I should clarify, I only really have experience with French language…sure, some sibilants have to be manually transferred. Still, I would surmise that regardless of language, inhales/ exhales will get treated the as we are all seeing.
Heck, frankly, I have been deep cleaning all the voice remnants back to the Speech layer(s); I was previously creating separate layers for voice/ speech “remnants”, but it is easy unwittingly create an additive condition of too many remnants with that workflow. Results can be excellent; user just needs to understand which layer to set Magic Wand focus to in order to create selections.
We would benefit greatly if we had a selection tool that would create a selection from already unmixed audio layer. At this time, that selection task remains mostly manual. I have been asking for such a functionality since week 2 of using SLP. I do foresee such a tool being preventative to add as I have talked about in the past…I hope to be incorrect.
The thing is - despite SLP would have the power to save me time, it just creates unnecessary manual work.
For example:
I have 6 interviews - with audio tracks for each interviewee and the interviewer.
Makes 12 audio tracks - each 7-12 min long.
I am using the batch mode for Unmix Noisy Speech and let the resulting layers be exported as “trackname - speech.wav” and “trackname - noise.wav”.
Because I can’t trust the Unmix tool I have to control every single audio track for false positives in the noise layer.
So I drag all 12 speech layers into SLP and let it open them as separate instances (My workaround because SLP does not recognize those separated speech and noise wavs as belonging together in a project).
Then drag each of the 12 noise layers individually onto its corresponding speech layer already open.
Then inspect the noise layer visually for those false positives. If there are none I just close that project window and move on to the next one. If there are false positives I have to listen back and compare, then move the parts back to the speech layer.
Then delete the noise layer and export the patched speech layer via mixdown.
(exporting as a layer takes more time because the UI doesn’t remember the settings and can mess with the export file name).
And this is just one workflow part of a more complex workflow. I really wish that SLP gets improved in this part to take away the stupid manual steps whenever possible.
Make my life easier not more complex.
Most of what I do with SL is manual editing.
For me, the most important update would be, being able to use the most basic tools like the gain tool, without it creating artefacts, that I have to edit out again. That would be a huge time saver.
Yepp, 100%
All and every function/module should do exactly what it is supposed to do. Not more, not less. At best no crashes, bugs and glitches all over the place.
And SL is a spectral editor first. For stem seperation, there are better tools and models already available.
So, SL should find back to its roots and it’ll be, by far, the best spetral editor on the market.
Stem separation is an essential part of why I bought SL.
As part of that I do a fair bit of spectral editing too.
If you don’t need Stem separation you can just ignore it but don’t think that just because you don’t need it it is by default redundant.
Sure, stem seperation is very nice to have, no question about that. But, and this is a very big “but”, the basic functions of the complete editing suite should run near 100% flawless and without any odd side effects, glitches or crashes. In that sense, stem seperation is just the cherry on the cake. I refer to the technical implementation and quality of software in general, not this or that nice feature. I do have some strong scientific requirements to fulfill and for this, SL can be expected to be numerical accurate to the bit.
Well, maybe, I am expecting to much and the software is just to cheap to meat the expectations. I would pay 1k or 2k €, if everything would work perfectly in SL. Maybe, it is just to cheap… I don’t know.
I also want stem separation, yet was not the deal-breaker for me when I purchased. For me audio post is king in what I do…I plan to do more music in near future.
Stem separating is not a spectral process.
In that regard, it has been a weight around the neck of the app’s core stability. and hugely escalates the core system requirements for the app.
OK @Robin_Lobel , I’ll add a simple one: contrast of currently selected tool on the Tool Settings bar.
I find it is very difficult to see, at a glance, which setting is currently in focus. I would love to be able to very easily see which setting is currently active/in-focus for these (especially for Replace, Add, Subtract, Intersect). A user set preference highlight color would be greatly well received by me…or just an option for high contrast on the Tool Settings bar would work. As it stands, I’m constantly just setting whatever I need as an additional mouse move because I cannot easily see the current active status.