Hi Folks,
I’m running Audacity 1.3 and wonder if someone can help me.
I’ve recorded a friends band live on a Zoom R16 then imported the individual tracks into Audacity.
So far so good. There is the perennial problem of bleed, guitar onto bass drum mic for example.
I’ve tried the normal Noise reduction and can’t find settings that allow me to reduce the bleed without affecting the bass drum sound.
Can anyone suggest another plugin or settings for standard Noise reduction.
Thanks
Mike Abbott
As you suggest in your title, this could be a good candidate for using a noise gate.
There is a noise gate plug-in for Audacity here: http://wiki.audacityteam.org/wiki/Nyquist_Effect_Plug-ins#Noise_Gate
(note this plug-in requires Audacity 1.3.x (1.3.13 recommended)
Instructions for installing plug-ins are here: http://audacityteam.org/download/plugins
I’m the author of that plug-in, so if you have any trouble using it, please don’t hesitate to ask. In fact, any feedback about plug-ins is very helpful for making future plug-ins useful and user friendly.
Many thanks for the quick reply,
It’s 1am here in the UK so I may well wait till tomorrow to try this.
I’ll let you know how I get on.
Mike
Hi Steve,
I’ve tried the noise gate both with recommended and guessed settings. The bleedover is such that I can’t reduce it to any great extent without adversley affecting the sound I want on the track.
I could try post a clip here but aren’t sure how to do it.
Mike
When you are creating a topic or a reply you should see a tab just under the Save Draft / Preview / Submit buttons at a tab which says “Upload Attachment”
WC
Thank you,
I’ve seen it now. I’ll upload an example later.
Please upload it in either WAV or FLAC format (lossless formats).
The forum only allows short files to be uploaded, but we only need a couple of seconds.
Thanks Steve,
Is there a file size limit?
Yes - I think it’s 1 MB.
(that’s about 6 seconds for a stereo WAV file, 12 seconds mono, 20 seconds mono FLAC format)
Hi Steve,
Here is the unprocessed example of the bass drum track.
By the time the signal is normalised and compressed there is quite a lot of bleed.
Any suggestions would be appreciated.
Mike
A noise gate isn’t going to work on that one.
If you have an isolated version of the guitar track (i.e. guitar only) which has bled into the kick drum track
you could try spectrally subtracting it from the kick drum track with the (free) Kn0ck0ut plug-in or similar.
[Kn0ck0ut is windows only]
That’s quite a tough one due to the relatively high level of bleed through (you’re not going to be able to get it perfect). For future reference you may want to look at your microphone positioning to try and lessen the bleed though at source.
For the gate, you really need to use a short gate attack time, and a slower release time, so I’ve made a custom version for you:
noisegate.ny (8.6 KB)
The help files have not been updated, but the only change is that the attack time (how quickly the gate opens) and the release time (how quickly it closes) are now separate controls. The gate opening (letting the sound through) uses lookahead, so that the gate should be open by the time the sound arrives.
As a starting point I’d suggest (after normalizing the track) something like this:
To “beef up” the sound a bit after using the noise gate (and to even out the volume a bit, you could try applying this soft clipping limiter.
Thanks to everyone for their assistance.
I’ll try both knockout and the modified gate. Certainly it would be easiest if there wasn’t any bleed through in the first place.
I had a call from the singer last night he’s listened to the tracks and thinks his vocals are aweful, I’d agree with that.
Can we re do the whole recording. At the same time I now find out that the drummer owns a headphone amp with 6 outputs so next time we won’t have to bother with vocal monitors.
The drummer also has an electronic kit. I wonder if it would be a cheat to ask him to use that and get away from drum mics?
So we will get a second chance.
Regards
For another use–audio book narration–
I’d been using noise gate for quite a while to reduce breath sounds (not eliminate!, just reduce) and found with the standard noisegate,ny that not only did it reduce the breath sounds, but that the ends of some words were also reduced. In fact reduced to unnoticeable. Not one of my listeners (proofers on the rights holder end) had commented on it, since it was usually clear that the -ing or -ed was “there” if only in the listeners mind. But it really bothered me that I couldn’t set the attack and decay times separately…until I found the “custom” version that allowed me to leave the quite ends of words by setting the release to a longer (.7 sec) time and the attack to a shorter time (.1 sec).
I’ve only played with a few samples, but am tremendously encouraged by what I’m hearing!!
Mac OS 10.10.1
Audacity 2.1 alpha
So, no help requested, just THANKS! (and a cross posting into the Narrators thread…where I perhaps others struggling with breath (not for breath 
, just with the sound of it) might start experimenting with it.
Most likely you need to set the “threshold” lower.
I’d been using noise gate for quite a while to reduce breath sounds (not eliminate!, just reduce) and found with the standard noisegate,ny that not only did it reduce the breath sounds, but that the ends of some words were also reduced.
That’s correct. It’s really difficult to patch up a ratty performance. It’s been my experience, however, that the only person who complains about “mouth effects” and respiration is the person whose mouth is doing the effects and respiring. There’s nothing that makes you more self-consious than listening to yourself on a good quality recording.
I would be surprised if anyone at ACX complained about it. They’ve said multiple times they are perfectly comfortable with natural reading style and normal noises like clothing rustle and swallowing. Precisely the sounds that would be present had you been sitting in the room over by the window reading to me.
What they don’t want is distracting noises like that dog barking behind chapter three or that annoying rain-in-the-trees hiss or gargling mosquitoes riding through everything.
Everybody also assumes buying the perfect microphone will solve everything, when recording in a nice quiet room with a modest microphone can work very well.
Another distracting effect is the talking into a wine glass or bad cellphone effect that you can get by overprocessing your voice. This is a very common problem. Everybody looks at all those tools and effects and thinks ‘surely I can use a bunch of them in my show.’
The attached is an after and before example of overprocessing. Note in the before (second and third cut) the natural clicks and breathing sounds.
Koz