Taperssection.com
Gear / Technical Help => Post-Processing, Computer / Streaming / Internet Devices & Related Activity => Topic started by: jagraham on September 09, 2011, 11:42:09 PM
-
I have a couple recordings from this summer i am trying to figure out. both have a bass heavy feedback sound from the stage that might be dangerous for the listener's speakers. keep in mind i am somewhat new to post processing, having just started working with HPF in post. i have attached snippets of the shows in question and was hoping someone could share their experiences with this and post processing. it seems to be they type of thing you cant "fix" but might be able to "make less terrible" somehow.
-
Attached are a before screenshot (spectral vision), and after, a view of the same original file in audacity (for familiarity), and the resulting wave file. I worked on the one around the 3-4 second mark.
The good thing about feedback is that it's generally a constant tone. There may be harmonics created in upper bands, but it's still flat and generally prominent, so finding it isn't hard once you hear it (as illustrated in the before shot). Once you figure out the frequency band (and any harmonic bands) you want to eliminate, it's just do a tight EQ/notch filter on that band. I was cheesy and selected the time and frequency stuff and then knocked 30db out of it for illustration. The spectral tool in Audacity works enough to tell you where it is (once you elongate the frequency band so you can see stuff better), but it takes some effort. Still, better then nothing if you get good at it. Best of luck.
-
Can you actually just draw the box shown in the screenshot directly in Audacity when viewing it in the spectral format? Which tool is that under, because it would solve some other problems that I have been having with certain recordings. TIA
-
Can you actually just draw the box shown in the screenshot directly in Audacity when viewing it in the spectral format? Which tool is that under, because it would solve some other problems that I have been having with certain recordings. TIA
I just got home and tried in Audacity, sorry, no box drawing. For reference, I did most of those shows in Isotope's RX since the concept was the same. However, it's still very possible and easy for single tone feedback (you'll need to figure out harmonics on your own) and when things change pitch you can sort of fix it, but it takes more effort as you do it in steps.
Attached are 3 screenshots, before (notice that I've zoomed in both on the timeline and the frequency scale), the (delivered) EQ with the notch filter I used, and after. You could use a rectangle instead of a spike, but I was lazy in the example. Best of luck.
-
I have used several notch filters like the one you have shown in your middle screenshot, but how does that do anything for the harmonics, or do you also have notches at other frequencies as well? Or is the reduction of the primary offensive frequency sufficient to make the harmonics relatively minor by comparison. I have used Soundforge's spectrum analyzer, but the view is not the same as the spectral view in your screenshots. It is more of a graph of the frequencies and their relative volumes. Is there a plug-in or other spectral viewing program that is readily available out there that you know of?
-
I have used several notch filters like the one you have shown in your middle screenshot, but how does that do anything for the harmonics, or do you also have notches at other frequencies as well? Or is the reduction of the primary offensive frequency sufficient to make the harmonics relatively minor by comparison.
In audacity it's done by hand, but yes, if there are harmonics I generally go after them, especially if they are noticeable. It's easier in RX since it has a magic tool that picks up on the harmonics and selects them (and works about 80% of the time in it's selection), but thats not to say it's carefree. The problem I've encountered is that it's tougher to knock out an astronomical amount when you have 3 or 4 harmonic bands without doing noticeable damage, so you end up just reducing the problem, not necessarily eliminating it.
I have used Soundforge's spectrum analyzer, but the view is not the same as the spectral view in your screenshots. It is more of a graph of the frequencies and their relative volumes. Is there a plug-in or other spectral viewing program that is readily available out there that you know of?
Dunno, sorry. Both Audacity and RX have native views. I wonder if the one that SF uses is buried somewhere. I can find =http://www.google.com/search?q=Soundforge%20spectral%20view&oe=utf-8&um=1&ie=UTF-8&tbm=isch&source=og&sa=N&hl=en&tab=wi&biw=1280&bih=821screenshots (http://=http://www.google.com/search?q=Soundforge%20spectral%20view&oe=utf-8&um=1&ie=UTF-8&tbm=isch&source=og&sa=N&hl=en&tab=wi&biw=1280&bih=821screenshots) that look right, but I'm not familiar enough with the product to recommend where to look.
-
can't find that spectral view in Audacity. Help?
-
here's a sample of the pitch:
files.me.com/brian.emerick/tfngxd.wav
I've run it through Izotope, and it doesn't pick up on it in denoise or hum removal. I can't seem to "see" the pitch line.
edit: nevermind, I found it. Awesome. I couldn't see it because I thought it was a hash line, LOL. It is so straight that you almost can't tell it's not supposed to be there. It's actually 3-4 separate frequencies all working together.
If you can't figure out another way to remove it, I highly suggest iZotope RX 2 Advanced. If you can afford it. Or acquire it....
all gone.
-
I couldn't see it because I thought it was a hash line, LOL. It is so straight that you almost can't tell it's not supposed to be there. It's actually 3-4 separate frequencies all working together.
Yeah, once you get familiar with looking at the spectral view for stuff, you start to pick up singing and all sorts of stuff like that. Those razor straight lines are a giveaway once you realize what you're looking at compared to "normal".
I highly suggest iZotope RX 2 Advanced.
RX is (.)(.) ;D
Even the base version is a strong product, there are a bunch of things from Izotope that I think are really valuable.