BTW I'd like to report this bug is getting worse and worse for me. In a project with 7 interviews, 5 of them were unusable in transcription mode. Every time I tried to edit a transcription it would corrupt passages of text. Weirdly there were also missing passages of text and text incorrectly exporting when using the .txt file format with no line breaks (only happened on two sequences out of 7 but no apparent reason) and one project even completely refused to export a .srt. In that project the sequence was 25P but the transcript time codes appeated to be in a 60 base. eg 00;01;15;59 which made no sense. There is something seriously wrong with the engine. I've reported this as a bug but nothing has been done since my OP. The only workaround is to translate to captions and edit there.  ... View more

Nevermind - it's under teh progress window (which I didn't even know was there in AE). ... View more

I've expereinced that a lot when re-rendering files from AE but it's a different error to the one in this thread. This can happen on any file format including .mov which I use. What happens is typically you have Premiere Pro open and you notice a problem with your render so you switch to AE, make a change to the file and re-render to the same file name.  If you have that sequence open in PP and you are parked on any frame of that clip, whether it be in the sequence or source window this will cause the ERF error. and the only way to fix it, is as you said to clear the cache. The way to avoid this happening is to ensure that the sequence or source with the problem clip is not open, or at the very least you are not parked with the playhead on any frame of the original when you re-render the clip. I've never thought to post this as a bug  - I always assumed what was happening was the re-rendered clip was causing cache to be corrupted because PP does not update cache when the file name doesn't change. However it is not specific to MXF files as I only used QT these days from AE. I guess we could make it a request that when the timestamp changes on a file PP automatically updates the cache (because it is a gotcha that still gets me occasionally)? ... View more

Check and see if the problems go away when you select software decoding in preferences for timeline playback. Your system is probably powerful enough to decode MXF on the fly using software. If you see improvement the most likely problem is the NVidia driver not playing well with Adobe. It's been a long time since I was on a PC but there used to be advanced settings in the NVidia driver specifically for NLEs and their also used to be a specific driver (not the Studio driver). Perhaps check the driver release notes? Just suggestions to try - understand your frustration. ... View more

Ignore that - it is not a beta specific issue. The beta will always tell you the version in "about Premiere Pro". Best way to check if you're on the latest version is to go to the Creative Cloud updates page and see if there is a beta update. Please also check and post your video driver version if you're on a PC. While I understand it's frustrating all these things help to troubleshoot. OS version, hardware, video driver, media type (be specific) and even posting an example project. FWIW I have seen projects that are problematic and copying all your sequences into a brand new project has resolved some issues. Not saying it will work but may be worth a try as I note a lot of these problem projects seem to involve years of work.  ... View more

Thanks Stan - please note this is a big project, 3 x 4K streams per multicam clip and probably a terrabyte's worth of clips imoported and I have had a memory crash while Premiere was open (I had another large project open which was probably a mistake). Put it this way - Apple's new "unified memory" is more marketing than substance. My machine has 32GB of RAM but Premiere was using 108GB! However just for sanity's sake, I have seen this happen on a much smaller project, albeit still multicam. ... View more

FYI - same machine specs as my OP but now working on 24.5.0 build 7 and now on Sonoma 14.5 (23F79) ... View more

Yes and I'd like to add to it with maybe a way to replicate? This doesn't seem to happen on single video timelines but i work mostly with multicam which may explain why it's worse for me. Here's an example I video'd today (apologies for the shaky cam but I just didn't have time for a tripod) and while it's not exactly the same bug it does seem that this is now happening across a couple of projects. This is the duplication error I've mentioned. To me this whole thing looks like a memory. problem. ... View more

Hi Steve I find the problem is it's random. I used the auto transcription yesterday and editing the text worked perfectly. I did it last week and for one particular video I'd edit two words, click out and the text engine would duplicate the sentence before it - repeatedly. I ended up with multiple copies of the same sentence. My workaround in this case was to turn the text int captions and I edit the actual graphics. It's not ideal because I prefer to start with an accurate transcription I can send the client but I can always give them the .srt as text. My computer specs are in the OP. Please post yours to see if we are using similar spec'd machines.  ... View more

Given all the troubleshooting steps you've already taken my only two suggestions are to check permissions on all your drives or to move the project file to a local drive. I guess the only other thing to try might be a new user. Workaround would be to export from AE in ProRes and just put the clip over the top of the dynamic link. ... View more

Error Retrieving Frame errors have been around for a long while (most of them attributable to older media or dodgy hardware/cables etc) but this problem seems more specific, affecting a particular type of very specific (and very popular) mxf file from the Sony FX6. ERF errors are simply timing errors which is why one drive might not exhibit the errors and another might, even if they seem very similar. Remember, an NLE has a lot of work to do, decoding a highly compressed file in an MXF wrapper. In an effort to improve something else, Adobe appear to have introduced a big problem. Working from a NAS, even via 10Gbit connection can exacerbate this as NAS connections include a certain amount of network overhead which introduces... timing issues. Even via a Thunderbolt NAS you are still using a network protocol. This problem is also showing up on local drives but for those experiencing it over a NAS I'd suggest at least trying a very fast local drive while Adobe sorts it. Or downgrading to a previous version that was working for you. One thing that will work is transcoding the source media to an intraframe codec like ProRes or DnxHD (each frame discrete). ... View more

Can you specify what camera, bit rate and data rate the media is from and I can try and replicate on my M1. ... View more

By dynamic link do you mean After Effects clips or Audition or something else? ... View more

That is the other possibility. Try transcoding the media in Encoder to Prores. I know it's a nuisance but Prores is a more editing friendly codec. ... View more

They are sxactly the conditions you would expect to see this error. I assume you were using the SSDs that you were editing from? So they are busy doing that transfer task filling up their limited bandwidth pipes with data and down those same "pipes" you're asking your computer to read data off files continuously if trying to playback media. Doesn't matter how powerful your computer is, you've run into a bandwidth problem. FYI .mov is just a container.  .mov files can be H264 files. For example both Nikon and Canon cameras create h264 files in a .mov container. Unless the error persists after the transfer is finished just ignore it.  ... View more

the full green multicam block in the top image is totally unweildy and hard to interpret. At the risk of oversharing (and not at all intending to impose my own personal preferences), I disagree.  To me this acutally looks cleaner as it is clear that the audio and picture assets are associated together in the timeline while maintaining access to all the audio ISOs, and if track organization is consistent across the workflow, it's still relatively clear which track each subject lives on.  It also gives me access to the source audio at the source clip level rathre than having to rely explicitly on the stringout sequence.  As you mentioned though, this is more a case of expectation than functional limitation.  That said you do lose the subject labeling per track in the Multi-Cam, and I concede that this is nice to have from the Editor's perspective, sepecially if mic'd subjects are changing setup-to-setup. @Ben Insler  Ben - I'm going to point out a flaw in your logic. The great thing about seeing the source audio as discrete clips is you can tell where the camera started and stopped recording to the frame. You can't do that with a continuous green line. I've not heard a single editor complain about the previous functionality and plenty complain about the change. Users in this thread include a variety of very distinct and advanced workflows. Maybe consider reversing this decision as it doesn't seem too popular. It's not like you had users knocking down your door saying this nest multicam thing is a nuisance. I do appreciate all your input here but could you at least put it to a vote? ... View more

Sorry, no - I thought youor the mods could just break it out into a new thread. ... View more

Some final things you could check:   In settings / color check that you have diplay color mgt ON - also try turning it off     In settings / Media check that both Enable hardware accelerated decoding is on     ... View more

Make "selection follows playhead" off a sticky selection rather than defaulting to on every time a project restarts? This is sticky.  However, your Lumetri Color panel requires it to be enabled and will turn it back on if you turn it off.  Try closing your Lumetri Color panel when you don't need it.   You're right - let's break this out into a separate thread as it is something that needs discussion. Because, like a lot of editors I have a dual screen setup where Essential Graphics, E Sound and the Lumetri Panel are always open. ... View more

No, this behavior has not changed.  The tendency here is to want to build a Multi-Cam and then crash everything down to just the necessary tracks that reflect the angles you need with stop and starts in those tracks.  But, as mentioned above, this is dangerous.  The correct way to approach this is to make a Multi-cam (A) from only the first recording from each angle regardless of how many stop/starts you had during the take.  This will build the Multi-Cam correctly without extraneous stop/start angle tracks.  Then, make a separate single Multi-Cam Source Sequence (B) of all stops and starts across all angles, and crash that down to just the angles/tracks you need to accommodate the correct number of angles with stops and starts.  Once that's done, copy all the crashed down clips from B and paste them into the first Multi-Cam A (overwritting the original three that are in there).  This now gives you all your stops and starts per angle, in a correctly configured Multi-Cam Source Sequence that you have not modified on a track level (and at this point you can delete B; A is your hero Multi-Cam).   Actually it does work - I tested it yesterday. I used the Camera Angle metadata tag to change all my shots to a single camera (BTW you can't do this all at once you have to do it individually clip by clip which seems odd). You can also use the camera label metadata. This is determined by the Audio: Sequence Settings option in the Create Multi-Camera Source Sequence dialog. Camera 1: Configures the Multi-cam to only use audio from Camera 1.  All other camera audio will be muted. All Cameras: Configures the Multi-cam to use all audio from all cameras at once.  No camera will be muted. Switch Audio: Configures the Multi-Cam to switch audio camera to camera, so your audio can follow video when camera angles are switched   Not really important but if you have not labelled your cameras how does PP "know" which is Camera 1. In this case it treated my Audio Recorder as camera 1 which gave the correct result.   I still want to go into the multicam and create extra video angles Unfortunately this is one of those things that you don't want to be doing inside a Multi-Cam.  Adding and/or subtracting video and audio tracks from an existing Multi-Cam Source Sequence can result in very unpredictable flattening results that are hard to troubleshoot.   Well I've been doing it for years and the results seem pretty predictable to me. In what way do they become "unpredictable"? For example for a live event with Powerpoint slides I've actually gone into the multicam and added the full frame slides in time with the slides changing on the screen and then used them as a multicam source. Saves hours. I've also done the 4K trick where I duplicate the camera track and have a close up and a wide. The only thing unpredictable is the way PP displays those reframed angles on playback in the multicam window. Not a huge issue - you still know which is the close up and wide.   You really do not have to be jumping through all these hoops - Multi-Cam handles it all for you.  You're shooting in 4K but editing in 1080p, so you make your Multi-cams, open up each one in the timeline, change its dimensions to 1080p, and scale down each angle (for simplicity sake, let's just assume your 4K footage is scaled to exactly 50%).  This way all your Multi-Cams cut into your edit at full frame.  This is all spot on.   Now you seem to be saying I can change things in the multicam like dimensions and colour correction.    From this point, if you want to push in on a shot (let's say 200%), you're right, you don't want to push in on the already reduced Multi-Cam.  So you can just flatten the Multi-Cam.     This is indeed a more elegant way to achieve my desired result. Thanks for the tip. This is the way I will tackle this situation in the future.   Thanks for working with me Ben and the detailed responses- I now have a totally new and more efficient multicam workflow. However I will continue to use the reframing 4K trick inside multicams. I'm comfortable with how PP treats those clips.      ... View more

I've been using AE since the early 2000's and have never seen this issue on AE. Error retrieving frame is normally a timing issue although it can be corrupt files. Anything that is encoded in DV/HDV mts or mpg is automatically suspect. I had an HDV recorder that encoded files that weren't recognised in any NLE and I was on Avid back then. The only solution was to transcode them in a program that didn't spit the dummy.  ... View more

You too Neil - always super helpful and ready to share. ... View more

"Insert or overwrite sequences as nests or individual clips is still set to individual clips, so the Multi-Cam does not cut in correcly.  Instead, since it's being treated like a sequence, all the components of the Multi-Cam are cut into my edit sequence.  This is the opposite of what we want, even though we've changed no settings whatsoever."   The question is why would this be a problem? If you don't want to work this way then you never have to set that switch the other way. Far from causing endless problems it's a one time thing. I've actually found this discussion illuminating. Having how the multicam audio waveforms explained will definitely change my workflow. I'd still like to have access to my source video tracks because I hate having to go into a multicam, introduce a splice at an in point, go back out of the multicam and to the outpoint and back into the multicam, add another splice and then adjust that indivdual part of the multicam, turn off the video tracks I don't want to adjust (as there is no solo on video tracks) and make sure I select the correct track to adjust. However I can (mostly) work around this by ensuring the source multicam is still in 4K. It won't help with random colour corrects but for re-frames should be fine.    BTW while I have your attention any movement on translating LTC to metadata timecode? Make "selection follows playhead" off a sticky selection rather than defaulting to on every time a project restarts? ... View more

Hi Ben - to answer your question I use 4K source media, so I might have two or three cameras at a live event all shooting 4K - I'll go into the multicam source and change it to 1080P and reframe my shots accordingly, duplicating tracks etc. But say none of them end up at 100% and there are two instances during the event where I need that full 4K resolution and need to reframe a shot differently. You lose resolution if you scale up something that has already been scaled down in a nest and of course your edges are cropped unless after re-framing the shot you reset your multicam source back to 4K. One way is to have access to the original source tracks. Simply matchframe the multicam switch to "un-nest" mode and cut the original clip in above the multicam. Hey presto - full 4K to work with as you please in your master, in sync - even track something if you have to. Another reason - colour correction. It's a lot faster to colour correct the cameras to match in the source multicam. You're not dealing with adding multiple colour corrects to the sequence timeline - instead it's all done for you in the multicam. But for some reason one of your operators during the shoot left their white balance on auto and at random spots the camera changed. The ony way to fix it is to go into the multicam, splice around the problem area and change the colour correction - or you could just cut the correct angle in above your sequence and change the colour correction there. Flexibility. That's what it's about. ... View more

I'm still not understanding why Adobe changed the functionality? It wasn't breaking anything and as you said gives a great deal of flexibility and is one of he main reasons I won't switch to Resolve because their multicam is inflexible.  I still want to go into the multicam and create extra video angles. The fact that I can no longer edit the untouched tracks individually into a master sequence is annoying. I can work around it, but I'm still not understanding the necessity to change something that wasn't broken.   Also if I have multiple button on and offs I don't want to end up with a multicam with a hundred different audio channels and a hundred different video tracks. Which is what happens because PP treats each clip as a new video and audio track. Has this behaviour changed? Can I now set everything from the same camera via angle metadata to end up on the same track?    In the project I'm working on the audio recorder had 4 discrete mono tracks (only one of which I needed) and the camera had a stereo audio track. Somehow the resultant multicam only had 4 channels of audio? And it was on automatic. PP seems to automatically assume the camera audio is not needed (it wasn't) because it mutes it in the multicam and does not provide access to it from the source. It's clever but inflexible. ... View more

Sam - if you're having this issue on AE perhaps it's the fact that you're using H264 as your editing format? Could there be a connection? I've never had this issue in AE. Like I said - it's fine if it works for you but it is NOT advice I would be giving to most editors. For a start, I individually colour correct shots and like most editors I do it on the edited material as to do it on the source would be crazy and involve hours of unnecessary work. I also make a living at this so if i have to spend money on storage that's a cost of doing business. Some of the projects I do the client has spent $AUD30K+ just on the shoots. I don't think they'd be too happy if I failed to archive the source material in what is already a compressed codec from the camera.  BTW - I'm not advocating you transcode all your shots into ProRes or DNxHD (although I know some editors who do and swear by it) just the problematic ones that PP and ME are having issues with. You don't even have to archive them with the project. This problem popped up for me and I could not transcode this source material to H264 because all the material is shot in log. And even if I did it might not work because you're swapping Sony or Canon MXF for H264 which still involves decoding GOPs. Both ProRes and DNxHD are intraframe formats - ie they compress individual frames - they don't rely on previous or following frames for information. It's an important distinction. ... View more