[Speex-dev] Signaling tones in Speex
Jean-Marc Valin
jean-marc.valin at usherbrooke.ca
Mon May 7 04:01:50 PDT 2007
Oh, hadn't realised you meant high-bitrate fax... At these rates, I'm
not quite Speex will be able to do much because the signal is pretty
much noise. I'd still be curious to hear how well/bad it does, though.
Jean-Marc
Mainak Chakraborty a écrit :
> In case a system is incapable of fax relay or if it is disabled, one of the
> easiest and safest options is to go for 40 kbps ADPCM compression (for fax
> upto 14.4 kbps)..even am new to this problem and the fair bit of seraching
> which i've done seems to suggest that the standard sloutions are to simply
> 'bypass' it else compress using ADPCM (40 k for fax upto 14.4 k, 32 k for
> fax upto 9.6 k and so on), both of which eat up a lot of bandwidth. So,
> speex sounds great!
> Once I start playing with the non-speech traffic will keep u posted.
> Cheers,
> Mainak
>
> On 5/7/07, Jean-Marc Valin <jean-marc.valin at usherbrooke.ca> wrote:
>>
>> Mainak Chakraborty wrote:
>> > Hi Jean,
>> > Thats great news for me to start off with as I was planning to go with
>> > 16 Kbps ADPCM keeping in mind the issues and options I had. Now,
>> whether
>> > the additional computation cost is worth the significant bandwidth
>> > savings, I have to see.
>> > Just wondering if it is possible to extend this logic to G3 and G4 fax
>> > as well, i.e. using a higher bit-rate and complexity mode for modem or
>> > fax instead of going for a higher bit-rate ADPCM.
>> > If so, a one stop solution for all kinds of traffic would be awesome.
>>
>> Not sure what these faxes are. It depends on how sophisticated they are.
>> In any case, I'm pretty sure that regardless of the test, Speex will
>> come up way ahead of ADPCM at any of the bit-rates it supports. BTW,
>> if/when you test Speex on DTMF and fax, please post the results here as
>> I'd be quite curious to see how it performs.
>>
>> Cheers,
>>
>> Jean-Marc
>>
>> > Thanks a lot for the prompt reply,
>> > Mainak
>> >
>> > On 5/4/07, *Jean-Marc Valin* <jean-marc.valin at usherbrooke.ca
>> > <mailto:jean-marc.valin at usherbrooke.ca>> wrote:
>> >
>> > Hi Mainak,
>> >
>> > Speex will definitely be able to handle DTMF. The only question is
>> what
>> > the minimum rate for that will be. I remember testing on a few
>> tones
>> at
>> > 8 kbps and I didn't hear too much distortion (though there have
>> have
>> > been). The only thing I can recommend is to use a higher complexity
>> > setting that one would normally use. For voice, I can't really tell
>> the
>> > difference between quality 1 and quality 3, but for DTMF, it
>> actually
>> > makes a big difference. I'd recommend using complexity 3 or 4 if
>> you
>> can
>> > afford it (in speexenc, you can change that with --comp and the
>> default
>> > is 3).
>> >
>> > Jean-Marc
>> >
>> > Mainak Chakraborty a écrit :
>> > > Hi all,
>> > > Has anyone tried conveying signaling tones (DTMF) through speex?
>> > If so, can
>> > > i get an idea of the lowest bit-rate at which I can do so with
>> > 'tolerable
>> > > distortion'. G.728 does so and G.729E (11.8 Kbps mode) is
>> 'claimed' to
>> > > do so
>> > > but even if true, G.729 (rather G.729I - 11.8/8/6.4 kbps) does
>> not
>> > allow me
>> > > to go below 6.4 kbps if needed. Basically I am looking for a
>> > single codec
>> > > which would allow me to play with multiple (low) bit-rates within
>> > voice and
>> > > switch to a higher bit-rate upon detecting a signaling tone
>> instead of
>> > > using
>> > > RFC 2833/out-of-band signaling methods. I hope speex is the
>> > solution am
>> > > looking for!
>> > > Regards,
>> > > Mainak
>> > >
>> > > PS: In any case I will have to test it out myself but a positive
>> > answer to
>> > > this query would definelty brighten up the road ahead :-)
>> > >
>> > >
>> > >
>> >
>> ------------------------------------------------------------------------
>> > >
>> > > _______________________________________________
>> > > Speex-dev mailing list
>> > > Speex-dev at xiph.org <mailto:Speex-dev at xiph.org>
>> > > http://lists.xiph.org/mailman/listinfo/speex-dev
>> >
>> >
>>
>
More information about the Speex-dev
mailing list