[Speex-dev] Quality degradation on new versions

[Jean-Marc Valin]

Hi,

Aviv Tamar a écrit :
> Hi Jim,
> First of all - thanks, turning the highpass filter off was what I needed,
> and the waveforms
> match now.
> But, when i did the PESQ tests again I found an interesting result :
> version 1.0.5 still got
> a slightly better average score, but the standard deviation on version 1.2
> beta1 was much smaller.

I found 1.2betaX to have better quality than 1.0.5, so maybe you can
send an example for which it's not the case (a case where it's audible,
not where PESQ says so).

> The cause for that is this - on some samples versions 1.0.5 and 1.2beta2
> produced a single
> noticable 'click' during the speech, while version 1.2beta1 didn't.
> Do you have an idea what can cause the click, and how to turn it off?

Can you send a sample that produces that effect? Also, were you using
fixed-point or floating-point?

	Jean-Marc

> Aviv
> 
> 
> speex-dev-bounces at xiph.org wrote on 12/07/2007 21:26:51:
> 
>> Hi Aviv,
>>
>> Does the audio sound bad?
>>
>> You can try turning off the highpass filter (which was not in 1.0.5).
> This
>> code is from ti/testenc-TI-C5x.c in the source tree:
>>
>>    /* Turn this off if you want to measure SNR (on by default) */
>>    tmp=0;
>>    speex_encoder_ctl(st, SPEEX_SET_HIGHPASS, &tmp);
>>    speex_decoder_ctl(dec, SPEEX_SET_HIGHPASS, &tmp);
>>
>> - Jim
>>
>> ----- Original Message -----
>> From: "Aviv Tamar" <AVIVT at il.ibm.com>
>> To: <speex-dev at xiph.org>
>> Sent: Thursday, July 12, 2007 10:18 AM
>> Subject: [Speex-dev] Quality degradation on new versions
>>
>>
>>> Hi,
>>> I have been using speex version 1.0.5 on a text-to-speech program.
>>> Recently
>>> I upgraded to version 1.2beta1
>>> and noticed that the waveform the I got after encoding and decoding on
> the
>>> new versions (beta1,beta2) is much
>>> more different than the original than on version 1.0.5. I also ran a
> PESQ
>>> comparison test on 700 voice samples
>>> and got better results in the older version  (I used quality 9, and
>>> disabled perceptual enhancement).
>>> Do you know what can cause these changes?
>>> My application requires the waveform after decoding to be as similar in
>>> shape to the original, so if the difference
>>> is the result of some new addition to the encoder i would like to know
> how
>>> to turn it off.
>>> Thanks,
>>>
>>> Aviv
>>>
>>> _______________________________________________
>>> Speex-dev mailing list
>>> Speex-dev at xiph.org
>>> http://lists.xiph.org/mailman/listinfo/speex-dev
>>>
>>
>> _______________________________________________
>> Speex-dev mailing list
>> Speex-dev at xiph.org
>> http://lists.xiph.org/mailman/listinfo/speex-dev
> 
> _______________________________________________
> Speex-dev mailing list
> Speex-dev at xiph.org
> http://lists.xiph.org/mailman/listinfo/speex-dev
> 
>