Using Audacity for a series of interviews, conversations and some other recorded dialogue tasks. Some of the audio has the voice very faint in the background, I can amp it to a n audible level but there is all that blechy garbage noise in the foreground. How do i extract what i want from these recordings? or clean up the noise? I have played with Noise reduction and many filters but to no avail. Audacity Noob by the way first post/time on this forum any and all help appreciated, I’ll check back later. If it helps one recording is a kindergarten play in a cafeteria, one an interview with a family member and several surveillance recordings of employees at a nursing home I supervise. All originally recorded as DSS files then converted to MP3 before editing in Audacity. If another audio format would serve my purposes better than MP3 let me know and again thanks.
The Audacity repair tools have limited usefulness for wild recordings. We can’t “dig a voice” out of trash, particularly if the trash is uncontrolled and louder than the voice. If you’re going to create a production, you do have to meet us half-way with a good microphone and reasonably quiet environment.
Also, since MP3 creates distortion and you can’t stop it, you should use WAV (Microsoft) 16-bit for all capture and mastering. If you need to, you can make an MP3 later, for playback in your portable music device or posting on-line. Fair warning, WAV may be a perfect format and good for editing, but its sound files are also many times larger than MP3. MP3 small files aren’t “free.”
some other recorded dialogue tasks.
Audacity can’t be used for surveillance, conflict resolution or law enforcement.
Koz
There’s not a whole lot you can do because voice is comprised of the same sounds/frequencies as the noise.
You can try some Equalization. For “experimenting” I recommend the Graphic EQ mode. The low frequencies (bass) are on the left and high-frequencies on the right. You can pull the sliders below ~100z all the way down because there is no useful low-frequency audio and it’s just noise. The main voice frequencies are around the middle so you can try pushing-up and pulling-down sliders one at a time to see if you can make an improvement. Boosting the higher frequencies can bring-out the “T” and “S” sounds to improve intelligibility, but cutting extreme-high frequencies may reduce some noise.
After boosting anything it’s a good idea to run Amplify (or Normalize). If you allow the default settings, either of these will bring down the overall volume (if necessary) to prevent clipping (additional distortion), or if the levels are low the defaults will bring the volume up. Run one of these effects before exporting the audio because that’s when it gets clipped and you can’t fix it later.
Beyond that, your brain is usually the best “filter” and if you can understand what’s being said you can make a typed transcript.
The “damage” done by MP3 is probably insignificant compared to the noise.
I don’t think Audacity will open .DSS files : it’s a proprietary format (used by Olympus),
but there are DSS->WAV online converters,
if you don’t mind sending the audio across the internet where it could be corrupted/intercepted.
In reality processing the audio does usually does not make the meaningful speech pop-out of incomprehensible noise,
like on CSI.
Thanks, ill try the conversion again to wma to see if its easier to pick apart in the EQ. @Trebar almost nothing messes with DSS files except Olympus software which you cannot use if device was obtained 2nd hand which was very frustrating in the beginning for me i have manageed to find one sound converter that does the trick but never thought of cleaning it up prior to conversion, rookie move lol this is a new medium for me but I’m diggin it.
Better to convert it to WAV if possible (Converting to WMA will reduce the sound quality a bit).
I did read NCH audio-editor, (which is allegedly free), can open some .DSS , but I’ve never tried it.
Olympus has free DSS player software, you could use Audacity to record the output of that …
