[Linux4christians] Audio tape to MP3 ideas

[Rob Mitchell]

Mike McMullin wrote:
<snip>
I'm looking at taking a 10 year catalogue of Sermons on Cassette Tape
and turning them into .mp3 files and archiving them on DVD's.  The OS'
involved will be both Linux and Windows (various flavours according to
the volunteers involved), and I was sort of figuring Audacity as the
common software of choice for all involved.  So my first question is,
has anyone dubbed from a tape source to an .mp3 file, and can you offer
some pointers?  </snip>

I only have experience with this question, not the other part of your
post.  My seminary study included several distance courses delivered 
on tapes.  I hated cassettes, but used Audacity to convert the tapes
to MP3s.  The bad news is that this not the highest fidelity way of 
digitizing audio; the good news is that because sermons are spoken 
word, they don't need to be as high quality as does music.

Here's the basic setup - you play the cassette on a tape deck, running
output from the preamp to the line in on your sound card.  Use whatever
sound recording application you wish - Audacity is Free and Open Source
and available for Windows and Linux, but there are other apps you can 
use.  For this purpose, however, Audacity is decent.

The default sampling rate for Audacity is 44100 Hz, which is fine for 
spoken word audio, but if you wish for better quality you can increase.
I suggest leaving it at this - the higher your sampling rate, the more
disk space you use, and since you're dubbing from a tape, chances are 
your source audio is not high enough quality to require a higher rate.

Once you have a project (an individual audio segment, which will generally
be a full sermon or a lecture) recorded, export it as an MP3 file.  You 
will need an encoder codec for this - the most common free encoder is 
called LAME, and you can use Synaptic to install if you are using a 
Debian/Ubuntu family distribution of Linux 
(use the GUI or "sudo apt-get install lame" in Bash).
You must have the encoder loaded before Audacity will export to MP3.
Once it is installed, you will be able to choose the Bit Rate of 
your MP3.  For spoken word, 64K will give you reasonably good quality
and reasonably small file size.  I've had good results with 32k.  Even
16k should accommodate spoken word well - the max frequency of the human
voice is about 8khz, and 16K is twice that, the minimum you need, but 
at that bit rate you will hear some artifact in the sound (it sounds 
like someone is talking through a high speed fan).  Stay at 32k or 64k.
For music you want 64k minimum - preferably 128 or 192 if you are using
digital input.

Adjust your audio individually for each tape you convert.  Watch
the waveform in the Audacity record window.  The wave form window
runs from 0 out to both +1 and -1.  You want your peaks to be out
at the ones, and your average waves to be between 0.5 and 1.  This
will give you loud enough volume without chopping off your peaks
(this is called flat-topping) which causes distortion (you can add 
this if you need to).

You might want to listen to these series on the Geeks and God
podcast, where you can learn a lot, but these guys tend to use
proprietary commercial tools rather than free ones, so ignore the 
brand names and listen to the tech discussion.

http://geeksandgod.com/podcast-series/podcasting-your-church
http://geeksandgod.com/podcast-series/audio-recording

Audacity includes a couple of Effects tools you may want to 
use.  Noise removal is a two-stage process - first you 
sample a silent stretch to get the ambient noise pattern,
then you apply noise removal that gets rid of most of that
background noise.  Use with caution - you don't have to keep
the change if it doesn't sound right.  The click removal
tool is actually for clicks and pops from vinyl records, 
but may help with mic pops as well.  Finally, you can 
edit out any long pauses, ahs and ers, plus (if you're
like me) any stupid things you said you'd like to get 
rid of, like that joke that didn't go over.  You edit out
just like a word processor - just highlight the segment 
you want to get rid of and hit delete.

If your track is offset, that is, if you notice that
your wave forms are higher on one side (toward +1 or 
-1) than the other, you can use Effects, Normalize to 
center them back up.  This gives a cleaner sound.

Another tool you will probably want to use is Effects, 
Compress.  This tool basically increases the volume of 
the quiet parts and decreases the volume of the louder 
parts, so that your volume is more uniform.  Actually
the compress tool in Audacity includes a normalize tool
that takes care of what is discussed in the previous 
paragraph as well as compressing.  Talk to the person
who runs your sound board about compression.

I only digitized a total of about 100 tapes this way, but 
it can be done.

It will occur to you to ask the question, "Can I run the 
tape faster to get the job done more quickly?"  The answer 
is, it depends.  If you have a really good quality tape deck 
that can output at full volume at a fixed increased speed, 
and you also have a really good quality sound card, then
the answer is yes.  If you have a lower quality deck and 
sound card (like the one integral to your system board) then
the answer is maybe. 

You must increase your sampling rate in Audacity by the same 
factor as your playback speed, e.g. if you playback at 200%, then
use 88000KHz as your sampling rate.  Use the Change Speed tool 
in Audacity (Effects, Change Speed) to slow the recorded audio
down, then remove noise, compress, and use effects as you would 
to produce an acceptable result.

You may be tempted to lower the pitch of your voice by using the 
pitch effect to give you a deeper voice, especially if your voice 
is reedy.  This is probably not a good idea - it is okay to use 
for comic effects, like if you want to do the voice of God saying, 
"Noah, I want you to build an ark," like in the 1970s Bill Cosby
routine.  The effect will probably be comic even if you don't
want it to be, so best bet is don't try to make your voice sound
different.  Also don't be tempted to use the echo effect or other
special effects - you will end up sounding cheesy and silly.
Instead, shoot for clean, clear audio at a uniform dynamic range.

Also, since you're digitizing your audio, why limit to burning the 
MP3 files to DVDs for posterity?  Put them on your website.  If your
hosting contract doesn't have enough space, that's okay - upload them
to http://www.archive.org - the only provision for their free unlimited
hosting is that you license your audio under Creative Commons and give
it away.  You can restrict derivative works and commercial use, and 
require attribution under Creative Commons.  The old "tape ministry" 
concept only generated a lot of income for really popular preachers
who either had really, really good content or were really, really 
good at self-promotion.  

Don't worry about people stealing your 
material - your problem is probably not infringement; if you're
like most preachers it's obscurity.  Give your stuff away.  

Hope this helps.

Rob in Memphis