Maker Pro
Maker Pro

Will mini MP3 recorders/players do subaudio?

R

Robert Martin

Jan 1, 1970
0
Can anyone advise if the capsule shaped 128MB USB connect MP3
recorder/players, commonly sold for under $100, will record and
playback frequencies in the 5-20 Hz range.

Of course, headphones would not be used in this application.

If not, what would be the most cost effective design for doing this
on-board and where might I find a schematic?

Robert
SCU
 
A

artie

Jan 1, 1970
0
Robert Martin said:
Can anyone advise if the capsule shaped 128MB USB connect MP3
recorder/players, commonly sold for under $100, will record and
playback frequencies in the 5-20 Hz range.

Of course, headphones would not be used in this application.

If not, what would be the most cost effective design for doing this
on-board and where might I find a schematic?

Robert
SCU

The problem isn't with the playback side, but with the encoding side.
It sounds like you want to use this for data recording. If you're
doing a single sine wave, pretty much full scale as far as amplitude
goes, and varying in frequency, I'd guess that most mp3 encoders will
encode it fine.

You could do some experiments using freeware/shareware audio editing
software. On Windoze, Cool Edit can generate tones, and save them as
mp3 files. Try encoding some test signals and then decoding them.
 
J

Jan Panteltje

Jan 1, 1970
0
Can anyone advise if the capsule shaped 128MB USB connect MP3
recorder/players, commonly sold for under $100, will record and
playback frequencies in the 5-20 Hz range.

Of course, headphones would not be used in this application.

If not, what would be the most cost effective design for doing this
on-board and where might I find a schematic?

Robert
SCU
I would use a Philips PCF8591 chip, connected with 4 wires to the parport
of the PC (scl, sda in, sda pulldown).
You can write the soft to do 30 samples / second for example, using the PC
timer, and record on the PC.

In 32kB you can then store 262144 bits / 12 = 21845 samples.
21845 / 30 = 728 seconds, so say 12 minutes.

If you need standalone small portable, a PIC with AD and some SRAM at
30 samples/s say 12 bits even for a 32 kB RAM would be 16 x 1024 seconds,
/ 100 so good for 160 seconds, 2.5 minutes.
More SRAM more minutes...
And would use little power, and cost next to nothing (a few $).
Youd have to make a 25 Hz lowpass in record to get rid of aliasing if
other frequencies are present.
If I did the math right.
JP
 
J

Jan Panteltje

Jan 1, 1970
0
On a sunny day (Sun, 04 Apr 2004 08:06:50 +1000) it happened Robert Martin

OK, I wanted to write (messed up the text):

I would use a Philips PCF8591 chip, connected with 4 wires to the parport
of the PC (scl, sda in, sda pulldown).
You can write the soft to do 30 samples / second for example, using the PC
timer, and record on the PC.

If you need standalone small portable, a PIC with AD and some SRAM at
30 samples/s say 12 bits even for a 32 kB RAM.
In 32kB you can then store 262144 bits / 12 = 21845 samples.
21845 / 30 = 728 seconds, so say 12 minutes.

More SRAM more minutes...
And would use little power, and cost next to nothing (a few $).
Youd have to make a 25 Hz lowpass in record to get rid of aliasing if
other frequencies are present.
If I did the math right.
JP
 
I

Ian Stirling

Jan 1, 1970
0
Robert Martin said:
Can anyone advise if the capsule shaped 128MB USB connect MP3
recorder/players, commonly sold for under $100, will record and
playback frequencies in the 5-20 Hz range.

Maybe.
At this sort of frequency, you run into problems with capacitive
coupling, and the encoder deciding that you can't hear 5Hz, so
leaving it out.

Buy one for evaluation, if it works, use it.
Some with other than MP3 voice encoding (ADPCM, ...) might suit
better.
I'd test the one I have here, but it'd mean searching for the line-in
jack, which I've misplaced.
 
J

James Meyer

Jan 1, 1970
0
Can anyone advise if the capsule shaped 128MB USB connect MP3
recorder/players, commonly sold for under $100, will record and
playback frequencies in the 5-20 Hz range.

Of course, headphones would not be used in this application.

If not, what would be the most cost effective design for doing this
on-board and where might I find a schematic?

Robert
SCU

If you use a voltage to frequency converter between the source and the
MP3 unit, you could extend the response down to zero Hz.

With a zero offset so that zero volts in gives you 10 KHz out and a
scale factor of 1 KHz per volt you might get pretty good recording from zero Hz
to 5 KHz. An anti-aliasing would be rrequired as well as a matching frequency
to voltage converter for playback.

Jim
 
J

Jan Panteltje

Jan 1, 1970
0
On a sunny day (Sun, 04 Apr 2004 08:06:50 +1000) it happened Robert Martin

OK, I wanted to write (messed up the text):

I would use a Philips PCF8591 chip, connected with 4 wires to the parport
of the PC (scl, sda in, sda pulldown).
You can write the soft to do 30 samples / second for example, using the PC
timer, and record on the PC.

If you need standalone small portable, a PIC with AD and some SRAM at
30 samples/s say 12 bits even for a 32 kB RAM.
In 32kB you can then store 262144 bits / 12 = 21845 samples.
21845 / 30 = 728 seconds, so say 12 minutes.
Still goofed here, for 20 Hz in you need to sample at 50 Hz, so 7 minutes
because of Niquist.
Must have been tired last night...
Watched too much alien movie perhaps ;-) and that was a good one :)
JP
 
I

Ian Stirling

Jan 1, 1970
0
James Meyer said:
If you use a voltage to frequency converter between the source and the
MP3 unit, you could extend the response down to zero Hz.

With a zero offset so that zero volts in gives you 10 KHz out and a
scale factor of 1 KHz per volt you might get pretty good recording from zero Hz
to 5 KHz. An anti-aliasing would be rrequired as well as a matching frequency
to voltage converter for playback.

I would expect that voltage to amplitude modulation at say 2Khz might be
more suitable.
I suspect that you'd tend to get distortion caused by the quantisation
of the frequency bins.
 
N

Nico Coesel

Jan 1, 1970
0
Robert Martin said:
Can anyone advise if the capsule shaped 128MB USB connect MP3
recorder/players, commonly sold for under $100, will record and
playback frequencies in the 5-20 Hz range.

I think they will, but you may need to do some minor modifications. I
suppose the mini MP3 players will have RC filters that may not pass
these low frequencies. These filters can be modified quite easely.
 
J

James Meyer

Jan 1, 1970
0
I would expect that voltage to amplitude modulation at say 2Khz might be
more suitable.
I suspect that you'd tend to get distortion caused by the quantisation
of the frequency bins.

Maintaining calibration of an amplitude modulated scheme might be more
problematical than a frequency modulated scheme. An amplitude based system
would suffer degradation if the amplitude of the signal out of the MP# recorder
wasn't constant. On the other hand, a 10 KHz signal is 10 KHz even if it is
distorted and severely attenuated.

I was envisioning a (more or less) continuous, analog,
modulator/demodulator system.

Jim
 
I

Ian Stirling

Jan 1, 1970
0
James Meyer said:
Maintaining calibration of an amplitude modulated scheme might be more
problematical than a frequency modulated scheme. An amplitude based system
would suffer degradation if the amplitude of the signal out of the MP# recorder
wasn't constant. On the other hand, a 10 KHz signal is 10 KHz even if it is
distorted and severely attenuated.

I was envisioning a (more or less) continuous, analog,
modulator/demodulator system.

But, you've got an A/D, a complex codec, a D/A and a complex decoder in there.
I was talking about the codec having problems with a slewing frequency,
which is practically never present in any normal audio signal, and much
less likely to have problems with a simple AM signal, as that's
practically what it's been designed to encode.
 
Top