From my experience, it works best when it's only one subject speaking. If there are any other people speaking in the background it has a hard time rendering those voices naturally. It also doesn't work well when the room is too echoey. Great tool, but only useful for the right situations.
you have to do some work. nothing has reached the point of one-click to fix everything. i don't think you remember what it was like just a few years ago to fix audio or video.
It would be nice if this feature actually worked in the program. No matter the length, it just keeps getting stuck in the analyzing phase. I wish they would give out the same version of the program to their entire CC user base and not just hand-selecting people to show that it "works." I have to go and use the website, and then it has an upload size limit, smh
Enhance speech is mind-blowing. However, I am finding that exporting out a .wav to Adobe Podcast is much faster than in Premiere Pro. A 25 minute interview took about three minutes in Adobe Podcast vs. ten minutes in Premiere. I liked the Podcast version better. I have been using this frequently, I shoot corporate video in many different acoustical environments, this feature is a whole lot easier than Audition.
I really don't like AI audio - I used the feature a lot in the past but now have become quite sensitive to hearing the AI 'sound' and I don't like it. I would prefer the audio you began with. There are more dramatic circumstances where this AI tool could help, but in your example I would prefer how it sounded out of camera.
This tool has been great no doubt about it, but I've found it to be hit and miss - depending on how bad the original audio is. It can also sometimes cut off words that are said more quietly. But over all its great.
I am on a MAC running Sonoma operating system. Downloaded the presets folder but instructions to open finder and navigate don't apply to this version of operating system. How do I install with Sonoma?
Yep, having the same issue, i previously had the presets installed and now they are gone and the instructions don't help due to no "presets" folder inside the "24" folder
You have to have pretty well-recorded audio to use this tool - otherwise it doesn't work. It can only enhance good audio - it can't repair poor audio. You can only polish a turd so much. The example of the groom reading the letter was pretty well captured audio - If you have something like a speech where you thought feed from the microphone receiver was good, but find out that there was a ton of feedback through the microphone and a lot of high-end frequencies captured, This tool can't really do anything for that. It's just like if you capture footage in low light and it's noisy - artificially increasing the exposure will only make the noise more noticeable. The point You made about capturing audio from 20 ft across the room is completely false - that will sound like complete gobbledygook with made up sounds that the AI program tries to interpret into words. I really appreciate the breakdown, but this tool is not Superman and you have to have decent recordings.
Where I find it comes in handy is for windy audio. I had unusably bad wind noise from a lapel on the beach during a proposal. It completely transformed how the video turned out. I've also used it on certain snippets to remove unwanted sounds. Sometimes that works and sometimes it doesn't. Overall it has a long way to go but I'd say it's a better tool to have than not to have.
@@ShawnCHeron also good at expos to take out background talking but my source mike is the legendary sennheiser reporter which is capturing the talent dialogue well
I use voice enhance in Premier pro but I don’t know why it takes such a long time to finalize. It is so frustrating. And I made a mistake one time of adding it to the main audio and also to a multi camera audio and that messed everything up my computer froze and I couldn’t do anything. Do you know why it takes so long for voice enhance to finalize? I have a very powerful computer as well. It’s been so frustrating that I just end up using voice enhance website instead.
When you still have to do a lot of work with Matt's free, excellent, enhanced speech, Lut, we have a problem. When I play it back, it sounds like the Minions are in my house. It sounds like AI & The Minions are burping, laughing, and having a party in my video. I spent 2 hours trying to fix my audition tape. It was the worst. Welp, back to not using enhanced speech.
I’ve been an evangelist for Adobe (formerly Shasta) Podcast AI. I went on the forums and begged for program integration. I will say, though, that it feels like a different algorithm than the browser-based one. It’s still good, but it feels… way more processed now. I have to turn down the effect quite a bit, but I’m not their target market - we have really nice mics and it has too much compression.
wait...you did a example of bad audio? that "Bad" audio sounds like all it needs is eqing, lol but side note the interesting part of this tool is that if you push it, it starts to add random words that the person didnt say!
Dxrevive does a much better job than premiere pro speech enhance. Davinci Resolve does it better and so final cut pro. Don't hate me, I'm using premiere too 🤣.
Hey Matt, great video. I was just using this feature but for some reason after applying the effect, when ever I made a cut it re analyzes the track. How do I fix this?
My only complaint is if I’m mid enhance & accidentally try to click anywhere before it’s done I will 100% get the spinning wheel of death + computer restart & I’ve got 128 gigs of Ram so 🤷🏻♂️
Please reply me asap 👇🏻 Hi brother please suggest me this thing- i'm a full time designer and complex video editor in future I will use this laptop for after effects works. The video clips I use for editing those are 4k 60FPS h.264/HEVE Currently I'm using Intel windows laptop Now I want to buy 14inch MacBook m3 pro with 11core CPU, 14Core GPU 18GB memory, 512GB SSD now can you please suggest me should I buy this laptop for next 5 years? I can't buy old products Advance thanks 🙏🏻
@@whoismatt so your suggestions it's I can buy MacBook m3 pro & I can edit 10-15 minutes complex videos with this m3 pro laptop smoothly? Right my brother?
This tool is gold for interviews or recordings on Zoom/Teams where the person is far from the mic, in a room with hard surfaces, or there is background noise. I see projects where I want to beg the producer to use Enhance.
Always great stuff Matt! Worth mentioning that with the newest release (vs 24.4 | dated May 14th, 2024) you no longer have to tell Premiere Pro what type of audio it is. We now also have what they're calling "Auto Tagging"; where the AI automatically recognizes a clip as Dialogue, Music, SFX, or Ambiance and adds a new interactive badge as well. I'm sure you recorded this earlier so just thought was worth a mention. Used on a recent wedding video myself and this new feature works so well!
Last time I tried this in premiere, a couple of months ago, I noticed that it sounded like two sound sources that were slightly out of phase. So I went old school and uploaded it to the web version, and blended it there. That sounded good. Have you had any issues with that problem?
Seems very similar to the tools in iZotope. They can work well but have been doing so for some years. Handy to have it built into Premiere though. But AI, in what way? Bit sick of this AI tag on everything, pretty meaningless marketing bs.
actually so helpful and generous nice work king, making a tutorial myself and felt like my audio was lacking but it was just what it needed. actual king shit
Great start, I hope Adobe continues to improve the rough edges: Why a slider with 7.0 as the max? Seems like no sane reason it isn't a 0% to 100% standard dry/wet slider. (yeah, first world complaint). When you apply the effect, after the processing, you can experiment with different slider setting and it doesn't have to reprocess the effect. That is really good. But when you render out the video, it has to re-process the effect each time you render. Also, I am not sure, but various edits to your timeline will "trigger" the effect having to process again. With 20-min to 40-min or longer audio track, processing time is still very significant. I would like to see this effect have an option to "render and replace" as if it was round-tripping out to Audition and back. A workaround I'm using is to do audio processing early in my edit, then render out a full ProRes 422 intermediate file and then import it back into the project for the rest of the edit. Pain of one long rendering task (and large file size on disk) is offset by not being constantly annoyed with unexpected processing delay and the final export is much faster.
I just noticed this option today. I don't usually do this sort of editing of just dialogue, but been doling a lot of interviews for a our theatre after a storm damaged it quite badly. And know learning and playing with these options and saw it about 2 hours ago for the first time. I think it makes the vocals sound too muffled though. Must play with it more though for now.
I shoot video podcasts in a studio with acoustic treatment, podcast XLR mikes and a mackie mixer with great preamps but as you mention still find this tool set at 20% takes the remaining room reverb out. Great video 😎
Thanks for the video. If I already cut up the one file and analyze one of the cut clips in the timeline, can I just copy paste the enhance audio effect on the other clips?
I use enhance audio on speeches and the issue im having is any time someone says something from the crowd that is very low fait. enhance speech boosts that audio also and it sounds very strange and inaudible. how can we stop enhance audio boosting low background voices?
I tried a clip with a lot of pop and 'splosives, and it did a surprisingly good job. There are a LOT of nuance settings to play with, but the one-click got everything in the ballpark. Crazy cool time saver.
I reviewed a motorcycle at a dealership that was right on the highway. People were in the background talking, motorcycles were being started. I used this tool and man did it clean that audio up!
