editing 30fps to 24fps?
- Thread starter chlsy8
- Start date
We always slate everything, and most editors I know of sync audio with picture before actually getting into editing.
If nothing else, then I'm sure the Director would like to see the edit come together with decent audio, rather than with scratch-track stuff that often times is barely audible... not to mention the issues I can imagine that would arise in the pacing of the editing when you can barely hear what someone's saying.
And considering that DSLRs can't be timecode-linked, that timecode often slips if it's not constantly re-checked, and the fact that timecode jam-slates are hugely expensive, I can't imagine timecode would be a reliable way to sync on most low-budget shows. Therefore, what does the sound editor use to sync audio if they have only a picture lock edit and the raw sound files? I can't see how that would work at all, not to mention the fact that unless you provide them with an extensive list of what take was used where, it would create issues in syncing as an editor may have had to wade through 10 takes to find the one used...
Just seems like a lot more work that you'd be paying for, that you could have your editor or assistant editor spend a day or two on before he starts editing and save a lot of hassles and issues and time down the line...
But then, I'm a DP not an editor or sound guy.
If nothing else, then I'm sure the Director would like to see the edit come together with decent audio, rather than with scratch-track stuff that often times is barely audible... not to mention the issues I can imagine that would arise in the pacing of the editing when you can barely hear what someone's saying.
And considering that DSLRs can't be timecode-linked, that timecode often slips if it's not constantly re-checked, and the fact that timecode jam-slates are hugely expensive, I can't imagine timecode would be a reliable way to sync on most low-budget shows. Therefore, what does the sound editor use to sync audio if they have only a picture lock edit and the raw sound files? I can't see how that would work at all, not to mention the fact that unless you provide them with an extensive list of what take was used where, it would create issues in syncing as an editor may have had to wade through 10 takes to find the one used...
Just seems like a lot more work that you'd be paying for, that you could have your editor or assistant editor spend a day or two on before he starts editing and save a lot of hassles and issues and time down the line...
But then, I'm a DP not an editor or sound guy.
Seriously, you guys are making a big deal out of nothing. Also, you're speaking in generalities, none of which apply to the specific subject at hand. You're also speaking authoritatively about a subject in which you are no authority.
Yeah, that was me. Sorry about that. I was using an export setting I'd never used before, and I missed that little important detail.
But that actually has NOTHING to do with what we're talking about. For the project you're describing, all that matters is the frame rate that is exported to. In the example you reference, I mistakenly exported to true 24p, instead of 23.98, and that caused headaches. The frame rate of the raw footage, however, would have absolutely NO bearing on your audio edit. No matter how much of my footage might have been 24/30/60, the final output would be the same.
Even if the OP were recording time-coded audio (I doubt it), there still would be zero issue in mixing frame rates, during the edit. Let's say all of the audio is time-coded for 24. The DP goofs and shoots some footage at 30. Uh-oh, now that mistake video and the audio have different time codes. What are we to do? The sky is falling!
Oh, wait. We can edit all of the video in a 24fps timeline, export to 24 and vioala -- now the old 30fps footage is time-stamped with 24fps time code. Doesn't help if you're syncing to the slate, but how many people on this forum are doing that? For most people on this forum, the scratch audio is all you need, and the slate gets practically no use.
All is well. The sky is not falling. Just edit the damn movie, no luck required to get the job done.
Yeah, that was me. Sorry about that. I was using an export setting I'd never used before, and I missed that little important detail.
But that actually has NOTHING to do with what we're talking about. For the project you're describing, all that matters is the frame rate that is exported to. In the example you reference, I mistakenly exported to true 24p, instead of 23.98, and that caused headaches. The frame rate of the raw footage, however, would have absolutely NO bearing on your audio edit. No matter how much of my footage might have been 24/30/60, the final output would be the same.
Even if the OP were recording time-coded audio (I doubt it), there still would be zero issue in mixing frame rates, during the edit. Let's say all of the audio is time-coded for 24. The DP goofs and shoots some footage at 30. Uh-oh, now that mistake video and the audio have different time codes. What are we to do? The sky is falling!
Oh, wait. We can edit all of the video in a 24fps timeline, export to 24 and vioala -- now the old 30fps footage is time-stamped with 24fps time code. Doesn't help if you're syncing to the slate, but how many people on this forum are doing that? For most people on this forum, the scratch audio is all you need, and the slate gets practically no use.
All is well. The sky is not falling. Just edit the damn movie, no luck required to get the job done.
Obviously the dialogue editor needs more than just a locked edit and the raw sound files. We need at least a scratch-track (with sync'ed camera sound). On the face of it, I agree it appears to be quicker/cheaper to have the assistant pic editor cut in the better mic's and in practice, on higher budget shows this is what happens but in reality this actually means paying for this job to be done twice. The problem is that a picture editor (or assistant) does not have the equipment to either accurately sync the sound or to accurately decide which mic is best. In reality, virtually all the scenes you hear in a finished professional film are a combination of various takes, many of which are unrelated to the take used in the visuals, plus of course the odd bit of ADR thrown in when nothing else usable can be found.
I agree, there's no point in recording time-code unless you can sync-lock the recorder and camera together and this is maybe too expensive for those with low budgets. On the other hand, not paying for this probably a false economy because you'll likely pay even more at the other end, during post. With a full length feature it's going to take months to find and go through all the alt takes (and mic's) amongst the thousands of files and this is also going to cost a fortune. The only other alternative is ADR, which is both costly and degrades the quality of the final product.
I'm afraid this is inaccurate, or at least, it is not always accurate. Fractional frame rates, like 23.976fps (or NTSC, 29.97fps) are actually 24fps (or 30fps in the case of NTSC) played slightly slow. One second of time-code with these frame rates does not equal one second of real time, which is why DF (drop frame) rates had to be invented. If you film at 23.976fps (and record audio at this rate) and then do a straight conversion from 23.976fps to 24fps then the picture will slightly speed up relative to the original audio and they will gradually drift out of sync with each other. The way around this is to convert using interpolation but then the issues I listed in my previous post apply. This is true when converting any frame rate to any other frame rate. In fact there is/was a very common workflow which involves this very problem, which is why "pull up" and "pull down" audio recording rates exist.
I perhaps wouldn't call myself an authority, I still occasionally have to seek more expert advise but after 20 years dealing professionally on an almost daily basis with frame rates and frame rate issues between picture and sound, I would say I've got a fairly decent understanding on the subject.
BTW, in the two different frame rate versions you sent me, I extracted the embedded audio from both and compared them with each other and the separate audio scratch track supplied, most of the time the sync was reasonably close although not perfect but in other places there was up to 4 frames difference between them. This is significant and why changing picture frame rates can be problematic.
G
Considering the primary edit, CF is right. Dropping 30p footage in a 24p timeline will before fine, in Premiere. Premiere doesn't speed or slow the footage to work, it drops or interpolates frames to keep it the same. Premiere does a fantastic job at it too, I've never been able to see any jitter or resolution loss.
FCP7 is now a 4 or 5 year old program, keep that in mind. It won't handle the difference quite so well. I'm fuzzy on it, but it might not let you put it on the timeline at all, and of it does it'll be render city. Everyone you make a cut or move a clip you'll have to render EVERYTHING. In this case, you can slow the 30p footage down to 24p (and slow down the audio by 20%). What I would try first is converting all your 30p footage to 24p ProRes or something with Compressor. It might drop frames for you and have it looking sharp.
OR... You could download the free 30 day trial of Premiere and finish your edit in a month with way fewer renders and more powerful tools and what not.
FCP7 is now a 4 or 5 year old program, keep that in mind. It won't handle the difference quite so well. I'm fuzzy on it, but it might not let you put it on the timeline at all, and of it does it'll be render city. Everyone you make a cut or move a clip you'll have to render EVERYTHING. In this case, you can slow the 30p footage down to 24p (and slow down the audio by 20%). What I would try first is converting all your 30p footage to 24p ProRes or something with Compressor. It might drop frames for you and have it looking sharp.
OR... You could download the free 30 day trial of Premiere and finish your edit in a month with way fewer renders and more powerful tools and what not.
Right. And that would be expected. Through the entire process, from shooting, to rough cut, everything was shot/edited/exported on/to 23.976. The 1080 version I originally sent you, with timecode, was the first time that true 24p was mistakenly introduced to this project. So, yeah, that would definitely cause sync issues, because it is playing back both video and audio at a slightly slower speed. But had I kept everything in the original 23.976, the shooting frame rate has absolutely no bearing on what is exported.
Regardless, I'm glad you noticed and figured out the problem. Hope it wasn't too big a pain in the rear for you.
I get what you're saying -- frame rate matters. And you pointed out a specific recent example in which frame rate matters. But the frame rate you're talking about is different from the frame rate the OP is asking about. You're talking about the frame rate of the exported rough cut. And that definitely matters. A lot. The OP is asking about the frame rate that the video was shot in, and will be edited in. And that doesn't matter, because by the time it gets to you, it will be conformed to the 23.976, and will match all other footage.
And that is all I meant when I asked whether or not you're an authority on this subject. You're an authority on audio post stuff, no question about that. But that's not the conversation we're having. We're talking about editing video, before it gets to you, and the stuff we're talking about won't effect your job, not ever so slightly.
OR... You could download the free 30 day trial of Premiere and finish your edit in a month with way fewer renders and more powerful tools and what not.
If I remember correctly, Vegas also handles all formats, in any timeline. So...same idea -- if FCP doesn't work it so well, Vegas would also be an option to download the free trial of. Import all of the footage into a 24p timeline, then export in 24p, and then use all of that converted footage in FCP.
I'm pretty sure FCP will handle it, I used to be an FCP buff before I bought Avid.
And I'm not even an editor
The other factor to consider, I would think, is that to assume the audio would be synced later is to assume that the OP has a pro audio post guy lined up. If he doesn't, then there may be an issue here, no?
And I'm not even an editor
The other factor to consider, I would think, is that to assume the audio would be synced later is to assume that the OP has a pro audio post guy lined up. If he doesn't, then there may be an issue here, no?
Ah but it does matter because in audio post we have to deal with all the production sound. So frame rate changes between filming and editing (before audio post) could cause exactly the same problem that we've experienced between editing and audio post. In audio post (whether done by a dedicated audio post guy or the picture editor), you frequently have to go back to the original production sound to find replacement/alternate dialogue, which is going to be a serious problem if the speed of the picture has changed in anyway since the filming because the original production sound will no longer sync with the picture.
With the tried and tested workflow of changing frame rates between filming on 35mm film and editing on video, the speed differences were known and predictable and the production sound mixer would use a pull up sample rate so the production sound could be pulled back down during audio post to sync with the picture edit.
As we have discovered though, when using interpolation in Premiere to convert between frame rates, there are slight discrepancies in picture speed/frame positions, nothing visibly noticeable but enough to cause sync issues with the sound in places. The problem is, that these discrepancies are no longer known or predictable and so cannot be compensated for as could be done with the standard workflow which involved frame rate changes.
Relating this all to the OP's question, he could have problems during dialogue editing when he goes back to the original production sound recorded when the picture was at 30fps.
One last point, Premiere is obviously interpolating both the picture and the linked audio (to maintain sync). Even expensive professional audio tools have difficulty with this type of process without causing some artefacts. You may not notice these artefacts under some conditions but in a cinema, say at a film festival, it's far more likely. It's certainly something I would be concerned about in a professional project.
Regardless, I'm glad you noticed and figured out the problem. Hope it wasn't too big a pain in the rear for you.
It caused a few headaches, half a day or so of extra work and a day or two of delay but it could have been much worse!
G
The only way this could effect audio post would be if the audio was recorded with a time stamp. And if you don't notice that your audio time stamp is 20% off from your edited video, you've got bigger problems. All you would have to do is ignore the time stamp. Audio has no frame rate.
Thank God we don't have to do that anymore. We can now just use the original audio, unmodified. How awesome is that? Hooray for the digital revolution!
When did we discover that? I see absolutely no reason why interpolation should even enter this conversation, neither for video nor audio.
He could? Why? The recording format of the audio is constant. You can put it in any timeline, and nothing changes.
Oh, it's obvious? I see no evidence of such (nor any logical explanation as to why this would be the case). Interpolation happens for pulldowns, and that's not what we're doing.
Yes, the production sound is constant but the picture isn't. If the picture and linked sound slows down by 20% the original production sound (not the audio in the timeline) will be 20% out of sync with the new picture. So we can't (easily) go back to the original production sound looking for alternative takes or mics to replace problem dialogue. This more serious problem is in addition to any problems caused time-code related issues.
Obviously when using Premiere we're not slowing the picture down by 20% because we are interpolating but:
From post #24: "BTW, in the two different frame rate versions you sent me, I extracted the embedded audio from both and compared them with each other and the separate audio scratch track supplied, most of the time the sync was reasonably close although not perfect but in other places there was up to 4 frames difference between them. This is significant and why changing picture frame rates can be problematic."
G
Last edited:
When you put 30p footage into a 24p timeline, you are neither slowing the footage down, nor interpolating. Every fifth frame is ignored.
And I already tried explaining to you how this has nothing to do with the concerns of the OP. The headaches caused by exporting a 23.976 project to true 24p is an entirely different conversation than the one we're trying to have to help the OP, who is concerned about using 30p footage in a 24p timeline. These two subjects are not related. At all. Nil. You're confused, and are confusing the conversation.
There's nothing more I can say about this; I'm done with this conversation.
I agree, I'm not an editor, I do find it confusing and I am confusing the conversation. But, I believe I am doing this with good reason. To start with you're assuming the OP is really talking about 23.976fps and 29.97fps. Even if this assumption is correct, in my experience, professional editors tend avoid just dropping every 6th frame due to motion stutter or artefacts from frame blending. I've heard of all sorts of weird and wonderful ways to get from 29.97 to 23.976 using Cinema Tools, Shake, After Effects and Optical Flow. As I said though, I am confused, I don't know or understand all the ins and outs of all the different pull-down, interpolation or other processes which can be used for this particular conversion. However, what I do know for sure is that they frequently cause problems during dialogue editing when it comes time to refer back to the original production sound.
I had assumed that Premiere was using some other, more convoluted processing than just dropping every 6th frame but in the past, all the picture/video editors I've worked with used FCP or AVID, so I've no experience of Premiere or of working with an editor using Premiere. So, I was assuming it was interpolating or doing some other process which would affect the audio because we've already discovered that it's not very accurate when it comes to audio sync with it's interpolation algorithms.
Baring in mind though that the OP is using FCP7 and not Premiere, I think my comments in general and my statement that the OP *could* have sync issues with the original production sound were appropriate. Hopefully, whatever method the OP chooses to convert from 30fps to 24fps, he will check that sync with the original production sound has been maintained before he gets to the stage where it causes a major issue.
G