Audio Waveform Generation got so much worse in Latest 2025 versions

Can confirm this behavior is still happening in the current 26.02 version--extremely slow generation of the visual waveform with no access to it until it reaches 100% completion.  If I want to know how it is progressing, I still have to open the now relatively massive “Progress” panel which needs to be about 17 times larger than the old lower right .25” by x 1.25”  temporary progress gauge. It can’t be viewed any narrower than 1.5” and can’t be shrunk any shorter than 5 inches tall in any case. 

Hi @ratzo_

Would you mind sharing a video of what you are seeing on your end? We have done internal testing with the new multithreading feature and are seeing faster results, so it would be really helpful to understand what you are experiencing so we can investigate and fix it.

Thanks so much!

I agree. For longer videos, such as livestreams that have sometimes upwards of 10 hours of footage, having to wait for the whole waveform to be generated adds so much dead time to the editing sessions...and having multithread doesn't seem to help much since the time it take to actually show the waveform seems to be considerably longer than on previous versions. I hope they manage a fix to this or at least make it so that we're able to see the waveform as they generate.

Thanks. I ended up reverting to an earlier version and will plan on stay there for the forseeable future.

Thanks for sharing the video, it was very helpful for understanding what you’re seeing.

Waveform generation itself has not become slower between recent versions. When comparing an older release and the current one using the same media and settings, waveform generation completed in the same amount of time, and waveforms appeared progressively as each clip finished processing.

What can impact the total time is when waveform generation runs at the same time as other background analysis tasks. If multiple analyses are happening in parallel, waveform generation may take longer to complete, which can make it feel slower overall.

If you prefer to prioritize waveform generation, you can reduce background processing by turning off Media Intelligence analysis and Audio Tagging in Preferences. This can help waveforms generate more quickly, especially on large or complex projects.

We know this can affect perceived performance, and improving how background analysis tasks is something we’re are looking at. Thanks again for taking the time to report this.

To me it definitely feels slower than before because of this delay showing it.   Here is the screen capture I started partway through the import process--dragged two .wavs totaling 2.5 hours into the 2nd audio track to sync with the original video 's audio, but can't start aligning them (I do it manually because the autosync matching is also slow). The preparing audio sits there at 12% for a long time, then several minutes later the waveform finally draws--you can see I am at 0/00 on the timeline and nothing is there until it finishes. 

https://drive.google.com/file/d/1C00mXcY5xo-8y_ZNmNXF-lhL77eQMKmv/view?usp=drive_link 

Hey @chaunceyc , thanks a lot for taking the time to write this up and for the very clear feedback. I really appreciate you calling out what worked well for you in the previous experience.

One of the main reasons this moved to the Progress panel is that we added multithreaded waveform generation. Under the hood, we are now processing waveforms in parallel, which makes the overall generation significantly faster, especially on longer clips and on machines with more cores. You can enable or verify this in Audio Preferences. With the old UI, we simply did not have a way to accurately represent multiple parallel processes happening at the same time, which is why the Progress panel became the right surface for this work.

That said, the Progress panel is fully closable, so it should not permanently take up screen real estate once you are done checking status. The intent was to make progress more transparent without forcing it to live on screen.

Regarding the behavior where waveforms only appear once the entire analysis is complete, that actually does not sound expected. You should still be able to see waveforms populating as they are generated, rather than having to wait for the full process to finish. If that is not happening on your end, it may very well be a bug. If you are willing, a short screen recording showing that behavior would be incredibly helpful for us to investigate and fix it.

Again, thank you for the candid feedback. It is exactly this kind of input that helps us improve the experience, and I really appreciate you flagging both the workflow impact and the possible regression you are seeing.