RX elements

[Stratman57]

I've just seen that RX Elements 11 is on sale with a 40% discount, and I'm wondering if it would give me a step up from the noise reduction/restoration than the tools I have in Adobe Audition 3.0.

Regards, Simon.

[IAA]

Well I've never used Auditon but I do use RX for cleaning up samples when I'm creating instruments and it works marvellously, certainly enough for me. Its also easy to use which is bonus when you're not sure what you're doing , I just close my eyes and twiddle - as they say and it works. no major artefacts or stuff. in one of my samples of a note on a piano i had recorded - (which I sampled whilst it was in a garage) a bird burst into song just on the sample decay - and it cleaned it up just fine.

Ian

[James Perrett]

I've just seen that RX Elements 11 is on sale with a 40% discount, and I'm wondering if it would give me a step up from the noise reduction/restoration than the tools I have in Adobe Audition 3.0.

The only improvement would be in the noise reduction. They seem to have stripped Elements right back so that you don't even get the audio editor these days and the current price is nearly double what I paid for Elements.

When it comes to click/pop removal Audition is much better, the clip removal is of a similar standard and you can get better de-reverb algorithms for free. If you really want to look at an alternative to Audition then Acon Acoustica would offer far more than RX Elements. RX only makes sense if you are prepared to go for the full version - and even then, Audition still wins out in some areas.

[alexis]

I just started a journey today using Steinberg's SpectraLayers (full version) well-regarded in SOS review pages IIRC.

It almost smoked my CPU (geriatric computer here), but it did a great job on its first test - stripping the vox out of a YouTube 2- track full song.

Looking forward to more time with it!

[James Perrett]

It almost smoked my CPU (geriatric computer here), but it did a great job on its first test - stripping the vox out of a YouTube 2- track full song.

You don't have to pay for SpectraLayers to do that - some of the best tools for vocal removal and stem separation are free.

[alexis]

It almost smoked my CPU (geriatric computer here), but it did a great job on its first test - stripping the vox out of a YouTube 2- track full song.

You don't have to pay for SpectraLayers to do that - some of the best tools for vocal removal and stem separation are free.

Right, thanks, James. That's just the first step in this journey

[James Perrett]

Right, thanks, James. That's just the first step in this journey

If you want to do stem separation then grab a copy of the Ultimate Vocal Remover

https://ultimatevocalremover.com/

which offers a variety of separation algorithms and also has a great de-reverb algorithm (far better than RX) although I think you have to download the de-reverb model separately. There's an option in the program that allows you to download new models.

[Stratman57]

Thanks for the replies, one of the things I'm interested in is the removal of room sound, which Audition 3.0 doesn't have. My current project is a "podcast" recording of several young people, whose families are refugees/immigrants to Wales. They told their stories to the writer, who then produced a script from their stories. We then did a round table recording of the young people reading the script as a guide, the plan being they come to another session where I recorded them individually.
Unfortunately not all of them turned up for the individual recording session.
So I have to use some of the lines from the round table recording, which some of the young people were not very loud, resulting in quite a lot of room sound when bringing the volume up to the levels of the individual recordings.

Regards, Simon.

[alexis]

Right, thanks, James. That's just the first step in this journey

If you want to do stem separation then grab a copy of the Ultimate Vocal Remover

https://ultimatevocalremover.com/

which offers a variety of separation algorithms and also has a great de-reverb algorithm (far better than RX) although I think you have to download the de-reverb model separately. There's an option in the program that allows you to download new models.

Thanks much again, James. I bought SpectraLayers a while ago, enjoying learning what it can do (far more than mix separation, if course)

[Tomás Mulcahy]

SpectraLayers sounds and looks like it is using the DeMucs algorithm, which is already free in UVR. I suspect Izoptope is using it as well, maybe with some tweaks.

It does not sound nearly as good to me as VB Splitz which you can use for a small fee on audiostrip.co.uk DeMucs does claim to separate BVs piano and guitar etc. but I've had zero success with that. It does a bit but the output has a lot of dropouts and warbles so is unuseable.

[James Perrett]

DeMucs does claim to separate BVs piano and guitar etc. but I've had zero success with that. It does a bit but the output has a lot of dropouts and warbles so is unuseable.

Are you sure that you are talking about Demucs? It doesn't claim to separate out backing vocals from lead vocals and the piano separation has always been "experimental" i.e. it rarely works, but I find it works well on guitar, bass, drums and vocals. Certainly far better than the Spleeter algorithm which is what RX used up until the latest version.

I'll have to check out the VB Splitz algorithm and see if it can be used with UVR.

[Tomás Mulcahy]

DeMucs does claim to separate BVs piano and guitar etc. but I've had zero success with that. It does a bit but the output has a lot of dropouts and warbles so is unuseable.

Are you sure that you are talking about Demucs? It doesn't claim to separate out backing vocals from lead vocals and the piano separation has always been "experimental" i.e. it rarely works, but I find it works well on guitar, bass, drums and vocals. Certainly far better than the Spleeter algorithm which is what RX used up until the latest version.

I'll have to check out the VB Splitz algorithm and see if it can be used with UVR.

Yes you're right, I mixed up the functionality. Let us know what you discover. AFAIK VB-Splitz is exclusive to Audiostrip? But I have been wrong before

[Matt Houghton]

Thanks for the replies, one of the things I'm interested in is the removal of room sound, which Audition 3.0 doesn't have. My current project is a "podcast" recording of several young people, whose families are refugees/immigrants to Wales. They told their stories to the writer, who then produced a script from their stories. We then did a round table recording of the young people reading the script as a guide, the plan being they come to another session where I recorded them individually.
Unfortunately not all of them turned up for the individual recording session.
So I have to use some of the lines from the round table recording, which some of the young people were not very loud, resulting in quite a lot of room sound when bringing the volume up to the levels of the individual recordings.

Regards, Simon.

Accentize do some great tools for this — ie specifically for cleaning up dialogue recordings. DxRevive might be all you need to make things workable. DeRoom Pro is slightly better at room removal. Chameleon allows you to impose room characteristics of one voice on to others. Not cheap but worth every penny.

You might also want to check out the Studio Sound feature in Descript. Last I checked, the automatic editing features in Descript were still a bit too off for serious use, particularly in terms of identifying breaths and 'erms' and making clumsy edits, but the sound processing isn't at all bad, and you can render the Studio Sound effect and import into your editor of choice, where you can blend to taste with the original as a sort of wet/dry control (meaning you don't have to choose the wet/dry setting when you render so can tweak that balance after your edits).

[James Perrett]

So I have to use some of the lines from the round table recording, which some of the young people were not very loud, resulting in quite a lot of room sound when bringing the volume up to the levels of the individual recordings.

If you are talking about room reverb then I'd definitely try the reverb removal available for UVR first. I've never had much luck with RX's De-Reverb tool.

One tip though - UVR gives you both a dry file and a reverb file. I find that adding a little of the reverb file back into the mix makes things sound a little more natural.

[alexis]

Thanks for the replies, one of the things I'm interested in is the removal of room sound, which Audition 3.0 doesn't have. My current project is a "podcast" recording of several young people, whose families are refugees/immigrants to Wales. They told their stories to the writer, who then produced a script from their stories. We then did a round table recording of the young people reading the script as a guide, the plan being they come to another session where I recorded them individually.
Unfortunately not all of them turned up for the individual recording session.
So I have to use some of the lines from the round table recording, which some of the young people were not very loud, resulting in quite a lot of room sound when bringing the volume up to the levels of the individual recordings.

Regards, Simon.

Accentize do some great tools for this — ie specifically for cleaning up dialogue recordings. DxRevive might be all you need to make things workable. DeRoom Pro is slightly better at room removal. Chameleon allows you to impose room characteristics of one voice on to others. Not cheap but worth every penny.

You might also want to check out the Studio Sound feature in Descript. Last I checked, the automatic editing features in Descript were still a bit too off for serious use, particularly in terms of identifying breaths and 'erms' and making clumsy edits, but the sound processing isn't at all bad, and you can render the Studio Sound effect and import into your editor of choice, where you can blend to taste with the original as a sort of wet/dry control (meaning you don't have to choose the wet/dry setting when you render so can tweak that balance after your edits).

Awesome info, thanks, Matt!

[alexis]

I've noticed Steinberg's SpectraLayers 11 also has De-Rev, Reverb match etc.

I'm just getting into all this, so I clearly can't compare quality of algorithms across programs. The SpectraLayers demos of course show quite impressive results in getting rid of the room reverb sound.

[alexis]

Interesting RX vs SpectraLayers comparison:

https://youtu.be/7KpyLsV7cxs?si=KsnedRte-nMJ1EOn

[James Perrett]

Interesting RX vs SpectraLayers comparison:

https://youtu.be/7KpyLsV7cxs?si=KsnedRte-nMJ1EOn

His Spectralayers vs UVR comparison is also interesting

https://www.youtube.com/watch?v=u3Ntpdt-pCU

although I'd like to hear how they compare when using the UVR Demucs 6 stem model on a guitar based track as I work with mainly guitar based music.

[beej65]

I use Dialogue Enhance 2 from Accentize which has noise reduction, dynamic control, spectral EQ, and gain control in one (relatively) economical plugin; including automatic modes for the latter three. However, I understand here the OP (or thread discussion) was also looking for 'room removal,' which to me usually means a separate plugin, eg. Waves Clarity De-Reverb (and others).

[Stratman57]

I was doing the editing in Reaper, and for the few lines that I needed to take from the round table reading I was able to reduce the room component using ReaFIR.

Regards, Simon.

[James Perrett]

ReaFIR is a remarkably useful tool and it is well worth investigating all of its different facets.

[alexis]

"What is ReaFir" Google result:
https://www.homebrewaudio.com/9603/reaf ... in-reaper/