Microsoft RIFF, WAVE audio data, 1 bit, mono 8000 Hz

Microsoft RIFF, WAVE audio data, 1 bit, mono 8000 Hz

[James H. Cloos Jr.]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I have a (free) voice mail service that provides the messages in a wav
file which file(1) specifies as:

        Microsoft RIFF, WAVE audio data, 1 bit, mono 8000 Hz

sox(1) doesn't recognise this file.  (I currently have
sox-12.15-5.i386.rpm installed, compiled on my box from the RH60 SRPM.)

Based on sox(1)'s wav.c, the fmt chunk size is 0x32, wFormatTag is
0x0022, wChannels is 1, samples per second is 8000, AvgBytesPerSec is
1067, wBlockAlign is 32, wBitsPerSample is 1 leaving 34 more octets of
header info.  They are (as chars):

        0x20,0x00,0x01,0x00,
        0xf0,0x00,0x00,0x00,
        0x00,0x00,0x00,0x00,
        0x00,0x00,0x00,0x00,
        0x00,0x00,0x00,0x00,
        0x00,0x00,0x20,0x00,
        0x00,0x00,0x00,0x00,
        0x00,0x00,0x00,0x00,
        0x00,0x00

So, does anyone have any tips on listening to this?

(And, yes, I have sent a note to the vendor about this; this note is
going out at the same time.)

- -JimC
- -- 
James H. Cloos, Jr.  <http://www.jhcloos.com/cloos/public_key> 1024D/ED7DAEA6 
<cloos@jhcloos.com>     E9E9 F828 61A4 6EA9 0F2B  63E7 997A 9F17 ED7D AEA6
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v0.9.7 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE3hEHCmXqfF+19rqYRAjzzAKC1L0DwupHBnr4M+12ihhc12eQ/CwCgwJ/l
zxE0KxVUn/CveWNOI8S/bXU=oKqX
-----END PGP SIGNATURE-----

Re: Microsoft RIFF, WAVE audio data, 1 bit, mono 8000 Hz

[Alan Cox]

> I have a (free) voice mail service that provides the messages in a wav
> file which file(1) specifies as:

Dial it up and send it a steady tone square wave at about 1KHz. 
If my guess to the format is right then you will get back something vaguely
like 
	FF,00,FF,00,FF,00,FF,00

THen you can figure the bitendianness they use. Basically its exactly what
it sounds like - single bit above/below threshold values.

So you'd do 

	while(len)
	{
		uchar i=*data++;
		int ct;
		for(ct=0;ct<8;ct++)
		{
			if(i&0x80)
				play8bit(0xFF);
			else
				play8bit(0x00);
			i<<=1;
		}
		len--;
	}

		

Re: Microsoft RIFF, WAVE audio data, 1 bit, mono 8000 Hz

[James H. Cloos Jr.]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

>>>>> "Eric" = Eric Mitchell <emitchell@altaira.com> replied:

Eric> There's a field in the wav header for any applied compression
Eric> algorithm.  (www.wotsit.com has decent file format docs).
Eric> What's that set to?  I don't know offhand of any linux .wav
Eric> players that deal correctly with compressed data.

Ah.  That helped.  From that file, the 0x0022 wFormatTag is
WAVE_FORMAT_DSPGROUP_TRUESPEECH.  Also, the 0x20,0x00 is the
size of the remaining raw header data I listed -- data that
is specific to the TRUESPEECH format.  The 0x01,0x00 is the
revision and the 0xf0,0x00 is the number of samples per block.

The rest of the info, esp including the format of the data and all
details of the compression are, according to that document, not
available.

Bummer.

Thanks for the pointer.  I'd managed to forget about wotsit.  *How* I
don't know.  But I had. :(

- -JimC
- -- 
James H. Cloos, Jr.  <http://www.jhcloos.com/cloos/public_key> 1024D/ED7DAEA6 
<cloos@jhcloos.com>     E9E9 F828 61A4 6EA9 0F2B  63E7 997A 9F17 ED7D AEA6

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v0.9.7 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE3hMEhmXqfF+19rqYRArOrAJ9/pstOEPVLKPO1dE9t97jvY8aUuACfZtV6
EPoOAw6cgePRU4wzOLXAyuI=bxZ6
-----END PGP SIGNATURE-----

Re: Microsoft RIFF, WAVE audio data, 1 bit, mono 8000 Hz

[James H. Cloos Jr.]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

>>>>> "Eric" = Eric Mitchell <emitchell@altaira.com> replied:

Eric> There's a field in the wav header for any applied compression
Eric> algorithm.  (www.wotsit.com has decent file format docs).
Eric> What's that set to?  I don't know offhand of any linux .wav
Eric> players that deal correctly with compressed data.

Ah.  That helped.  From that file, the 0x0022 wFormatTag is
WAVE_FORMAT_DSPGROUP_TRUESPEECH.  Also, the 0x20,0x00 is the
size of the remaining raw header data I listed -- data that
is specific to the TRUESPEECH format.  The 0x01,0x00 is the
revision and the 0xf0,0x00 is the number of samples per block.

The rest of the info, esp including the format of the data and all
details of the compression are, according to that document, not
available.

Bummer.

Thanks for the pointer.  I'd managed to forget about wotsit.  *How* I
don't know.  But I had. :(

- -JimC
- -- 
James H. Cloos, Jr.  <http://www.jhcloos.com/cloos/public_key> 1024D/ED7DAEA6 
<cloos@jhcloos.com>     E9E9 F828 61A4 6EA9 0F2B  63E7 997A 9F17 ED7D AEA6

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v0.9.7 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE3hMEhmXqfF+19rqYRArOrAJ9/pstOEPVLKPO1dE9t97jvY8aUuACfZtV6
EPoOAw6cgePRU4wzOLXAyuI=bxZ6
-----END PGP SIGNATURE-----


-- 
         To unsubscribe: mail sound-list-request@redhat.com with 
                       "unsubscribe" as the Subject.