NOTE: dxRevive's Studio algorithm is not a direct equivalent of iZotope RX's dialogue isolate or Waves Clarity Vx. dxRevives Retain algorithm is closer to those others. Just wanted to clarify and note that all but the last sample in this video were processed with the Studio algorithm which attempts to first clean the audio clip and then make it sound like it was recorded in a studio.
iZotope RX 11 should be released very soon. It would be useful to see a comparison between RX, especially RX Advanced, and Accentize. The latter has several components, priced separately with a discounted price tag for the full bundle of US$837. Currently, iZotope RX Advanced is $1200. Standard is $400. Accenture's deRevive Pro component, purchased by itself, is $300.
Thanks Curtis, here's a detailed comment. Grab a sandwich. 😁 I do TV/Film post and have the arsenal of the older Waves X and Z restoration tools and the Izotope RX post tools as well now as Clarity VX. I've recently been trying out Clear, Revive Pro and Hush. These last three are what my post prod friends at Warner Brothers are using for all the content coming out of there (TV/Film) the past many months. Neural Network type voice de-noising is changing so quickly now but I've found there can't and won't be an overall "winner" because the number of layers used in the neural net deep learning and the number of neurons and how the algorithms are built completely affect how each of these behaves/responds to various content. It all completely depends on the content as to which Neural Net recipe will work the best for each different audio file. The one thing that the current update of Revive does that no one else offers is the re-synthesis of frequencies that should be there but are missing. For example, the file I'm testing RevivePro on right now has horrible hum. Revive didn't do a good job at all on it, but Wave hum removal removed the hum completely and without leaving any artifacts behind. But this old 8mm recording of an off broadway production in an echoey theater sounded really thin with a bunch of other noise issues. Revive did an amazing job restoring the sound of what the voice likely sounded more like to the audience (noise removal and actual voice restoration including restoring fullness and closeness of the voice). But, where all AI noise removal falls short and fails in right now is music. There just isn't anyone who has created an AI noise removal option for music because there simply hasn't been enough AI training in music to know the difference between noise and overtones and harmonic content etc. So, currently as of March 1st 2024, Spectral de-noising (like RX or Waves or Cedar etc) are still the best option for removing noise after feeding is a sample of the noise you are wanting to remove. BUT, spectral noise removal leaves horribly noticeable artifacts behind if you push it too hard or the wrong way. Final thought for now is fortunately but also unfortunately, each of these plugins I listed are amazing and where one is completely weak another saves the day, BUT I just don't want to spend $300 + $200 + 200 + $150 etc etc on RevivePro and other options when it really "takes a village" and several are really needed in the toolbox so when one doesn't do the job, you've got another ready to try. Clear is much better at separating reverb and noise separately from the voice. Well enough that Clarity deverb most likely isn't needed. Others are equally good at removing reverb (without having to only offer deverbing) ClarityVX has been awesome and I use it on almost every project with dialogue, but I'm also still using about $1200 of other tools ALSO.. If I buy RevivePro, I will be using a new $300 plugin mostly to restore fullness of dialogue but will also still have to use RX for de-click, dx isolate, decrackle, spectral repair to remove birds and whistles and wind etc), but my big complaint with Revive is currently you can't turn off the synthesis revitalizing engine and simply remove noise. That's actually a big deal, because sometimes I only want to remove noise but retain the character of the voice. The re-synthesis comes along for the ride. A quandary because I know in another short 6 months something else will come out that likely combines a bunch of capabilities and offers more independent control over features and then I will look at the hundreds of $$ that I can't ever get back and then no budget left for the upgrade or switch to the better "generalist" option. Thanks again for your video! Alex
One thing I noticed pretty early on is that although the noise floor drops significantly, the voice gets very garbely, almost wobbly. I'd say the built in voice isolation in resolve is going to be a bit cleaner, although it too has a limit. I hope black magic improves this even more in time.
I was using compressor, de-esser and EQ for year of dealing with home recording stuff. Those plugins are doing great job, but it is also fun to do it manually. It helps to know what those plugins are replicating and replacing. Also helps with identifying problems at recording time and not until post.
This demo is not using dxRevive for what its intended -- dxrevive is not a noise reduction tool - it is a tool that analyzes and restores crappy audio. I had some dialogue which I tried to save with some heavy izotoping. I was able to clean it some but it was just a bad recording - not useable. Then I hit the izotoped line with 75 % dxrevive - and like magic - dxrevive created a new line which translated the weak recording into a useable recording - and it worked well. Is it studio quality? No - but it is now USEABLE which is good since the actress is not available for ADR. Apparently the program analyzes the vocal portion of the sound and creates a new recording with a computer processed version of what it thinks the vocal should sound like. Very cool - especially to those of us who used to work with analog dialogue recordings before digital editing was born.
The company then needs to be much clearer about the purpose of this plugin. I took them at their word and the first two things they list on their website are noise removal and reverb suppression. www.accentize.com/dxrevive/
Great demos, thanks. I'm definitely noticing a rise in these type of plugins and excited for how good they are getting. Some time soon, creators will have no idea of the pains we went through to repair, less than ideal, audio lol. Stay well, Curtis!
Thanks Ray! I'm finding them useful for constant fan type noise (which is plentiful in many non-studio environments). Stay well and great to hear from you!
I don’t know about this particular one, but I find a lot of the time these tools each have their strengths and so it’s a matter of the correct tool for the job. A lot of the time it can be as simple as making the noise sound pleasing rather than trying to remove it. I think AI voice regeneration can be useful as well.
I'm glad to see that the most important audio engineer rule still applies: Garbage in: Garbage out. This helps, and I'm quite impressed and would be thankful to have it, but it isn't perfect.
Having been in the forensic audio realm for the last 25 years, it is amazing what has emerged recently in the arena of clarification/noise removal. RX has become the standard for our industry, replacing packages that were $20,000. Others are now joining in and even improving on it at a fraction of what we used to spend.
Thanks for the video. The de-reverb sounds pretty good. I would have liked to see how well it works against more extreme noise and more extreme reverb.
@@curtisjudd Lots of people use bad microphones or very distant microphones built into their camcorder or phone. So the reverb is often much worse than what you demoed here.
@@curtisjudd I am not the one doing this bad audio. I know how to get good audio at the source. The issue is that I am often tasked with fixing other people's stuff.
This seems similar to the "studio" effect in Descript that is pretty good at cleaning up audio from webcams/zoom recordings but definitely isn't perfect. When I was producing podcasts with remote guests, that was pretty useful, and something similar that's a standalone plugin is definitely handy.
Presumably adaptive, given it has no explicit (user-visible) "Learn" phase. In which case, several questions arise: 1) How quickly does it (_initially_) adapt/match (presumably to patterns it was trained on during development)? 2) Or does it perform some kind of (hidden) seek-ahead (enabling apparently immediate adaption)? 3) How quickly/well does it adapt to _changes within_ a clip/track, e.g. different person speaking or switching same person to a different microphone, or an event P.A. ("public address") system (mic/amplifier/speaker) suddenly failing (hence camera/recordist microphone suddenly only receiving presenter's voice direct-from-mouth at-a-distance). (I do realise such sections could/should be treated separately - I'm only interested in "kicking the tyres" here, getting a feeling for what this plugin is exactly and what it can/can't do).
I've noticed there was a proliferate use of AI speech enhance (probably Adobe Speech Enhance specifically) in the last few months among podcasters, especially when interviewing someone over Zoom with poor notebook camera and mic. I recognize immediately the AI artifacts and it drives me nuts. I doubt whether such an AI processed interview is more pleasant to listen to instead of the typical poor mic quality we got used to over the last several years. Perhaps normies do find it better, I don't
My take is to only use these to "take the edge off". When pushed, they all sound horrible in a different way than the originals. And, as we hear here, it is still best to get the recording right WHILE recording.
I've been noticing it too. I've used the Adobe product at times when a client hands me really bad audio. But, over time, I'm losing my taste for it. It often mixes up 't' and 's' and starts to make it hard to understand what people are saying. I am finding that I prefer to clean up stuff mostly in RX10 where I have control over how the audio is getting changed.
Like all non soundies. Podcasters, video editors RUclipsrs et al will take audio tools and just cook everything with them. These tools are meant for us who know how and when to use them.
Davinci's Voice Isolation and Dialogue Leveler have been suiting me much better than iZotope RX. I think this dxRevive tool would be nice to add into the pipeline for a cherry on top, but I'd want Davinci to do the heavy lifting. 90% for Dialogue Leveler has been incredible with little artifacts and all noise gone.
In a workflow, could presumably apply on its own for productivity or as a refinement (final polish-up/clarify) stage, following heavy-lifting by non-synthesising processes (De-Reverb/Room, De-Hum/Notches, EQ). Experiments/demos of its benefit in such a refinement stage would be interesting.
Ah I see that @user-ik8vy1rg8f (2 weeks ago, as of this post) liked the idea of a "heavy lifting" initial stage of processing. And Curtis agreed with that comment.
Doesn't seem any better than the built-in FCP one, and sometimes worse. It's not doing any "AI" like the adobe one which tries to "re-buid" the vocal rather than just de-noise. I'm not sure what this really offers. "Retain" was the only example that kinda worked.
I used it on a film recently. In one scene the boom mic was too far away from the talent which can happen because of Boom Shadow forcing the boom op to pull away. Revive fixed it brilliantly. Curtis' demonstration here was extreme and you'll have artifacts if the voice is too thin or too far away. But it is brilliant at restoring Dialogue where the boom is a few feet further away than is ideal and you start hearing the room. A good alternative from switching over to a lav in these situations. Like most audio tools it's not intended to fix crap sound. But rather make usable audio sound great.
I like its use as a tool when I need to remove/attenuate a noise that was masking a chunk dialogue frequencies. It's "reintegrating missing frequency components" as well as separating noise from the dialogue. Sort of like a RX Dialogue Isolate and Spectral Recovery tool all in one. I've been using it in the RX editor with the frequency selection tools. Any thoughts on the newish Cedar VoicEX denoiser?
Love this channel as always - BUT - FYI .. No IOS for DVR 18.6 users. This plugin like most, only works in OS / Windows. Pointless for future editors on Ipad 12.9 M2. Get with the program & give us IOS versions.
There is one that is even better, but it's nowhere near as known. It's the Unmix Noisy Speech in Spectralayers 10. It's way beyond what any other VST plugin does. But it's a bit capricious, ignoring some small parts of the vocal information out of the blue. @@curtisjudd @darkechoproductionsllc9559
I'm currently experimenting with dxRevive, based on an event videography session where my gun mic failed and only the camera INT mic recording is available. In that test, dxRevive essentially re-created the INT Mic compensation EQ profile that I previously established manually from overall audience clapping (BTW: anyone got links to using that technique? - first time I've guessed&used it), just with more (and possibly excessive) spectral domain detail.
how does it works in other languages? I understand this a english spoken channel and most fo the solutions are designed that way, but... can it work in spanish? It would be nice if you could acquire a sample audio to test it!
LOL good point, I recently used Adobe AI Enhance Tool for a movie in Chinese, it might sound like English or Russian if the audio clip is not that clear. I call it AI-generated ART
I don't like what it does the treble in voices. Makes them sound wobbly and metallic. Both Resolve, Go-Yo and Clarity do a much better job in that regard. When it comes to super old, distorted recordings, it does a great job on the other hand.
I thought adobe podcast (now enhance) was amazing when it first came out. Now that it's out of beta (and for the last month or so), the quality has nosedived. Its practically useless at the mo. I'm about to hop off to see what this one costs
@@curtisjudd SpectraLayers Unmix Noisy Speech is incomparably more powerful than anything else with faint speech in really noisy situations. But it's also very inconsistent, and you need to work on the separated noise layer when it fails and bring some ignored chunks back to the speech layer with Acon or Waves. If they make it act consistent, it will be hands down the best tool on the market for dialogue separation. Waves is finnicky too, and there's always a tradeoff between Broad 1 and 2. Sometimes I combine them and redo the EQ. Acon is very useful because it's rather consistent, but it's also aggressive and very blunt, leaving out spectral holes. What was once the crown jewel, RX dialogue isolate is the least used for me, by far.
@@cosmingurau my further experience is, acon and waves are the same (so-call AI-) level, acon is more cpu friendly, sometimes waves is better depending on the audio clip. They are all better than izotope（my rx version is 9 btw, never use rx for dialogue separation after clarity was released🤣). Quite often, Adobe Podcast AI Tool could be a miracle.
Nope. RX Advanced is superior but is a bit more than the $800 bundle these developers are asking. For the price, I would get RX on sale and have superior control.
LOL! Good on you! dxRevive seems like it would be an easy solution if you're starting with a pretty good, clean recording already and want to quickly make it sound like a studio recording.
Next level? Nah, i'm not impressed at all! There are some great tools coming out all the time and this has the same level of restoration as my RX7 from 5 years ago. Sounds like a step back! Makes every voice flutter and robotic when cranked up above 60%. Waves clarity vx makes this just look dumb..
@@curtisjudd I saw a review of the pro version where they used the frequency band specific filtering and there you can get some very good results, even on things where you say it doesn't do a real good job. It's the pro tools expert review. If you didn't see it yet, you should. I don't say this to critic you, just to add a bit of extra info
@@eneasmentzel9758 Thanks for the note! Yes, that works in *some* cases where the noise is within a limited frequency range. Sadly, many noise sources are more broadband.
NOTE: dxRevive's Studio algorithm is not a direct equivalent of iZotope RX's dialogue isolate or Waves Clarity Vx. dxRevives Retain algorithm is closer to those others.
Just wanted to clarify and note that all but the last sample in this video were processed with the Studio algorithm which attempts to first clean the audio clip and then make it sound like it was recorded in a studio.
Impressive!
iZotope RX 11 should be released very soon. It would be useful to see a comparison between RX, especially RX Advanced, and Accentize. The latter has several components, priced separately with a discounted price tag for the full bundle of US$837. Currently, iZotope RX Advanced is $1200. Standard is $400. Accenture's deRevive Pro component, purchased by itself, is $300.
👍
As always, this was nice, Curtis! Great to know where things are currently in the realm of Ai improved Audio. Thank you!
🙏
Thanks Curtis, here's a detailed comment. Grab a sandwich. 😁
I do TV/Film post and have the arsenal of the older Waves X and Z restoration tools and the Izotope RX post tools as well now as Clarity VX. I've recently been trying out Clear, Revive Pro and Hush. These last three are what my post prod friends at Warner Brothers are using for all the content coming out of there (TV/Film) the past many months. Neural Network type voice de-noising is changing so quickly now but I've found there can't and won't be an overall "winner" because the number of layers used in the neural net deep learning and the number of neurons and how the algorithms are built completely affect how each of these behaves/responds to various content. It all completely depends on the content as to which Neural Net recipe will work the best for each different audio file.
The one thing that the current update of Revive does that no one else offers is the re-synthesis of frequencies that should be there but are missing. For example, the file I'm testing RevivePro on right now has horrible hum. Revive didn't do a good job at all on it, but Wave hum removal removed the hum completely and without leaving any artifacts behind. But this old 8mm recording of an off broadway production in an echoey theater sounded really thin with a bunch of other noise issues. Revive did an amazing job restoring the sound of what the voice likely sounded more like to the audience (noise removal and actual voice restoration including restoring fullness and closeness of the voice). But, where all AI noise removal falls short and fails in right now is music. There just isn't anyone who has created an AI noise removal option for music because there simply hasn't been enough AI training in music to know the difference between noise and overtones and harmonic content etc.
So, currently as of March 1st 2024, Spectral de-noising (like RX or Waves or Cedar etc) are still the best option for removing noise after feeding is a sample of the noise you are wanting to remove. BUT, spectral noise removal leaves horribly noticeable artifacts behind if you push it too hard or the wrong way.
Final thought for now is fortunately but also unfortunately, each of these plugins I listed are amazing and where one is completely weak another saves the day, BUT I just don't want to spend $300 + $200 + 200 + $150 etc etc on RevivePro and other options when it really "takes a village" and several are really needed in the toolbox so when one doesn't do the job, you've got another ready to try.
Clear is much better at separating reverb and noise separately from the voice. Well enough that Clarity deverb most likely isn't needed. Others are equally good at removing reverb (without having to only offer deverbing) ClarityVX has been awesome and I use it on almost every project with dialogue, but I'm also still using about $1200 of other tools ALSO.. If I buy RevivePro, I will be using a new $300 plugin mostly to restore fullness of dialogue but will also still have to use RX for de-click, dx isolate, decrackle, spectral repair to remove birds and whistles and wind etc), but my big complaint with Revive is currently you can't turn off the synthesis revitalizing engine and simply remove noise. That's actually a big deal, because sometimes I only want to remove noise but retain the character of the voice. The re-synthesis comes along for the ride. A quandary because I know in another short 6 months something else will come out that likely combines a bunch of capabilities and offers more independent control over features and then I will look at the hundreds of $$ that I can't ever get back and then no budget left for the upgrade or switch to the better "generalist" option.
Thanks again for your video!
Alex
One thing I noticed pretty early on is that although the noise floor drops significantly, the voice gets very garbely, almost wobbly. I'd say the built in voice isolation in resolve is going to be a bit cleaner, although it too has a limit. I hope black magic improves this even more in time.
I hope the Resolve plugin gets better. It didn’t do better than dxRevive in our tests.
I was using compressor, de-esser and EQ for year of dealing with home recording stuff. Those plugins are doing great job, but it is also fun to do it manually. It helps to know what those plugins are replicating and replacing. Also helps with identifying problems at recording time and not until post.
I agree. 👍
This demo is not using dxRevive for what its intended -- dxrevive is not a noise reduction tool - it is a tool that analyzes and restores crappy audio. I had some dialogue which I tried to save with some heavy izotoping. I was able to clean it some but it was just a bad recording - not useable. Then I hit the izotoped line with 75 % dxrevive - and like magic - dxrevive created a new line which translated the weak recording into a useable recording - and it worked well. Is it studio quality? No - but it is now USEABLE which is good since the actress is not available for ADR. Apparently the program analyzes the vocal portion of the sound and creates a new recording with a computer processed version of what it thinks the vocal should sound like. Very cool - especially to those of us who used to work with analog dialogue recordings before digital editing was born.
The company then needs to be much clearer about the purpose of this plugin. I took them at their word and the first two things they list on their website are noise removal and reverb suppression. www.accentize.com/dxrevive/
Great demos, thanks. I'm definitely noticing a rise in these type of plugins and excited for how good they are getting. Some time soon, creators will have no idea of the pains we went through to repair, less than ideal, audio lol. Stay well, Curtis!
Thanks Ray! I'm finding them useful for constant fan type noise (which is plentiful in many non-studio environments). Stay well and great to hear from you!
I don’t know about this particular one, but I find a lot of the time these tools each have their strengths and so it’s a matter of the correct tool for the job. A lot of the time it can be as simple as making the noise sound pleasing rather than trying to remove it. I think AI voice regeneration can be useful as well.
Completely agree.
I'm glad to see that the most important audio engineer rule still applies: Garbage in: Garbage out. This helps, and I'm quite impressed and would be thankful to have it, but it isn't perfect.
Helpful but not perfect - good way to sum it up.
Having been in the forensic audio realm for the last 25 years, it is amazing what has emerged recently in the arena of clarification/noise removal. RX has become the standard for our industry, replacing packages that were $20,000. Others are now joining in and even improving on it at a fraction of what we used to spend.
Indeed!
I've been doing so much audio restoration for short films that I've wondered if I could actually get into audio forensics.
Very nice review. Excellent job picking out some of the typical problems and testing it on them.
Thanks Gregg.
Thanks for the video. The de-reverb sounds pretty good.
I would have liked to see how well it works against more extreme noise and more extreme reverb.
Wow, that was pretty intense reverb.
@@curtisjudd Lots of people use bad microphones or very distant microphones built into their camcorder or phone. So the reverb is often much worse than what you demoed here.
@@GeorgeOu the fix it in post mindset which is pretty risky if one is working for clients or
Is some other professional capacity.
@@curtisjudd I am not the one doing this bad audio. I know how to get good audio at the source. The issue is that I am often tasked with fixing other people's stuff.
@@GeorgeOu I feel that and have been there. How do you handle the situation when there just isn't a post solution?
This seems similar to the "studio" effect in Descript that is pretty good at cleaning up audio from webcams/zoom recordings but definitely isn't perfect. When I was producing podcasts with remote guests, that was pretty useful, and something similar that's a standalone plugin is definitely handy.
👍
Presumably adaptive, given it has no explicit (user-visible) "Learn" phase. In which case, several questions arise: 1) How quickly does it (_initially_) adapt/match (presumably to patterns it was trained on during development)? 2) Or does it perform some kind of (hidden) seek-ahead (enabling apparently immediate adaption)? 3) How quickly/well does it adapt to _changes within_ a clip/track, e.g. different person speaking or switching same person to a different microphone, or an event P.A. ("public address") system (mic/amplifier/speaker) suddenly failing (hence camera/recordist microphone suddenly only receiving presenter's voice direct-from-mouth at-a-distance). (I do realise such sections could/should be treated separately - I'm only interested in "kicking the tyres" here, getting a feeling for what this plugin is exactly and what it can/can't do).
Correct, fully adaptive.
Thanks so much for this. it's a pity the pro is such a jump in price
You’re welcome.
Thank you, Curtis.
👍
I've noticed there was a proliferate use of AI speech enhance (probably Adobe Speech Enhance specifically) in the last few months among podcasters, especially when interviewing someone over Zoom with poor notebook camera and mic. I recognize immediately the AI artifacts and it drives me nuts. I doubt whether such an AI processed interview is more pleasant to listen to instead of the typical poor mic quality we got used to over the last several years. Perhaps normies do find it better, I don't
My take is to only use these to "take the edge off". When pushed, they all sound horrible in a different way than the originals. And, as we hear here, it is still best to get the recording right WHILE recording.
I've been noticing it too. I've used the Adobe product at times when a client hands me really bad audio. But, over time, I'm losing my taste for it. It often mixes up 't' and 's' and starts to make it hard to understand what people are saying. I am finding that I prefer to clean up stuff mostly in RX10 where I have control over how the audio is getting changed.
Like all non soundies. Podcasters, video editors RUclipsrs et al will take audio tools and just cook everything with them. These tools are meant for us who know how and when to use them.
Davinci's Voice Isolation and Dialogue Leveler have been suiting me much better than iZotope RX. I think this dxRevive tool would be nice to add into the pipeline for a cherry on top, but I'd want Davinci to do the heavy lifting. 90% for Dialogue Leveler has been incredible with little artifacts and all noise gone.
👍
In a workflow, could presumably apply on its own for productivity or as a refinement (final polish-up/clarify) stage, following heavy-lifting by non-synthesising processes (De-Reverb/Room, De-Hum/Notches, EQ). Experiments/demos of its benefit in such a refinement stage would be interesting.
Ah I see that @user-ik8vy1rg8f (2 weeks ago, as of this post) liked the idea of a "heavy lifting" initial stage of processing. And Curtis agreed with that comment.
Yes
wow, also any plans for a review of the Hollyland Lark Max?
Yes!
Hi! Thank you for your test! What about BFX CrumplePop?
Haven’t tried it yet.
Doesn't seem any better than the built-in FCP one, and sometimes worse. It's not doing any "AI" like the adobe one which tries to "re-buid" the vocal rather than just de-noise. I'm not sure what this really offers. "Retain" was the only example that kinda worked.
It is definitely doing some re-synthesis on the watch recording.
I used it on a film recently. In one scene the boom mic was too far away from the talent which can happen because of Boom Shadow forcing the boom op to pull away. Revive fixed it brilliantly. Curtis' demonstration here was extreme and you'll have artifacts if the voice is too thin or too far away. But it is brilliant at restoring Dialogue where the boom is a few feet further away than is ideal and you start hearing the room. A good alternative from switching over to a lav in these situations. Like most audio tools it's not intended to fix crap sound. But rather make usable audio sound great.
I like its use as a tool when I need to remove/attenuate a noise that was masking a chunk dialogue frequencies. It's "reintegrating missing frequency components" as well as separating noise from the dialogue. Sort of like a RX Dialogue Isolate and Spectral Recovery tool all in one. I've been using it in the RX editor with the frequency selection tools.
Any thoughts on the newish Cedar VoicEX denoiser?
Haven’t tried Cedar yet, but would like to try!
Love this channel as always - BUT - FYI .. No IOS for DVR 18.6 users. This plugin like most, only works in OS / Windows. Pointless for future editors on Ipad 12.9 M2. Get with the program & give us IOS versions.
Hopefully Accenture see's your request here. I, myself, don't edit on an iPad yet, but more probably will as time marches on.
Wonder how does it compare to their realtime solutions for voice
Slightly better for more use cases.
Clarity is the best on yet from waves
Mmm, came out about the same in my tests.
@curtisjudd I used on my show that's on Amazon prime, Dark Echoes paranormal. Works great
There is one that is even better, but it's nowhere near as known. It's the Unmix Noisy Speech in Spectralayers 10. It's way beyond what any other VST plugin does. But it's a bit capricious, ignoring some small parts of the vocal information out of the blue. @@curtisjudd @darkechoproductionsllc9559
I'm currently experimenting with dxRevive, based on an event videography session where my gun mic failed and only the camera INT mic recording is available. In that test, dxRevive essentially re-created the INT Mic compensation EQ profile that I previously established manually from overall audience clapping (BTW: anyone got links to using that technique? - first time I've guessed&used it), just with more (and possibly excessive) spectral domain detail.
Good luck. 👍
how does it works in other languages? I understand this a english spoken channel and most fo the solutions are designed that way, but... can it work in spanish? It would be nice if you could acquire a sample audio to test it!
Accentize claims it also has been trained with multiple languages. Check out their site for details.
LOL good point, I recently used Adobe AI Enhance Tool for a movie in Chinese, it might sound like English or Russian if the audio clip is not that clear. I call it AI-generated ART
Would have likes to hear a sample.of lav mic clothes rustling during dialogue
Doesn't do a great job with that.
how can i fix a shitty zoom recording? I have struggle with that. dxRevive sounds weird on that unfortunately
Try Adobe enhance and see if that’s any better? But my guess is it won’t be a lot better.
I don't like what it does the treble in voices. Makes them sound wobbly and metallic.
Both Resolve, Go-Yo and Clarity do a much better job in that regard.
When it comes to super old, distorted recordings, it does a great job on the other hand.
The retain algorithm did well in regard to the higher frequencies.
I thought adobe podcast (now enhance) was amazing when it first came out. Now that it's out of beta (and for the last month or so), the quality has nosedived. Its practically useless at the mo. I'm about to hop off to see what this one costs
👍
I would not expect a Restorator to act as an Isolator.
It allegedly does both.
acon dialogue> waves clarity>izotope>this company
🤷
no. it's spectralayers > waves clarity vx > acon dialogue extract > izotope rx dialogue isolate
@@cosmingurau Definitively? My experience is that it depends on the audio clip.
@@curtisjudd SpectraLayers Unmix Noisy Speech is incomparably more powerful than anything else with faint speech in really noisy situations. But it's also very inconsistent, and you need to work on the separated noise layer when it fails and bring some ignored chunks back to the speech layer with Acon or Waves. If they make it act consistent, it will be hands down the best tool on the market for dialogue separation. Waves is finnicky too, and there's always a tradeoff between Broad 1 and 2. Sometimes I combine them and redo the EQ. Acon is very useful because it's rather consistent, but it's also aggressive and very blunt, leaving out spectral holes. What was once the crown jewel, RX dialogue isolate is the least used for me, by far.
@@cosmingurau my further experience is, acon and waves are the same (so-call AI-) level, acon is more cpu friendly, sometimes waves is better depending on the audio clip. They are all better than izotope（my rx version is 9 btw, never use rx for dialogue separation after clarity was released🤣). Quite often, Adobe Podcast AI Tool could be a miracle.
Nope. RX Advanced is superior but is a bit more than the $800 bundle these developers are asking. For the price, I would get RX on sale and have superior control.
LOL! Good on you!
dxRevive seems like it would be an easy solution if you're starting with a pretty good, clean recording already and want to quickly make it sound like a studio recording.
Boooo! Apple/iOS only! No windows version available
🙁
Next level? Nah, i'm not impressed at all! There are some great tools coming out all the time and this has the same level of restoration as my RX7 from 5 years ago. Sounds like a step back!
Makes every voice flutter and robotic when cranked up above 60%. Waves clarity vx makes this just look dumb..
The retain algorithm is more like the tools you mentioned.
@@curtisjudd I saw a review of the pro version where they used the frequency band specific filtering and there you can get some very good results, even on things where you say it doesn't do a real good job. It's the pro tools expert review. If you didn't see it yet, you should. I don't say this to critic you, just to add a bit of extra info
@@eneasmentzel9758 Thanks for the note! Yes, that works in *some* cases where the noise is within a limited frequency range. Sadly, many noise sources are more broadband.