[hts-users:01782] Re: Improve audio quality

["Jenny Palmy" <jenny.palmy@xxxxxxxxx>]

Thanks Simon and Pieter for your help.

I generated these sounds by Dr. Daniel Erro's voice convertion toolkit
http://gps-tsc.upc.es/veu/soft/soft/vc_toolkit/, it is an excellent
toolkit. And I want to use the outputs in my application. However, the
buzziness is not satisfing. So I try to find some postprocessing
method to remove the buzziness. Could you give me some idea about
this?

Cheers,
Jenny

On Thu, Nov 13, 2008 at 5:53 PM, Pieter Scholtz <pscholtz@xxxxxxxxxxxxx> wrote:
>
> Hi Jenny
>
>
>
> That really depends on how the speech was generated. So more details would be handy. I would try postfiltering, or increasing the postfiltering intensity. It's hts_engine's beta parameter. You could also try using parameter generation considering global variance, although this is only really effective if you use MLSA, not so good for LSP. Using STRAIGHT would also improve quality. Also a mixed excitation scheme to eliminate the buzziness.
>
>
>
> Cheers
>
> Pieter
>
>
>
>
>
>
>
> On Thu, 13 Nov 2008 15:37:35 +0800, "Jenny Palmy" <jenny.palmy@xxxxxxxxx> wrote:
>
>> Dear All,
>
>>
>
>> Two speech audios are attached. Their quality are not very good. Could
>
>> anyone give me some idea on how to improve their quality? Is there
>
>> some software or function I can use?
>
>>
>
>> Cheers,
>
>> Jenny
>
>>
>
>>
>
>
>