@SP. So something like Steinberg SpectraLayers Pro would be useless for such a project?

There is nothing in Sound Forge to handle this? I think SF Pro can handle this...

@Former user OK, if you own SpectraLayers that is good choice for manually removing noise and unwanted sounds from your recording and in this case RX would be somewhat redundant. But it will not remove reverb which is your problem if the recorder was sitting in the middle of the room. If think that Acon Digital Deverberate works good for removing reverb.

How would this be done. I am sort of new to audio repair and each recording has different issues, some have a continuous tone, others it is hiss, or noisy environs.

@Former user I think it is a good idea to post some snippets like you suggested so we can look into it.

Okie, here is a sample. Let me know if you need it to be longer. https://www.filemail.com/d/dysczxtiuvuibpa

@Former user Thanks, here is how you can do it with SpectraLayers:

Select the Time-Range-Selection tool and select a part of the audio when nobody is talking so you get only the noise sound.

Then register the selected noise.

After that you can open the Noise reduction and play with the settings. Setting it to 100% will remove the hiss and noise completely and it will sound much better.

Before / After

@SP. Thanks for that. Much appreciated.

SpectraLayers Pro (SLP) rocks my world. RX Advanced is amazing as well but neither or any of the other de-verb plug-ins currently available can 'totally' fix a mic placement issue. Excessive de-verbing and restoration will cause artifacts that can be worse than the original problem. SF Pro-15 includes the legacy (NR-2.0) noise reduction suite tools as well as RX Elements plug-ins (not to be confused with RX Standard or RX Advanced, which differ vastly in the number of tools.. and price). Tape hiss and other 'din' type noise attenuation can be quick and easy with the built in SFP tools..
The Sound Forge Pro Suite version includes SLP and the aforementioned SFP NR tools. SLP has a 'slight' learning curve, so you will need to learn the UI to get up to speed, especially surgically working with the spectral tools. If you have experience with photo editing software will be somewhat familiar.

I will check out the sample clips when time allows.

Here is another audio clip: https://www.filemail.com/d/gvoogockpiperxg

I had a cassette recorder and put the mic in the middle of the room because there were many interviewees. I am not interested in deverbing so much as trying to get each speaker sound clear enough for the listener to not struggle too hard to hear. I want to grab and elevate the voices. If there is *some* noise and reverberation, that is fine. But more important is that the interviewees need to be heard clearly (as possible).

@Former user For your next clip try the Voice Denoiser inside SpectraLayers. It works wonders here.

@SP. I did as you instructed for the first sample and while the noise is gone, there are now artifacts. The interviewee sounds like a Decepticon (from the 80's cartoon, never saw the new films). Is there anything I can do about that? Same with the second sample, there is also bit of artifact, sort of reminds me of "Howmework" era Daft Punk with some filter effects and a bit of early Laurie Anderson robotic-sheen vocals here and there.

@Former user It looks like the noise removal deletes too much of the frequencies which are needed for the human voice to sound natural. I would suggest you write down the timestamps where it doesn't work good enough and remove less noise in these spots instead of removing the noise from the whole file in one go. Thats a bit of manual labor.

@Former user

First, I must say that I don't have Spectralayers, but I do have, and use, iZotope RX8 Advanced.

The outcome you are describing is always, in my experience, caused because the user has applied "too much" processing at one go. I don't know how much user control SL allows but I would suggest making several passes with it, each time just reducing the "noise" by a small amount, perhaps as little as 2 or 3dB. Also, if you have access to anything like an FFT setting, experiment with different amounts of this on each pass.

HTH

Jeff

Is it possible that Sound Forge Pro might have all the required tools for audio repair/noise reduction? Maybe iZotope makes noise reduction/audio repair easier and more streamlined but Sound Forge Pro might be able to achieve the same results, perhaps in a longer, more complicated way. I checked and there seems to be A LOT of tools in SFP, maybe enough for what is needed. Or is it that the likes of iZotope has exclusive tools one cannot find anywhere else...?

@Former user Are you mixing up Sound Forge and SpectraLayers?

@SP. No I am referring to SFP, specifically. I take it, by your question, the answer is a definitive, "No."

@Former user You can see a list of features here: https://www.izotope.com/en/products/rx/features.html#compare

@SP. Okay, I am going to stop all audio work and save up one of these years for iZotope, the end all be all of all software. The alpha, the omega. All you need is iZotope. thanks.

@Former user Be sure to get the iZotope everything bundle. 😉

It's iZotope's world, we just live in it.

Or I could go for Acon Digital Post instead and get a pretty good suite that gives iZotope a run for it's money at a fraction of the cost. It also goes on sale so it can be acquired cheaper.

Here is another piece of 'deep problem audio' https://www.filemail.com/d/uejodisrxboudon Does anyone here think this can be salvaged?

Here is another piece of 'deep problem audio' https://www.filemail.com/d/uejodisrxboudon Does anyone here think this can be salvaged?

If/when you read my other post in regards to noise reduction, it would be in your best interest that if you have these files and have audio which starts prior to the voices, you should include it and not trim up the file.

One thing I dislike about the iZotope De-Noise, the plugin asks for a minimum of 1.5 seconds of audio for the "learn" function. However, that should provide you a good target of what these noise reduction plugins are expecting in regards to the amount of noise without voice present to include so the learn function has the best chance of being effective.