Cleaning up recordings / recording formats

[oakcroft13]

In answer to Alan's question: the software package I use for sound 
editing is Syntrillion's Cool Edit, which I have found very powerful. 
It is shareware, but you can download a trial version and use it for 
30 days (with restrictions) before you have to purchase it. It is 
available from www.syntrillion.com. Its most useful facilities are:
* it can record via the PC soundcard from whatever recorder you use
* it can record from a directly attached microphone, especially 
useful for laptops
* it can translate between all the standard formats, especially .wav 
and .mp3
* it has extensive capabilties for editing, transformation, selective 
amplification, enveloping, noise removal etc.

There are other packages too, I'm only recommending this one because 
I use it and it works. If you only wish to digitise recordings, not 
edit them, Windows Sound Recorder (free!) will work fine, but there 
are some tricks not covered in the documentation required to make it 
use the file format you want. Also, you may have to download a 
seperate mp3 codec.

(I apologise to Macintosh users that I am only covering PC work here, 
perhaps someone who understands Macs better could join in!)

As regards recording formats, I always record at 44,100 samples per 
second, 16 bit data, mono. This can lead to very large files 
(fortunately I have a big hard-drive!). When I add recordings to my 
database I reformat them as mp3, which gives a reduction in file-size 
of over ten times. I also edit out initial silence, trim the 
recordings to length, etc. I archive the big .wav files off to CD in 
case I ever need them again.

Here are some thoughts on recording formats.

Sampling rate: the Nyquist theorem, grossly summarised, says that you 
need to sample at double the maximum frequency present in the sound. 
In practice I find with recordings you need to go higher than that. 
To sample at 22,050 runs the risk of losing frequencies somewhat 
below 10 kHz. Therefore, I always use 44,100. Do NOT digitise at 
48,000 as some Windows drivers have bugs at this sampling rate.

Number of bits: 8 bit data gives disappointing results, 32 bit data 
is complete overkill and is not universally supported. Use 16 bit.

Number of channels: except for recordings in exceptional 
circumstances (e.g. recording a swinging bell close up) I do not 
believe stereo adds any value - save the disc space and use mono.

File format: .wav files using PCM format are guaranteed to give you 
back exactly what went in, but at the expense of large amounts of 
disc (over 5 Mbyte per minute).
.wav files using the ADPCM format are also guaranteed to give you 
back almost exactly what went in, and are 25% of the size of PCM 
format files. In trials I conducted last year, I could detect no 
difference in the sound of PCM and ADPCM recordings. Disc usage is 
1.3 Mbyte / minute.
.mp3 files are very compressed. The format I use (56k bits/sec on the 
Cool Edit menu) reduces file sizes by a factor of 12. If the original 
recording is high quality then this mp3 format gives moderately good 
reproduction. However, if the original recording is distorted (i.e. 
the bells were too loud for microphone or recorder), or there is a 
big wash of background harmonics, mp3 gives a very poor result. Disc 
usage is about 420 kByte / minute.

The only exception to the above rules, of course, is if the original 
recording is of such poor quality (e.g. severely bandwidth limited) 
that digitising at 44,100 is not warranted.

I've written enough now, more (if people think it is useful) in a 
follow-up post on recorders, microphones, loudness, distortion, and 
how to get Windows Sound Recorder to record and save in your chosen 
file format.

Bill H