[Speex-dev] FW: Re: Fwd: Re: Fixed Point on wideband-mode: Single Frame loss on 2000 Hz sine causes "freak off"

Frank Lorenz Frank_wtal at web.de
Fri Jan 15 05:23:49 PST 2010


Did not send this to speex-dev mailing list by error...

-----Ursprüngliche Nachricht-----
Von: "Frank Lorenz" <Frank_wtal at web.de>
Gesendet: 14.01.10 14:03:22
An: Jean-Marc Valin <jean-marc.valin at usherbrooke.ca>
Betreff: Re: [Speex-dev] Fwd: Re: Fixed Point on wideband-mode: Single Frame loss on 2000 Hz sine causes "freak off"
Hi Jean-Marc,

The codec mostly does not recover, it stays instable. I have attached a screenshot of the 2 waveforms of the output for quailty 7, complexity 2: The upper one is for a 2000 Hz, -15 dB input, the other one for 2200 Hz input. You see the frame loss 2 seconds after start, and approx. 1.5 seconds later this "freak out" starts. In this picuture, 2200 Hz looks worse, but in fact, I can stabilize the 2200 Hz output by setting complexity to >= 4.
In our real time system, we had the situation that the codec did only recover after we reduced the input signal to very small level.

One thing to note is the very different "fade in time" after the frame loss for the two frequencies. I do not get the point what happens here...

I will test for smaller values of cumul_gain (already tested a few vaules without success, but perhaps I missed a good one...)

best regards,
Frank


Jean-Marc Valin <jean-marc.valin at usherbrooke.ca> hat am 14. Januar 2010 um 13:15 geschrieben:

> What happens if you change that line:
>
> if (cumul_gain > 262144)
>
> to use a smaller value? What value works OK (if any)?
>
> One more thing, when things go wrong, do they eventually go back to
> normal or does the codec never recover? It's unavoidable that the audio
> goes bad for a short period of time because of the long-term predictor.
>
>         Jean-Marc
> 


______________________________________________________
GRATIS für alle WEB.DE-Nutzer: Die maxdome Movie-FLAT!
Jetzt freischalten unter http://movieflat.web.de



More information about the Speex-dev mailing list