[Subject Prev][Subject Next][Thread Prev][Thread Next][Date Index][Thread Index]

[hts-users:02054] Re: Problem Feature Extraction


Thanks, Heiga.
But I used Straight synthesizer but got also silence of the synthesized wave
and I found something strange in the mgc-lsp coefficient, the first
dimention of one frame is a strange number like -8 or -10

frame 1:
-11.9328
0.0764273
0.152513
0.226703
0.300016
0.379501
0.464736
0.543695
0.612467
0.68565
	0.767955
	0.845911
	0.916864
	0.991657
	1.06629
	1.14632
	1.22783
	1.30185
	1.37696
	1.45615
	1.53342
	1.61157
	1.68378
	1.75613
	1.83215
	1.91693
	1.99326
	2.06442
	2.14271
	2.22008
	2.30074
	2.3722
	2.44746
	2.52588
	2.60563
	2.68158
	2.76039
	2.8352
	2.91576
	2.98508
	3.06541

frame 2:
-8.43078
0.0728854
0.146196
0.220731
0.294602
0.373948
0.462149
0.538513
0.602996
0.671176
0.756476
0.838068
0.908864
0.99264
1.07358
1.14956
1.22247
1.2913
1.36791
1.44872
1.52548
1.61396
1.68564
1.75945
1.83578
1.92646
1.99256
2.0641
2.14209
2.22718
2.30408
2.37924
2.45094
2.52726
2.6021
2.68094
2.75865
2.83269
2.91605
2.98337
3.06966
-8.23503
0.069078
0.138786
0.215642
0.295721
0.374385
0.453902
0.528237
0.602931
0.683715
0.757493
0.82814
0.911094
0.991399
1.06001
1.14412
1.22597
1.29659
1.37006
1.45312
1.53067
1.60339
1.67846
1.76637
1.83279
1.9154
1.98356
2.07207
2.14471
2.22209
2.29533
2.37675
2.44591
2.53723
2.6035
2.68642
2.75861
2.83389
2.90947
2.99184
3.05938

2009/6/24 Heiga Zen (Byung Ha CHUN) <heiga.zen@xxxxxxxxxxxxxxxxx>
>
> Hi,
>
>
> Heiga ZEN (Byung Ha CHUN) wrote:
>
>> This is the reason why dynamic range of synthesized waveform is re-normalized in HTS-demo/data/scripts/synthesis.m as
>>
>> [sy] = exstraightsynth(f0,sp,ap,16000,prm);
>> wavwrite( sy/max(abs(sy)), 16000, 'temp.wav');
>
>
> Oops, this explanation is a bit wrong.  Anyway, MATLAB STRAIGHT normalizes the amplitude of input waveform so I recommend you to reconstruct waveform from generated parameters using MATLAB STRAIGHT and check whether the waveform is properly synthesized or not.
>
> Best regards,
>
>
> Heiga ZEN (Byung Ha CHUN)
>
> --
> --------------------------
> Heiga ZEN (Byung Ha CHUN)
> Speech Technology Group
> Cambridge Research Lab
> Toshiba Research Europe
> phone: +44 1223 436975
>
> ______________________________________________________________________
> This email has been scanned by the MessageLabs Email Security System.
> For more information please visit http://www.messagelabs.com/email ______________________________________________________________________
>



-- 
Xiang Wang
Institue Of Acoustics, CAS

Follow-Ups
[hts-users:02055] Re: Problem Feature Extraction, Heiga ZEN (Byung Ha CHUN)
References
[hts-users:02034] Re: Problem Feature Extraction, Geoffrey Wilfart
[hts-users:02038] Re: Problem Feature Extraction, wangxiang
[hts-users:02039] Re: Problem Feature Extraction, Keiichi Tokuda
[hts-users:02040] Re: Problem Feature Extraction, wangxiang
[hts-users:02041] Re: Problem Feature Extraction, Geoffrey Wilfart
[hts-users:02043] Re: Problem Feature Extraction, wangxiang
[hts-users:02045] Re: Problem Feature Extraction, Geoffrey Wilfart
[hts-users:02046] Re: Problem Feature Extraction, wangxiang
[hts-users:02050] Re: Problem Feature Extraction, Heiga ZEN (Byung Ha CHUN)
[hts-users:02051] Re: Problem Feature Extraction, Heiga Zen (Byung Ha CHUN)