Background "breathing" noise

Comments

shroadster wrote on 6/17/2005, 11:57 AM
No I mean I recorded a new clip in 16 bit mode...and the background noise is still there, just a little less noticable.
jaegersing wrote on 6/17/2005, 6:34 PM
Can you post a sample of this wave file, one you captured that does not have breathing until you play it in Vegas? I'd like to try and isolate this to the file or else to something on your system.

Richard
shroadster wrote on 6/17/2005, 8:03 PM
I'm not sure how I would post a wave file of the captured audio that only has the breathing in Vegas, because once I render it out in Vegas, the sound is there permanently.

I've finished filming a project, and need to start editing. I captured some footage with Pinnacle Studio, but unfortunately it's Scene-Detection does not divide each clip into a separate file--so I'm having to manually do that in Vegas. I really could use Vegas Capture here, I just can't have this breathing noise. It's getting very frustrating.

Thanks for all the help, any more suggestions would be more than appreciated.

jaegersing wrote on 6/17/2005, 8:25 PM
What I meant was to post a captured, (not rendered), file so that some of us could try to reproduce the effect by playing it in our copy of Vegas. If we can't repro this, then it would point to something at your end. However if others can achieve the same effect just by playing it in Vegas, we should be able to help pin down the problem.

Richard
shroadster wrote on 6/18/2005, 3:40 PM
Ok,

I captured a short clip with WMM that only has the background breathing in Vegas. If anybody would like it just send me an email and I'd be glad to let you look at it.

Thanks for the help.
jaegersing wrote on 6/18/2005, 7:27 PM
I received the WMM clip and tried it out. As you say, it plays OK in media player, but in Vegas and Sound Forge there is the breathing effect. I'll try to describe the sounds in a bit more detail.

I noticed that in media player, there is a soft background roar under the hiss, and this low frequency sound (it's in the 120 - 240 Hz region) is fluctuating periodically (dips at approx 1 sec period). When playing back in Vegas (or Sound Forge), the hiss level drops in sync with the low frequency dips. In media player, the hiss level stays constant and only the low frequency roar dips.

What follows now is pure conjecture on my part.

The audio in the avi file is 32KHz, so I believe it is recorded at 12 bit resolution in the camera. I am suspecting that Vegas is not treating the 12 bit files correctly, and is interpreting them as linear instead of non-linear. It could be that the audio in the file is being tagged by Windows incorrectly as 16 bit instead of 12 bit, but in that case I'm not sure how WMM can get it right.

If you could please send me a file that was captured at 16 bit/48KHz, it would be interesting to see if the results are the same.

Richard Hunter
jaegersing wrote on 6/18/2005, 7:49 PM
I posted the WMM clip on the web if anyone else wants to take a listen.

Richard

Breathing Audio
farss wrote on 6/18/2005, 8:51 PM
As far as I know 12 bit and 16 bit audio files are linear.
Bob.
fwtep wrote on 6/18/2005, 9:07 PM
Unless I just don't know what to listen for, the file sounds the same in media player as it does in Vegas 4 and 6. Do you mean that thing that sounds like an owl? If so, I hear that with the Windows Media Player too. I also listened to the MP3 file listed in the initial post. Again, it sounds the same in each app.
farss wrote on 6/18/2005, 9:22 PM
OK,
had a listen to it. Here's the thing, neither Vegas nor SF do 12 bit but as far as I know that shouldn't matter, a value of 0 = odBFS, in other words a sample of maximum value is the same in an 8,12, 16 or 24 bit file i.e. 0. Sample rates are another matter but nothing in that file sounds like sample rate errors, they show up as nasty clicks or ticks and there's nothing like that in there.

I can listen to the file in Vegas 5, SF7 or WMP and it sounds the same. Firstly the noise floor is only at -30dBFS. that's pretty horrible and it sounds like plain old noise from a mic preamp , just sounds like what I've heard all too often recorded by cheap cameras that have permanant AGC. Buried in amongst it (but at a much lower level than the last sample I listened to) is what sounds for all the world like some mechanical noise. THis could be anything of course but if it's there no matter where you make the recording then it could well be the camera itself, the heads in DV cameras create quite a bit of noise, combine that with how high the mic gain is wound up by the camera and hey presto, wirring sounds in your audio track, most noticable during silence.

Easy enough to check this, get a minipin plug with all the pins shorted together and try recording with that plugged into the mic socket, bet the noise goes. If that sounds too difficult try poking the mic inside a block of foam rubber.

Also as I've suggested before, try Printing To Tape an AVI with absolute silence in it and capture that back, I'll bet the silence stays that way.
Bob.
John_Cline wrote on 6/18/2005, 9:45 PM
12-bit audio is non-linear, it uses a perceptually-coded non-linear compression. The dynamic range is still 96 dB just like 16-bit. In the "Blue Book" DV spec, the low levels are encoded 1:1 and higher levels are compressed a bit in a non-linear fashion. The Blue Book spec states that the compression is a table lookup, with a corresponding expansion on playback.

Quasi-logarithmic compression is a perfectly valid way to squeeze more dynamic range into a given number of bits. At high audio levels, it takes a greater change in level for the level change to be detected by the human ear than at low volumes. The human ear is unlikely to be able to tell the difference between 16-bit linear and 12-bit non-linear if they were both using a 48k sample rate. Unfortunately, 12-bit DV audio is sampled at 32k. So, while you may not notice the dynamic range compression, you will probably notice the limited frequency response if your original source audio had a lot of high frequency information before encoding to 32khz 12-bit.

The possibiity exists that Vegas is not actually using the lookup-table to expand the 12-bit audio to 16-bit audio and is merely mapping the compressed 12-bit audio to the upper 12-bits of the 16-bit audio file. This would easily account for the breathing and there would be some low-level quantizing noise which would vary with signal level, too.

John
Grazie wrote on 6/19/2005, 12:49 AM
John - interesting. Would this level of complexity result in what I would call a type of phasing modulation? Where one "wave" - of a certain bit - is affecting and modulating of the other - ending up with volume variations .. and hence the "breathing"? I can imagine some phased sinusoidal variations going on. I'm wondering if the maths within Vegas 6 is creating such fluctuations?

Grazie
jaegersing wrote on 6/19/2005, 1:17 AM
The part that sounds different is the level of the hiss. In WMP, it stays constant, but in Vegas and SF, it dips along with the low frequency noise. I've tried it on 2 systems and it's totally repeatable.

Richard
farss wrote on 6/19/2005, 1:44 AM
First time I've heard of this and that's a very interesting piece of information, learn something new every day. What's more interesting is Vegas insists on creating a proxy file when it opens the sample, this I have seen happen before but not everytime I've opened a 32K audio file, most of the ones I've come across have been out of FCP and recorded to a DSR-11, Vegas reports them as 16 bit 32K and they playout just fine. I've even switched my DSR-11 into 12/32, recorded some audio and captured it back and it sounds just fine, no wierd noises from memory.

Grazies theory is possible but I'm assuming the non linear 12 bit system is just that, there's no dynamics involved, just a non linear quantization?
Bob.
jaegersing wrote on 6/19/2005, 2:44 AM
Bob, there should not be any dynamic compression, just a straightforward mapping of bits. The link below actually lists the mapping. However, if the playback device interprets the bits wrongly, I'm thinking it is possible that a change in level in one frequency band may cause a level change in a different band.

Also, I just tried stripping the audio from the AVI using Virtualdub, and the resulting wave file plays OK in SF. I don't know how to check whether it's just due to Virtualdub handling the nonlinear conversion properly.

I've just received the 16 bit AVI file, but it's a different clip altogether and the breathing effect is only just noticeable. Don't know what to make of this yet.

Richard

http://rfc3190.x42.com/
farss wrote on 6/19/2005, 3:24 AM
Took trusty old D8 camera, hooked it up to PC with camera set to 12/32K and captured live using VidCap, did same again with camera in 16/48K.
Tried opening both AVIs with Vegas, results are the same, odd thing is Vegas doesn't build an audio proxy file with the 12/32K file, I'm thinking maybe VidCap converts it to 16bit, Vegas certainly reports it as such. We'd really need someone from Sony to explain that.
I don't have WMM on the system to try capturing with that.

However BOTH files contain noise remarkably similar to what's in Shoadster's file. Even the Sonograms look the same, the period of the variation in the noise is different but thats easy to explain. Where the noise is coming from is the fans in my nearby PC, as there's about 6 of them I'd expect there's quite a few LF beats happening.
Bob.
jaegersing wrote on 6/19/2005, 4:08 AM
Hi Bob. The low frequency beats are audible in WMP too, so I am interpreting these as simply part of the recording. As you say, they could be due to motors or other devices.

The part that is strange is that the hiss level behaves differently when the file is played back in Vegas compared to WMP. From what you posted earlier, I'm not sure that you are actually hearing (or noticing?) the hiss in the WMM file. It's probably from the recording mic, it certainly couldn't be produced by PC fans.

Richard
farss wrote on 6/19/2005, 6:36 AM
The 'hiss' is undoubtably coming from the mic preamp, no mic could be THAT bad surely. It's a bit hard to tell what's going on and for my money far from anything to get too excited about. The bigger issue is having that much noise, it means anything recorded that way is pretty well nigh useless for any serious work anyway.
You'll get away with it in many situations as the actual environmental noise keeps the cameras AGC from winding the gain (and hence the noise) up too high but try using this kind of gear to record VOs or music in a quiet environment and you run into serious problems, been there done that.
Even when you get the noise down another 20dB as soon as you start dialing in a moderate amounts of compression bingo, the noise floor comes up, so then you add a noise gate and kill the room tone which makes the compression sound 'breathy'.
There's certainly an interesting side issue here though, just how Vegas handles 12/32K audio, I've asked about this some time ago and no real answer was forthcoming.
Bob.
shroadster wrote on 6/19/2005, 7:24 PM
Well I just don't get it. Even with 16 bit the captured audio has a slight breathing effect. I just don't understand why some programs (Vegas) capture it with the noise, and others don't. Is there a possibility it's just something with my camera that Vegas doesn't agree with?

Perhaps I'll just buy Scenalzyer and use it for my capturing needs. That's the only fix I can see at the moment.

Thanks everyone so much for your help. You are the experts, and I've learned quite a bit through this...well...at least tried to :)
Spot|DSE wrote on 6/19/2005, 8:13 PM
Since I've forgotten the beginning of the thread.....did you disable all FX in the audio trax?

shroadster wrote on 6/19/2005, 8:24 PM
It sounds the same (breathing present) with or without the FX on the audio tracks.
jaegersing wrote on 6/20/2005, 2:07 AM
Your camera isn't a VX1000 by any chance, is it?

Richard
farss wrote on 6/20/2005, 5:15 AM
Shroadster,
I don't understand your logic, the sample that was posted recently and supposedly captured using WMM has the same 'breathing' noise as the one you sent me and I can create pretty much the same noise using my own camera. The noise is coming from my room, there's not a thing unusual about it and capturing with a different application isn't going to change anything.

Even if I put two closed doors between the noisy stuff in my room and my mics and going into my pretty decent audio gear, guess what, same noise, just about 20dB quieter. Sure there's a lot less hiss, my mics and preamps are good for around 90dB S/N but outside of a purpose built studio it's pretty hard to find an environment quiet enough to do justice to them. What my gear doesn't have is AGC, if I decide to wind the gain up as high as it gets on most consummer cameras I can record noise just as bad as that and it sounds the same.

The only answer to your problem is to get better kit and use a quieter environment to record in. That doesn't mean a more expensive camera either, very few cameras have anything like the performance of even mid range audio gear.

Bob.
shroadster wrote on 6/20/2005, 11:50 AM
Nope, not a VX1000.

Bob, the audio that I capture with Pinnacle Studio or Scenalyzer does NOT have the breathing noise in Vegas or any other program. And when I render it out in Vegas, it is NOT there, I guarantee it.

However the audio from Vegas-captured clips, WMM, and another cheap program have the breathing noise in Vegas, and when rendered out it is very noticeable and annoying. So in my logic, getting Scenalyzer would fix the problem. I just would rather use Vegas capture, however right now it doesn't look like an option.

I would get a better kit if I could, trust me. But I'm just a kid lucky to have what I have. Thanks for the help everyone.