KVR Audio

jackson · Post by **jackson** » Fri Aug 11, 2006 7:09 am

I have an audio recording I need to denoise. Using Waves xnoise by sampling the noise signature, I thought I'd successfully removed the high end noise from the sample. However, it apparently only does a good job of removing the noise around centre pan. There's still audible noise at the left and right extremes of the signal. Using further reduction just degrades the audio far too much, and repeated attempts at removing the noise from the original sample still leave the left/right extremes unfiltered... Regular lowpass filters simply destroy the sound.

Is this odd? I've certainly never come across this issue before. I'd love to be able to apply a little more noise reduction to only the left and right extremes of the "panoramic spectrum" (or whatever it's called), but the technique is eluding me at this hour

Cheers!
Mike

Blue Cat Audio · Post by **Blue Cat Audio** » Fri Aug 11, 2006 1:02 pm

Maybe you could try side channel EQing (low pass filtering the sides of the signal only).

BertKoor · Post by **BertKoor** » Fri Aug 11, 2006 1:43 pm

Split up the stereo file in two mono files and do noise reduction on them seperate. Then join them again.
How does Adobe Audition do it?

Blue Cat Audio · Post by **Blue Cat Audio** » Fri Aug 11, 2006 2:30 pm

Split up the stereo file in two mono files and do noise reduction on them seperate. Then join them again.
How does Adobe Audition do it?

I think the artefacts of the noise reduction process will produce an unbalanced stereo track (removed frequencies will be potentially very different on L and R channels), so it might be even worse (human hear is sensitive to stereo balance)...

BertKoor · Post by **BertKoor** » Fri Aug 11, 2006 2:53 pm

If you can reduce noise that way without audible artefacts, aren't they inaudible??

Blue Cat Audio · Post by **Blue Cat Audio** » Fri Aug 11, 2006 3:04 pm

It really depends on your de-noising algorithm and your content. Usually what it (very basically) does is remove some frequencies that are supposed to be noise, and let the others...

Anyway, some content is removed and/or modified. If it's not done the same way on the left and right channels (or at least balanced by the algorithm), the artefacts you usually do not hear might be audible, because different on left & right channels, what will disturb your ear.

It's not a general rule, it depends on both the algorithm and the content: you should give it a try, but I think it might be worse than leaving the noise or a bit of EQ.

I'd be interested in hearing the result of the test though

jackson · Post by **jackson** » Fri Aug 11, 2006 3:47 pm

I'll give it a try and let you know. If I have time, I'll post the results...

Now that I'm thinking about it though, I'm pretty confused by this. Last night I was just thinking of the sounds in terms of a single file with the sound placed in the panorama. But obviously this is a skewed view, as pan positions are just determined by the relative volumes of the left/right channels, yeah?

Hmmm... I'll get a chance this weekend to hear the sound on a large-ish club sound system, so I'm looking forward to seeing if the noise is bad or even noticeable there.

Blue Cat Audio · Post by **Blue Cat Audio** » Fri Aug 11, 2006 3:51 pm

ut obviously this is a skewed view, as pan positions are just determined by the relative volumes of the left/right channels, yeah?

Yes that's it. Very simple

BertKoor · Post by **BertKoor** » Fri Aug 11, 2006 4:01 pm

jackson wrote:I'll get a chance this weekend to hear the sound on a large-ish club sound system, so I'm looking forward to seeing if the noise is bad or even noticeable there.

Do you know the RMS level of the noise in solo? If it's below -70dB probably no one will notice. You can also try listening in a car CD player while driving.

Listening with headphones reveals tiny irritating things like that, which will go unnoticed in real-world listening environments.

jackson · Post by **jackson** » Fri Aug 11, 2006 4:13 pm

bluecatonline wrote:
ut obviously this is a skewed view, as pan positions are just determined by the relative volumes of the left/right channels, yeah?
Yes that's it. Very simple

So now I'm thinking about what you said about the possible problems with processing each channel separately. Maybe xnoise "accidentally" knocked the noise frequencies out of phase, which would explain why they can only be heard hard panned. Hmm, that makes sense...

BertKoor wrote:Do you know the RMS level of the noise in solo? If it's below -70dB probably no one will notice. You can also try listening in a car CD player while driving.

Listening with headphones reveals tiny irritating things like that, which will go unnoticed in real-world listening environments.

Yeah, I was doing detailed editing late last night with my cans on, and the noise was really bugging me. I'm pretty sure the noise won't be too audible on most systems, but we're going to be sending the song out to labels in a couple of weeks so I'm trying to get a cleaner sound.

kylen · Post by **kylen** » Fri Aug 11, 2006 6:23 pm

Hopefully X-noise isn't looking at the mono and sides separately. In my experience that can really re-phase up your audio which can make sounds breathe across the stereo soundfield, dynamically, if you know what I mean.

I would think that Adobe Auditon and other NR tools would have a solid L/R linkage otherwise you'd get seasick. I haven't tested this so I'm just guessing.

But if the extreme sides are too noisy, and you're playing this in a club, narrow the width of the stereo field a little till it goes away - I have to do that sometimes. With something like Ozone3 you can do a multiband narrow, for example just narrow the mids, or just narrow 4K on up a bit.

jackson · Post by **jackson** » Fri Aug 11, 2006 6:37 pm

I left a message for the promoter last night to find out if the sound system will be running in mono as a lot of clubs often do. Hopefully not, for this case, as I want to hear it in stereo...

I don't have Ozone3, but that sounds like a fantastic feature. I'm using another stereo plugin on some other drum tracks to get a wide sound, so I guess I don't need all that stereo width on the noisy sample.

I'm really starting to think that either xnoise caused a phase shift or the sample was never correctly aligned when I recorded it... It's actually a "sample" of a variety of percussion elements (ie a homemade drumloop) recorded from vinyl, through a DJ mixer and then into the AD's... Plenty of room for error along the way, I guess. I know the clock on my soundcard is rock solid though, so it shouldn't be jitter on the recording that caused it.

Blue Cat Audio · Post by **Blue Cat Audio** » Fri Aug 11, 2006 6:59 pm

narrow the width of the stereo field a little till it goes away - I have to do that sometimes. With something like Ozone3 you can do a multiband narrow, for example just narrow the mids, or just narrow 4K on up a bit.

Yes that's what I was proposing.

BTW [start self promotion] we have three equalizers that can do Mid/Side EQinq for such purposes: the Widening Parametr'EQ, Widening Liny EQ or Widening Triple EQ can help you to do it easily for a very nice price [end self promotion]

Maybe xnoise "accidentally" knocked the noise frequencies out of phase, which would explain why they can only be heard hard panned. Hmm, that makes sense...

Did you make a comparison with the original? (you can also try to substract the xnoised one from the original)

Good luck!

jackson · Post by **jackson** » Fri Aug 11, 2006 7:38 pm

bluecatonline wrote:
narrow the width of the stereo field a little till it goes away - I have to do that sometimes. With something like Ozone3 you can do a multiband narrow, for example just narrow the mids, or just narrow 4K on up a bit.
Yes that's what I was proposing.

BTW [start self promotion] we have three equalizers that can do Mid/Side EQinq for such purposes: the Widening Parametr'EQ, Widening Liny EQ or Widening Triple EQ can help you to do it easily for a very nice price [end self promotion]

Maybe xnoise "accidentally" knocked the noise frequencies out of phase, which would explain why they can only be heard hard panned. Hmm, that makes sense...
Did you make a comparison with the original? (you can also try to substract the xnoised one from the original)

Good luck!

hahaha, I was waiting for the plug

I'm really not in the market for new plugins right now, got some other things I NEED first. I'll definitely take at look at your range when I am though.

Ugh, no time for the studio right now with work today and two gigs this weekend

Blue Cat Audio · Post by **Blue Cat Audio** » Fri Aug 11, 2006 8:09 pm

hahaha, I was waiting for the plug

It was difficult not to mention them given what was said in the previous posts

Ugh, no time for the studio right now with work today and two gigs this weekend

Have a nice Week End. It's night time here... Good night!

eq'ing panned audio / pan placement noise reduction