🚀 go-pugleaf

In article <1992Apr8.185930.12694@berlioz.nsc.com> you write:
>The book says (on Pg. 8-40)
><WAVE-form>  ->  RIFF(  'WAVE'			//Form type
>			<fmt-ck>		// Waveform format
>			<data-ck> )		// Waveform data
>
><fmt-ck>  -> 	fmt(	<wave-format>		// WaveFormat Structure
>			<format-specific>	// Dependent on format category
>
><wave-format> -> struct {WORD formatTag;		// Format Category
>			 WORD nChannels;		// No. channels
>			 DWORD nSamplesPerSec;   // Sampling Rate
>			 DWORD nAvgBytesPerSec;  // For buffering
>			 WORD  nBlockAlign;}	// Block alignments
>
><PCM-format-specific> ->  struct {WORD   nBitsPerSample;}  //Sample Size
>
><data-ck> -> data( <wave-data> )

>As you can see, I found where the text is that says RIFF, WAVE, fmt, etc...
>But I am a bit confused because I can't find the parantheses, etc...

<smirk>  Okay.  In the RIFF specification documents, they say that the
notation  XXXX( <term> ...) means the following occurs in memory:

	XXXX   = 4 characters
        <size> = 4 byte length of chunk in bytes
        <term> = whatever data should follow

>I would really appreciate it if anyone could explain this differences to me
>or send me any current and more detailed information on the WAVE file format.  
>Since I rarely, if ever, read news, please send email to me personally...

>ding
>52 49 46 46 			RIFF      <- Right
>46 2D 0 0                                <- Length of rest of file
>57 41 56 45 			WAVE      <- Right
>66 6D 74 			fmt       <- Close it's  "fmt "(note the space)
>20 10 0 0 0 1 0 1 0 22 56 0              <- fmt chunk size (4 bytes)
>0 22 56 0 0 1 0 8 0                         fmt type (2 bytes)
					     num channels (2 bytes)
					     samples_per_sec (4 bytes)
					     ave_bytes_per_sec (4 bytes)
					     block_size (2 bytes)
					     bits_per_sample (2 bytes)
>64 61 74 61 			data      <- Right
>22 2D 0 0                                <- Length of this chunk
>80 80 80 80 80 80 80 80 80....           <- The Data

Hope that clears things up some.  If you are really curious, I wrote a program
from the specification to parse WAVE files and convert them to SND files.

Also, be aware that even though people usually format the file the way you have
it above that the RIFF spec does not state that the <fmt> chunk must precede
the <data> chunk.  Oddly enough, they can be reversed!  The size information
is provided so that people reading the file can know how much memory to
allocate to read in the chunk, and for people who are scanning the file to
skip past a chunk they are not interested in.  It is legal to put bitmaps
(and other stuff) in a RIFF Wave file!

Good luck,

Stephen.