Renovator for noise elimination

[DCLR-Records]

Hello,
Anyone using ReNOVAtor here ?
Case :
Recordings done during live acoustic concerts, people coughing, doors closing, ...
Complex orchestral music... Needless to say, these noises should be eliminated but the music MUST remain unaffected.
I got the trial version from Algorithmix...
Big frustration :
When trying to remove one occurrence ( I spent some hours), I either get an even worse synthetic type of noise instead, or the music during that time when the cough is removed just becomes synthetic and useless.
And, I might have people coughing or sneezing in the audience each xx second, resulting in hundreds of occurrences in a one hour program.
Questions :
Anyone here being an expert in handling such cases ?
Any "automatic" type of noise removal procedure, avoiding to spend trillions of hours cleaning a concert recording ?
And... The music MUST remain unaffected.
Best regards,
Paul

[allemano bernard@neuf.fr]

Hi,
I sympathize, and I can confirm that there is no 100% effective program in this area. Maybe RX 5 from izotope ..... but you always hear something. The only effective solution and inaudible (if done well) is to replace the fragment polluted by a similar passage. In Pyramix, we can succeed in "reparations" amazing accuracy.

So good luck

and good Sunday

Best regards,

Bernard

[DJS]

Yes I am frustrated by the audibility of these so called miracle spectral repair software packages for classical concert and session recording cleanup. One gets a synthetic replacement of the original event, and then it is a weigh up of which is less objectionable. I have found the pattern replacement the best in RX3 Adv, this is great for clicks from piano stools, etc. But often I leave the live sound events in, as they are usually less objectionable to the software replaced sections. Coughs, due to their wide spectral content always sound naff when replaced with some synthetic thump or wave of some approximation from around the event. I find the cough more appealing.

The most problematic aspect of all this is for a multichannel source, one has to edit the noises out of all the constituent multichannel wavs BEFORE editing if one wants to change mind later. So often I have noise removed an edited "final" master and someone wants a change. Sigh, repeat and rinse.

I guess its still early days for a lot of this stuff.

[allemano bernard@neuf.fr]

Hi,

I find the cough more appealing.

I agree ! ! ! I prefer a noise than a syntethic replacement ! !

Best regards,

Bernard

[DCLR-Records]

Thanks David and Bernard,
Indeed what I was experiencing.
And on some recent jobs I had people caughing, about 100 times during one concert.
Customers seem to think that we sound engineers should just have some magical software or black box that removes noises... Not always easy to tell them that we .... Don't.
Best regards,
Paul.

[phaseboy]

It is absolutely possible to achieve impressive results with reNOVAtor, however it does take a lot of getting use to, and can become very time-consuming - depending on the project it may not always be worth the time and expense. There are times when I'll spend 20 minutes fussing on some complex noise and give up as I can't achieve something satisfactory - only to decide to try again the next day and fix it in a couple of minutes. Most things I can deal with pretty quickly, but it took a lot practice.

I would encourage you to give it another shot and offer the following thoughts:

- in many cases copy/paste will provide a much better result than interpolation
- most things (unless they are tiny in frequency content and separated from any music content) require multiple steps to fix without audible artifacts - ie: it is rare to place a box around a problem and fix it with one click.
- multiple small well placed interpolations can wield superior results to large ones.
- you may not always be able to completely remove a distracting noise without ending up with unsatisfactory artifacts, but I have yet to come across something that I wasn't able to minimize to some degree and make less distracting.

Best Regards,

Mark

[Classic]

There are times when I'll spend 20 minutes fussing on some complex noise and give up as I can't achieve something satisfactory - only to decide to try again the next day and fix it in a couple of minutes.

Mark

I too!

[DJS]

Customers seem to think that we sound engineers should just have some magical software or black box that removes noises... Not always easy to tell them that we .... Don't.

People who cough with an uncovered mouth in a classical music concert should be taken out with a sniper. I've often fantasized about seats that swing away and drop the occupant into the abyss after they cough uncovered. Could be driven my a mic and a spectral content test to see if the cough is uncovered or not.

[fl]

Back when I worked in radio recording a lot of concerts, I spent way too much time refining my design for a ceiling mounted Cough Seeking Laser (TM), a freely rotating installation equipped with precision targeting and facial recognition technology. Someone just has to look like they're about to cough, and all that remains is a puff of acrid smoke and some particulate matter. It remains an open question whether or not the resulting explosion would be more or less disturbing than the cough would have been, as well as its net benefit/detriment to subscription ticket sales.

I have to agree with Mark about copy/pasting being extremely useful, even when more temptingly impressive or exotic tools are available - it was a real revelation when I discovered the various modes of Pasting available within Izotope RX. You can Paste to replace, or you can Paste to overlay on top - highly applicable when imprecise choirs exercise their democratic rights regarding the timing of final consonants, particularly "t"s and "k"s.

[avi]

Hi Paul,
FWIW, I find ReNOVAtor an excellent tool for working on classical music - I've used it to take out thousands of noises over a few years, and wouldn't be without it. (I remember being reluctant to make the investment years ago, but having it in the workflow it can be very useful; if I was starting again it would be one of the first things I would buy.)

I'd echo Mark's observations above, and it's certainly usually possible to mitigate the worst of a noise if not remove it. I generally prefer to denies in multitrack as you can scroll through the tracks and identify the main source of the noise, then treat each track separately (particularly in an orchestral situation with player noise close to a percussion spot, for example, the time delay from there to the main pair means that zapping a noise on track 1 completely misses it on track 22). On some occasions I've found that a noise comes out more easily in the mixed stereo than the multitrack, but that's rare.

I'd add that
- sometimes if a noise is particularly difficult, exiting the ReNOVAtor window and reselecting the same passage of audio but slightly differently and trying again often works; I suspect the different selection changes the maths behind the display and processing a bit and where it sat awkwardly before it then sits better.
- vertical interpolation (and large block sizes) can be useful if you are well zoomed in in log display trying to find the lower ends of a rumble, or remove a mis-plucked viola string.

I grant you though that some coughs can be particularly troublesome due to the nature of the sound!