Re: VPIM Minutes - IETF 50


jutta@panix.com
Thu, 12 Apr 2001 10:42:03 -0400 (EDT)


[James P. Salsman]
> If the receiver doesn't understand the audio/wav header, and tries
> to play it as raw GSM, it is short enough, but are there alignment
> issues; which is to ask: If the audio/wav header is prepended to
> the data it encapsulates, can it move the rest of the data out of
> codec frame alignment, distorting the audio decoding?
>
> Jutta: Do you know the answer to that question?

It's possible, but not worth it.

Two problems:

    (1) GSM frames themselves come in two incompatible
        flavors -- Microsoft WAV and plain (Berlin-toolkit style).
        Almost everything with a WAV header is Microsoft-style;
        almost everything without is Berlin-style.
        Arguably, allowing MS without a header only makes
        things worse and more confusing.

    (2) The MS GSM frame size is 65 bytes (containing two
        bit-packed GSM frames). Unless we add padding
        (a padding that wouldn't be added by non-VPIM GSM
        recorders), the WAV header will be smaller than
        that, and, yes, its presence would throw a reader
        out of alignment.

Other than that, having a "noisy" frame in the transmission would
not be a problem; although GSM does build on state established by
previous frames, it recovers quickly from errors.

Personally, I couldn't name a system that understands the MS WAV
format but not the MS WAV header; I don't think this is fixing
a problem that actually exists, while adding a fair amount of
confusion to the spec.

Cheers,

Jutta <jutta@pobox.com>



This archive was generated by hypermail 2.0b3 on Thu Apr 12 2001 - 17:42:15 IDT