1) Removing clipping that isn't clipped and 2) Good clip removal software

[[email protected]]

Is there anything that can be done about audio that has distortion from overloaded circuitry but doesn't display a "clipped" wave form?

Would you agree that iZotope is the best distortion removal software? Is there something that's almost as good but lower in price?

Thoughts on, experience with SoundSoap?

Thanks.

[Nate Najar]

On Friday, June 6, 2014 2:28:51 AM UTC-4, wrote:
> Is there anything that can be done about audio that has distortion from overloaded circuitry but doesn't display a "clipped" wave form?
>
>
>
> Would you agree that iZotope is the best distortion removal software? Is there something that's almost as good but lower in price?
>
>
>
> Thoughts on, experience with SoundSoap?
>
>
>
> Thanks.

I used izotope to do just that. I had a singer who actually clipped a distressor, but the part was so good there was no way I was going to retrack it.. izotope took it right out.

in the daw I like to setup an extra track for edits like that. on the bad track cut out the section with the problem and move it to the "edit" track but in the same position on the timeline. do the repair on the edit track )which is of course just a small snippet of audio) and then bus the output of the edit track to the input of the original track. Set the DAW record mode to only record the selected region/item/whatever. Then you just arm your original track and press record. it's a seamless edit and you don't hurt any of the rest of your track.

[Scott Dorsey]

> wrote:
>Is there anything that can be done about audio that has distortion from overloaded circuitry but doesn't display a "clipped" wave form?

Retrack it.

>Would you agree that iZotope is the best distortion removal software? Is there something that's almost as good but lower in price?

You can try using a the LP declicking algorithms in a lot of different
software. I have used the CEDAR one to clean up clipped tracks before.

But if you don't see flattopping on the envelope display, the cleanup
software probably won't be of that much use.

It depends a lot on the signal. You can often make a slightly clipped vocal
into something usable. You can take a clipped horn and it'll be fine. But
forget about a clipped piano, it's not going to be fixable.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Peter Larsen[_3_]]

On 06-06-2014 07:28, wrote:

> Is there anything that can be done about audio that has distortion from overloaded circuitry but doesn't display a "clipped" wave form?

That's a tough one, but you have gotten advice from people that know
more than I.

> Would you agree that iZotope is the best distortion removal software? Is there something that's almost as good but lower in price?

I don't know, but I know that Audition 3 can unclip stuff that is
clipped more than 6 dB and produce probable sounding audio. I think it
is Izotope OEM ... but it always did have a good unclip. If it clip is
neat, then perhaps make it neat by clipping an extra half dB works by
producing the curve shape the algorithm expects.

> Thoughts on, experience with SoundSoap?

Huh - Whatsat?

> Thanks

Kind regards

Peter Larsen

[hank alrich]

Scott Dorsey > wrote:

> > wrote: >Is there anything that can be done about
> audio that has distortion from overloaded circuitry but doesn't display a
> "clipped" wave form?
>
> Retrack it.

One ight way, but not always available. Remember when Geoge Massenburg
posted in response to someone outraged that there was a touch of
clipping on an EmmyLou, Dolly, Linda track?

--
shut up and play your guitar * HankAlrich.Com
HankandShaidriMusic.Com
YouTube.Com/WalkinayMusic

[hank alrich]

Peter Larsen > wrote:

> > Thoughts on, experience with SoundSoap?
>
> Huh - Whatsat?

http://www.soundness-llc.com/soundsoap3.html

--
shut up and play your guitar * HankAlrich.Com
HankandShaidriMusic.Com
YouTube.Com/WalkinayMusic

[Peter Larsen[_3_]]

On 06-06-2014 17:45, hank alrich wrote:

> Peter Larsen > wrote:

>>> Thoughts on, experience with SoundSoap?

>> Huh - Whatsat?

> http://www.soundness-llc.com/soundsoap3.html

Ah, thanks, I like the configurability of Audition 3's noise reduction,
but perhaps worth testing anyway even if it is marketed as dumbified.

Kind regards

Peter Larsen

[[email protected]]

Most of these are described as working in the time domain, rounding off the waveform etc.

You can also try this working in the frequency domain with most any DAW as follows:

Split the track into two identical tracks.

Take the first track and apply low pass filtering or 6 dB per octave slope downward or whatever EQ sounds good in terms of minimizing the distortion without regard to the loss of highs. We will be putting the highs back in the next steps.

Take the second track and high pass filter it by setting the EQ more or less complimentary to the first track.

Take this high pass filtered track and pass it through a gate. Set the gate threshold so that it stays closed most of the time but opens only when there is actual high frequency stuff you want i.e. the s sounds.

If the high frequency energy due to the distortion exceeds the level of the desired high frequencies, then you need the more advanced gate threshold....

A bit more advanced method adds a compressor with a side chain input and use the low pass track to "duck" the high pass track before the gate. Thus the gate will open only when the low freq track is low amplitude and the high freq track is high amplitude.

Basically you want to keep the gate closed when the high freq energy is the distortion and open it when the high freq energy is real.

This can work pretty well on an isolated Vox track but obviously gets messy on a more complex source.


Have fun
Mark

[[email protected]]

On Friday, June 6, 2014 2:39:38 AM UTC-4, Nate Najar wrote:
> On Friday, June 6, 2014 2:28:51 AM UTC-4, wrote:
>
> > Is there anything that can be done about audio that has distortion from overloaded circuitry but doesn't display a "clipped" wave form?
>
> I used izotope to do just that.


The problem is the distortion isn't from clipping of the file, the clipping sounds like it happened downstream and the sound of clipping is what was recorded, but the actual signal doesn't display clipping.

Here's a short sample. I've tried the RX3 demo, haven't made any headway on it.

https://drive.google.com/file/d/0B0zvAZXgfLgiWmpKNlF1RTFjVUU/edit?usp=sharing

[Peter Larsen[_3_]]

On 09-06-2014 04:15, wrote:

> On Friday, June 6, 2014 2:39:38 AM UTC-4, Nate Najar wrote:
>> On Friday, June 6, 2014 2:28:51 AM UTC-4, wrote:

>>> Is there anything that can be done about audio that has distortion
>>> from overloaded circuitry but doesn't display a "clipped" wave form?

It kinda does, left has a plateau at -4.5 dB and right at -10.5 dB. Why
do you supply a stereo file of a mono signal? But it sounds distorted
rather than clipped.

>> I used izotope to do just that.

> The problem is the distortion isn't from clipping of the file,
> the clipping sounds like it happened downstream

No, it appears to be is mic or mic pre. Perhaps made less obvious by
channel eq. What mic was it?

> and the sound of clipping is what was recorded, but the actual signal
> doesn't display clipping.

It quite easy to clip cleanly and then unclip and roll-off gently - I
tried a 5 kHz lowpass first order bessel, but it doesn't really alter
the sound. Add some verb and don't put it too loud in the mix and leave
it be, seriously. If somebody asks say "Yes, but we wanted that effect
for the 1960's atmosphere it provides". Or use a lot of time with some
vinyl tools and don't really get it good anyway.

> Here's a short sample. I've tried the RX3 demo, haven't made any headway on it.

> https://drive.google.com/file/d/0B0zvAZXgfLgiWmpKNlF1RTFjVUU/edit?usp=sharing

Here are the results of the 5 minutes you got for your 10 cent:

https://dl.dropboxusercontent.com/u/34064013/clipped.wav

https://dl.dropboxusercontent.com/u/34064013/unclip.wav

https://dl.dropboxusercontent.com/u/34064013/unclip-lowpass.wav

The will not be online for long, about time I turn that box upside down
and dust it ...

Kind regards

Peter Larsen

[[email protected]]

On Monday, June 9, 2014 2:48:54 AM UTC-4, Peter Larsen wrote:

> It kinda does, left has a plateau at -4.5 dB and right at -10.5 dB. Why
> do you supply a stereo file of a mono signal?

It's what's on the DVD. It's a DVD of a webcast.

> What mic was it?

No idea. This is as good a shot of it as I can find on the video. The guy who sent me the video knows nothing about the tech details. I have no access to the source material. It may have been straight from the mics to video. The one thing he knew is it was recorded to analog video. Don't know how it was digitized.

https://drive.google.com/file/d/0B0zvAZXgfLgiVzFwNVJDMUpoNGs/edit?usp=sharing

> Here are the results of the 5 minutes you got for your 10 cent:
>
> https://dl.dropboxusercontent.com/u/34064013/clipped.wav
>
> https://dl.dropboxusercontent.com/u/34064013/unclip.wav
>
> https://dl.dropboxusercontent.com/u/34064013/unclip-lowpass.wav


Can't access them. Any way you could upload them to Google Drive and make the sharing for anyone who has the link? It's free.

[hank alrich]

> wrote:

> On Monday, June 9, 2014 2:48:54 AM UTC-4, Peter Larsen wrote:
>
> > It kinda does, left has a plateau at -4.5 dB and right at -10.5 dB. Why
> > do you supply a stereo file of a mono signal?
>
> It's what's on the DVD. It's a DVD of a webcast.
>
> > What mic was it?
>
> No idea. This is as good a shot of it as I can find on the video. The guy
>who sent me the video knows nothing about the tech details. I have no
>access to the source material. It may have been straight from the mics
>to video. The one thing he knew is it was recorded to analog video.
>Don't know how it was digitized.
>
>
>https://drive.google.com/file/d/0B0zvAZXgfLgiVzFwNVJDMUpoNGs/edit?usp=s
>haring
>
> > Here are the results of the 5 minutes you got for your 10 cent:
> >
> > https://dl.dropboxusercontent.com/u/34064013/clipped.wav
> >
> > https://dl.dropboxusercontent.com/u/34064013/unclip.wav
> >
> > https://dl.dropboxusercontent.com/u/34064013/unclip-lowpass.wav
>
>
> Can't access them. Any way you could upload them to Google Drive and make
>the sharing for anyone who has the link? It's free.
>

All those links come right up on my machine. Dropbox is also free, those
appear to be public links, so I'm curious what's going wrong on your
end.

--
shut up and play your guitar * HankAlrich.Com
HankandShaidriMusic.Com
YouTube.Com/WalkinayMusic

[[email protected]]

This clipping sounds really bad. Sounds like it is not just clipping but reversal also. Some input stages don,t just clip when overloaded, they reverse phase. You are going to have to really roll off the top end to smooth that out.

Mark

[[email protected]]

Okay, I got the files to work under IE - Firefox wasn't cooperating.

Basically you killed the highs to sort of mask the distortion.

I sincerely appreciate you giving it the old college try but it takes too much away from the rest of the audio. Most of it isn't badly distorted like that. Guess I'll just have to live with the distortion.