips you're dealing with here, is to edit the EDL to have an underscore in the tape name between the reel and the clip (i.e., make A001C002 into A001_C002). This will allow you to use the format string in Resolve's conform options to match the tape name without much difficulty (try "*/%R_*_*", that should give you what I'm referring to). That should "resolve" some of the conflicts that you might be seeing by making the source selection less ambiguous. It sounds like the way you have it set up now, Resolve is only using time code as its match criteria because it doesn't see a tape name or source file name match. This will clear that up.

I'll be damned... it worked!!!

When I got into the office today, I started in on it. First thing I did was to edit one of the erroneous clips in the EDL with an underscore. Then I tried the conform to no avail. So then I changed the source clip file pattern in the conform setting to "*/%R_*_*" as you suggested and gave it a go and it worked perfectly.

In fact, it conformed the entire project. I didn't even have to edit the individual clip names in EDL at all... this is huge!

I always thought that the default source clip pattern was the best, most comprehensive, but clearly it is not.

Thank you so much M Most. You just saved me hours and I'm sure I'll benefit from to applying this technique to future projects. I suspect this has been the missing link in my conform workflow all this time.. though I can't be sure just yet... I'll report back in a few weeks.

Thanks again!


dezzy

 

[M Most]

I'll be damned... it worked!!!

Glad it helped.

As I said earlier, conforming is really a matter of getting metadata to match. Using EDLs makes it much easier to change tape names if necessary, and using the pattern string allows you to match those names in Resolve. If you ever get confused, just look at the tape names in the EDL, and compare them to the Reel names in the Media Pool. If you've set up everything correctly they should match. And if they do, conforming should be straightforward.