Bug Report: Auto ducking not generating key frames correctly
Since updating to premier pro 2020 (windows) from 2019 I have experienced the same problem at random in every project I've created.
I create simple projects with a voice over, and a music track and use the auto ducking feature as intended.
I am finding randomly half way through every project I will encounter a point where generating key frames for the background music will ignore the voiceover track. Typically this means it just doesn't duck no matter what you do.
I have tried deleting and re-importing the media to no effect.
It will only be one random clip or some effected.
Most recently I had it in reverse. I have a segment following a voiceover about 20 seconds long with no other audio content other than the music itself, and it just stays ducked no matter what I do. It does not generate the key frames to take it from the ducked state. The same piece if music starts ducking again as it should when the voiceover continues. It's almost like it thinks the in camera audio (which I deleted) is still present.
Unfortunately, we have not been able to reproduce these issues in-house.
Please provide project and media as this will help us reproduce the bug and have a developer investigate/fix it. Please send Dropbox, google drive, creative cloud link to the files to firstname.lastname@example.org attention Michael.
Thanks for your help.
I- TeaM commented
THE MAIN CAUSE OF THIS ISSUE IS FADING OUT IN THE DIALOGUE TRACK LIKE YOU HAVE CUTTED DIALOGUE INTO SECTION NOT A SINGLE TRACK AND ADEED FADE OUT EFFECT IN THE DIALOGUE TRACKS. TO AVOID THIS, SIMPLY NEST THE DIALOGUE TRACKS INTO ONE SINGLE TRACK THEN SELECT THE NESTED DIALOGUE TRACKS AS DIALOGUE. YOU CAN KEEP THE BACKGROUND MUSIC INTO MULTIPLE OR NESTED AND RESELCT THEM AS MUSIC TRACKS AND DUCK AGAINST DIALOGUE. IT WILL WORK.
@Danny Kahler Could you be a little more descriptive on when it fails (or even better, get a video of the failure happening and post it here). I'm unable to replicated a failure like this on my end.
Danny Kahler commented
On a Mac. Same issue. It works at random and decides not to work at random. I've decided not to even show it in my classes until these bugs are fixed.
I'd like to add a few problem solving steps to this, since Premiere doesn't generate an error if no keyframes are added.
1. Always double check that there is something to duck against that is properly tagged. Autoduck works by analyzing the audio levels of the source (usually dialog) and calculating the keyframes from that. So, a simple workflow is to select your dialog, tag it as dialog in the Essential Sound panel, then tag your music, turn on auto-ducking and then press the generate button. Always double check that the "duck against" buttons are lit up correctly (they turn blue when active). That is how the algorithm decides what to use for ducking. You can turn off all of the duck against buttons and the generate keyframes will still work. No keyframes will be generated in that case. Basically, always look for a blue icon under duck against.
2. The ducking algorithm will add fades to the duck. If the fades are larger than the ducking amount, no keyframes will be added. When determining the timing, set the fades relatively short, then increase them for the final product.
3. After any changes the keyframes will need to be regenerated. if you move a clip, or change a slider, press the generate keyframes button again. In Audition the keyframes will auto-generate, in Premiere they will not.
4. Start with very aggressive settings, then back them off. This allows you to determine the timing of the ducking then dial it back so that it sounds good. When starting a new ducking sections, set the Sensitivity and Amount to the maximum (all the way to the right) and the fades to the minimum (all the way to the left) and generate. Then inspect the result, and set the sensitivity down and regenerate until it seems that the sensitivity is correct. Then repeat with the amount. Finally do the same with the fades. Remember to play the autoduck and listen with your ears. You cannot set audio parameters by just looking at your timeline. And if it sounds right it most likely is right. "That looks good" is bad. "That sounds good" is good.
"Bear in mind my dialogue is at -6DB from max so it is more than loud enough to trigger at 6."
It doesn't work that way, though I can understand why you would make that assumption based on the UI. The meters are reading the audio as a peak measurement, the Auto-ducking is measuring the audio using an averaged RMS measurement. The numbers in Auto-duck sensitivity are just for reference, and are not measuring any sort of decibel level. That is why they are unlabeled. Just think of it as more sensitive or less sensitive.
Barbara Briggs-Davies commented
I'm having the same issue. Even if I push sensitivity all the way up to 10, nothing happens - keyframes don't even generate.
Jakub Poniatowski commented
I'm also having a similar issue. The keyframes are being generated but the music level is not changing. Even when i duck by -48, it generates the keyframes but ignoes them during playback.
arthur busco commented
Ducking is not working for me either. I'm using the same preset I use everyday for months, and now it's not working, and when I up the sensitivity, it animates all wrong. It's a bug, and it is making my workflow 2x worse
Victoria Scullion commented
Further info:.. It seems setting the sensitivity of the ducking above the default of 6 causes it to calculate correctly, both in the instance where its not un-ducking when there is no audio track to duck against, and also when its not ducking against dialogue. Bear in mind my dialogue is at -6DB from max so it is more than loud enough to trigger at 6.
I also notice that the key frames calculate way quicker when doing this.