|
Dictaphone software...
|
View this Thread in Original format
| Jackson |
Figured someone here might know.
I need to clean up some mp3 recordings of lectures. There's a lot of background noise that I want to try to get rid of. Anybody know of any software that I can use to do this? |
|
|
| EddieZilker |
Dragon takes dictation.
Wavosaur is free and will let you run a plug-in stack to work with audio. |
|
|
| Jackson |
| quote: | Originally posted by EddieZilker
Dragon takes dictation.
Wavosaur is free and will let you run a plug-in stack to work with audio. |
Just downloaded...wow, so many editing options (never used any sound software before).
Is this something that will take a while rather than just at the click of a few buttons? |
|
|
| ZeJayMan |
| Taking notes generally works:wtf: |
|
|
| Jackson |
| quote: | Originally posted by ZeJayMan
Taking notes generally works:wtf: |
Cant read notes when driving...or to fall asleep to.
Trying Soundsoap to see if its easier |
|
|
| EddieZilker |
| quote: | Originally posted by Jackson
Just downloaded...wow, so many editing options (never used any sound software before).
Is this something that will take a while rather than just at the click of a few buttons? |
It can take a while to isolate the speech from the sound you don't want - yes. The process you're trying to engage in is very roughly akin to a mastering process where you're filtering out the back-ground noise while focusing on amplifying the frequencies you want. The more noise you filter, however, the more likely it is you'll wind up with a recording where the prosody (the frequency range and cadence where the human speech is discernible as such but without clear distinction between words) outmatches consonance, sibilance, and other plosive sounds (that actually distinguish words and phrases). In the very best case scenario, you're going to wind up with a distinguishable recording featuring a lot of processing artifacts.
If you want, I could have a crack at processing the material. It's probably going to take me something on the order of 30 to 40 minutes what it might take you hours to do. |
|
|
| Jackson |
| Yeah that sounds awesome! Although there are like 25 I wanted to do lol. |
|
|
| EddieZilker |
| quote: | Originally posted by Jackson
Yeah that sounds awesome! Although there are like 25 I wanted to do lol. |
I'll do one and come up with a procedure for you to follow that shouldn't take too long. The alternative is first one free but I charge you $25.00 per hour, for the rest. :disbelief |
|
|
| Jackson |
Wow thank u so much!!!
Heres the audio:
[[ LINK REMOVED ]]
|
|
|
| EddieZilker |
This is a screen-cap of a female vocal I'm working with. You'll notice a nice, full spectrum with a couple of very small notches in the wave form where I had to lower a few frequencies in the H and K (as in 'H' at or 'K'at - words - lower notch - left to the center line between 2186 Hz and 4120 Hz) and S/T (higher notch - aka sibilance - as in Mi'ss'i'ss'ippi or 'T'op - to the left of the center line between 4720 & 10190 Hz).
Yours, up here, is completely shot dead.
Your wave form only peaks, incidentally, above 4720 Hz at a level of -70 db related to the -10 db peak located at the 469 Hz (250-750 Hz being a range for musically useful bass tones). Most of the time it's on the floor - meaning there are no dynamics in that range, to speak of. One can make out words that the speaker is saying, only occasionally, and processing to remove the sound of the person who is coughing, nearby, may also mean removing the remaining sounds which make the speaker's voice discernible above prosody.
I'll try to keep working on it but I've got other projects which are time-sensitive and I absolutely cannot guarantee that the end result is anything that will be more listenable. If anything, I might be able to get it cleaned up enough for a jury to make clear distinctions about what was said, during a trial, but it probably won't be anything you'll want to listen to (kind of like my last album - badump tssss). |
|
|
|
|
