diff --git a/docs/audio/female/sample-1/golf_p361_004.wav b/docs/audio/female/sample-1/golf_p361_004.wav new file mode 100644 index 0000000..1eca94b Binary files /dev/null and b/docs/audio/female/sample-1/golf_p361_004.wav differ diff --git a/docs/audio/female/sample-1/gt_p361_004.wav b/docs/audio/female/sample-1/gt_p361_004.wav new file mode 100644 index 0000000..2a50bd7 Binary files /dev/null and b/docs/audio/female/sample-1/gt_p361_004.wav differ diff --git a/docs/audio/female/sample-1/nhv_p361_004.wav b/docs/audio/female/sample-1/nhv_p361_004.wav new file mode 100644 index 0000000..514766a Binary files /dev/null and b/docs/audio/female/sample-1/nhv_p361_004.wav differ diff --git a/docs/audio/female/sample-1/pysptk_p361_004.wav b/docs/audio/female/sample-1/pysptk_p361_004.wav new file mode 100644 index 0000000..306eec0 Binary files /dev/null and b/docs/audio/female/sample-1/pysptk_p361_004.wav differ diff --git a/docs/audio/female/sample-1/world_p361_004.wav b/docs/audio/female/sample-1/world_p361_004.wav new file mode 100644 index 0000000..d8b174c Binary files /dev/null and b/docs/audio/female/sample-1/world_p361_004.wav differ diff --git a/docs/audio/female/sample-2/golf_p361_006.wav b/docs/audio/female/sample-2/golf_p361_006.wav new file mode 100644 index 0000000..f14a3b3 Binary files /dev/null and b/docs/audio/female/sample-2/golf_p361_006.wav differ diff --git a/docs/audio/female/sample-2/gt_p361_006.wav b/docs/audio/female/sample-2/gt_p361_006.wav new file mode 100644 index 0000000..757f7aa Binary files /dev/null and b/docs/audio/female/sample-2/gt_p361_006.wav differ diff --git a/docs/audio/female/sample-2/nhv_p361_006.wav b/docs/audio/female/sample-2/nhv_p361_006.wav new file mode 100644 index 0000000..e29de06 Binary files /dev/null and b/docs/audio/female/sample-2/nhv_p361_006.wav differ diff --git a/docs/audio/female/sample-2/pysptk_p361_006.wav b/docs/audio/female/sample-2/pysptk_p361_006.wav new file mode 100644 index 0000000..fc33c68 Binary files /dev/null and b/docs/audio/female/sample-2/pysptk_p361_006.wav differ diff --git a/docs/audio/female/sample-2/world_p361_006.wav b/docs/audio/female/sample-2/world_p361_006.wav new file mode 100644 index 0000000..ee0e2f8 Binary files /dev/null and b/docs/audio/female/sample-2/world_p361_006.wav differ diff --git a/docs/audio/female/sample-3/golf_p361_008.wav b/docs/audio/female/sample-3/golf_p361_008.wav new file mode 100644 index 0000000..d09fa4e Binary files /dev/null and b/docs/audio/female/sample-3/golf_p361_008.wav differ diff --git a/docs/audio/female/sample-3/gt_p361_008.wav b/docs/audio/female/sample-3/gt_p361_008.wav new file mode 100644 index 0000000..b3e5853 Binary files /dev/null and b/docs/audio/female/sample-3/gt_p361_008.wav differ diff --git a/docs/audio/female/sample-3/nhv_p361_008.wav b/docs/audio/female/sample-3/nhv_p361_008.wav new file mode 100644 index 0000000..ee08c35 Binary files /dev/null and b/docs/audio/female/sample-3/nhv_p361_008.wav differ diff --git a/docs/audio/female/sample-3/pysptk_p361_008.wav b/docs/audio/female/sample-3/pysptk_p361_008.wav new file mode 100644 index 0000000..07afdc2 Binary files /dev/null and b/docs/audio/female/sample-3/pysptk_p361_008.wav differ diff --git a/docs/audio/female/sample-3/world_p361_008.wav b/docs/audio/female/sample-3/world_p361_008.wav new file mode 100644 index 0000000..b0b410b Binary files /dev/null and b/docs/audio/female/sample-3/world_p361_008.wav differ diff --git a/docs/audio/female/sample-4/golf_p361_011.wav b/docs/audio/female/sample-4/golf_p361_011.wav new file mode 100644 index 0000000..fb7563b Binary files /dev/null and b/docs/audio/female/sample-4/golf_p361_011.wav differ diff --git a/docs/audio/female/sample-4/gt_p361_011.wav b/docs/audio/female/sample-4/gt_p361_011.wav new file mode 100644 index 0000000..95cbef5 Binary files /dev/null and b/docs/audio/female/sample-4/gt_p361_011.wav differ diff --git a/docs/audio/female/sample-4/nhv_p361_011.wav b/docs/audio/female/sample-4/nhv_p361_011.wav new file mode 100644 index 0000000..36e927a Binary files /dev/null and b/docs/audio/female/sample-4/nhv_p361_011.wav differ diff --git a/docs/audio/female/sample-4/pysptk_p361_011.wav b/docs/audio/female/sample-4/pysptk_p361_011.wav new file mode 100644 index 0000000..e5a9baf Binary files /dev/null and b/docs/audio/female/sample-4/pysptk_p361_011.wav differ diff --git a/docs/audio/female/sample-4/world_p361_011.wav b/docs/audio/female/sample-4/world_p361_011.wav new file mode 100644 index 0000000..b2ba6c0 Binary files /dev/null and b/docs/audio/female/sample-4/world_p361_011.wav differ diff --git a/docs/audio/female/sample-5/golf_p361_024.wav b/docs/audio/female/sample-5/golf_p361_024.wav new file mode 100644 index 0000000..4f1488c Binary files /dev/null and b/docs/audio/female/sample-5/golf_p361_024.wav differ diff --git a/docs/audio/female/sample-5/gt_p361_024.wav b/docs/audio/female/sample-5/gt_p361_024.wav new file mode 100644 index 0000000..7bf8f2c Binary files /dev/null and b/docs/audio/female/sample-5/gt_p361_024.wav differ diff --git a/docs/audio/female/sample-5/nhv_p361_024.wav b/docs/audio/female/sample-5/nhv_p361_024.wav new file mode 100644 index 0000000..d5ebdbe Binary files /dev/null and b/docs/audio/female/sample-5/nhv_p361_024.wav differ diff --git a/docs/audio/female/sample-5/pysptk_p361_024.wav b/docs/audio/female/sample-5/pysptk_p361_024.wav new file mode 100644 index 0000000..ce9e773 Binary files /dev/null and b/docs/audio/female/sample-5/pysptk_p361_024.wav differ diff --git a/docs/audio/female/sample-5/world_p361_024.wav b/docs/audio/female/sample-5/world_p361_024.wav new file mode 100644 index 0000000..ecc5d70 Binary files /dev/null and b/docs/audio/female/sample-5/world_p361_024.wav differ diff --git a/docs/audio/male/sample-1/golf_p360_003.wav b/docs/audio/male/sample-1/golf_p360_003.wav new file mode 100644 index 0000000..15dd0ee Binary files /dev/null and b/docs/audio/male/sample-1/golf_p360_003.wav differ diff --git a/docs/audio/male/sample-1/gt_p360_003.wav b/docs/audio/male/sample-1/gt_p360_003.wav new file mode 100644 index 0000000..c31f1ac Binary files /dev/null and b/docs/audio/male/sample-1/gt_p360_003.wav differ diff --git a/docs/audio/male/sample-1/nhv_p360_003.wav b/docs/audio/male/sample-1/nhv_p360_003.wav new file mode 100644 index 0000000..9d11640 Binary files /dev/null and b/docs/audio/male/sample-1/nhv_p360_003.wav differ diff --git a/docs/audio/male/sample-1/pysptk_p360_003.wav b/docs/audio/male/sample-1/pysptk_p360_003.wav new file mode 100644 index 0000000..263dda9 Binary files /dev/null and b/docs/audio/male/sample-1/pysptk_p360_003.wav differ diff --git a/docs/audio/male/sample-1/world_p360_003.wav b/docs/audio/male/sample-1/world_p360_003.wav new file mode 100644 index 0000000..5186bcc Binary files /dev/null and b/docs/audio/male/sample-1/world_p360_003.wav differ diff --git a/docs/audio/male/sample-2/golf_p360_005.wav b/docs/audio/male/sample-2/golf_p360_005.wav new file mode 100644 index 0000000..1082df4 Binary files /dev/null and b/docs/audio/male/sample-2/golf_p360_005.wav differ diff --git a/docs/audio/male/sample-2/gt_p360_005.wav b/docs/audio/male/sample-2/gt_p360_005.wav new file mode 100644 index 0000000..5132c9a Binary files /dev/null and b/docs/audio/male/sample-2/gt_p360_005.wav differ diff --git a/docs/audio/male/sample-2/nhv_p360_005.wav b/docs/audio/male/sample-2/nhv_p360_005.wav new file mode 100644 index 0000000..533cfd6 Binary files /dev/null and b/docs/audio/male/sample-2/nhv_p360_005.wav differ diff --git a/docs/audio/male/sample-2/pysptk_p360_005.wav b/docs/audio/male/sample-2/pysptk_p360_005.wav new file mode 100644 index 0000000..cef9930 Binary files /dev/null and b/docs/audio/male/sample-2/pysptk_p360_005.wav differ diff --git a/docs/audio/male/sample-2/world_p360_005.wav b/docs/audio/male/sample-2/world_p360_005.wav new file mode 100644 index 0000000..88c1916 Binary files /dev/null and b/docs/audio/male/sample-2/world_p360_005.wav differ diff --git a/docs/audio/male/sample-3/golf_p360_007.wav b/docs/audio/male/sample-3/golf_p360_007.wav new file mode 100644 index 0000000..3432c4f Binary files /dev/null and b/docs/audio/male/sample-3/golf_p360_007.wav differ diff --git a/docs/audio/male/sample-3/gt_p360_007.wav b/docs/audio/male/sample-3/gt_p360_007.wav new file mode 100644 index 0000000..925ba41 Binary files /dev/null and b/docs/audio/male/sample-3/gt_p360_007.wav differ diff --git a/docs/audio/male/sample-3/nhv_p360_007.wav b/docs/audio/male/sample-3/nhv_p360_007.wav new file mode 100644 index 0000000..ce0d54f Binary files /dev/null and b/docs/audio/male/sample-3/nhv_p360_007.wav differ diff --git a/docs/audio/male/sample-3/pysptk_p360_007.wav b/docs/audio/male/sample-3/pysptk_p360_007.wav new file mode 100644 index 0000000..483bfe9 Binary files /dev/null and b/docs/audio/male/sample-3/pysptk_p360_007.wav differ diff --git a/docs/audio/male/sample-3/world_p360_007.wav b/docs/audio/male/sample-3/world_p360_007.wav new file mode 100644 index 0000000..1730776 Binary files /dev/null and b/docs/audio/male/sample-3/world_p360_007.wav differ diff --git a/docs/audio/male/sample-4/golf_p360_009.wav b/docs/audio/male/sample-4/golf_p360_009.wav new file mode 100644 index 0000000..d10c1df Binary files /dev/null and b/docs/audio/male/sample-4/golf_p360_009.wav differ diff --git a/docs/audio/male/sample-4/gt_p360_009.wav b/docs/audio/male/sample-4/gt_p360_009.wav new file mode 100644 index 0000000..b06a336 Binary files /dev/null and b/docs/audio/male/sample-4/gt_p360_009.wav differ diff --git a/docs/audio/male/sample-4/nhv_p360_009.wav b/docs/audio/male/sample-4/nhv_p360_009.wav new file mode 100644 index 0000000..98d1c81 Binary files /dev/null and b/docs/audio/male/sample-4/nhv_p360_009.wav differ diff --git a/docs/audio/male/sample-4/pysptk_p360_009.wav b/docs/audio/male/sample-4/pysptk_p360_009.wav new file mode 100644 index 0000000..0d072fa Binary files /dev/null and b/docs/audio/male/sample-4/pysptk_p360_009.wav differ diff --git a/docs/audio/male/sample-4/world_p360_009.wav b/docs/audio/male/sample-4/world_p360_009.wav new file mode 100644 index 0000000..5794f64 Binary files /dev/null and b/docs/audio/male/sample-4/world_p360_009.wav differ diff --git a/docs/audio/male/sample-5/golf_p360_019.wav b/docs/audio/male/sample-5/golf_p360_019.wav new file mode 100644 index 0000000..8363a17 Binary files /dev/null and b/docs/audio/male/sample-5/golf_p360_019.wav differ diff --git a/docs/audio/male/sample-5/gt_p360_019.wav b/docs/audio/male/sample-5/gt_p360_019.wav new file mode 100644 index 0000000..6496e91 Binary files /dev/null and b/docs/audio/male/sample-5/gt_p360_019.wav differ diff --git a/docs/audio/male/sample-5/nhv_p360_019.wav b/docs/audio/male/sample-5/nhv_p360_019.wav new file mode 100644 index 0000000..b090090 Binary files /dev/null and b/docs/audio/male/sample-5/nhv_p360_019.wav differ diff --git a/docs/audio/male/sample-5/pysptk_p360_019.wav b/docs/audio/male/sample-5/pysptk_p360_019.wav new file mode 100644 index 0000000..dbc3f11 Binary files /dev/null and b/docs/audio/male/sample-5/pysptk_p360_019.wav differ diff --git a/docs/audio/male/sample-5/world_p360_019.wav b/docs/audio/male/sample-5/world_p360_019.wav new file mode 100644 index 0000000..a1d0c6d Binary files /dev/null and b/docs/audio/male/sample-5/world_p360_019.wav differ diff --git a/docs/index.html b/docs/index.html index 9c0c9d7..64f5dbf 100644 --- a/docs/index.html +++ b/docs/index.html @@ -2,6 +2,15 @@ + + + @@ -66,7 +75,8 @@

Abstract



Figure 1: Flow diagram of the proposed end-to-end analysis-by-synthesis - experiment.

+ experiment. S(z) is the speech, G(z) is the harmonic oscillator, C(z) is the noise filter, and + H(z) is the vocal tract filter.

@@ -78,15 +88,14 @@

Listening Samples

This section contains the listening samples from the VCTK Corpus used for subjective evaluation in - the paper. - Reference is the ground truth audio, and the other columns are the copy-synthesis speech using the - same clip with different models, which are the traditional linear predictive coding (LPC), - the neural homomorphic vocoder (NHV), the differentiable WORLD vocoder (▽World), - and the proposed source-filter GOLF vocoder (GOLF). + the paper. Reference is the ground truth audio, and the other columns are the copy-synthesis speech + using the same clip with different models, which are the traditional linear predictive coding + (LPC), the neural homomorphic vocoder (NHV), the differentiable WORLD vocoder + (▽World), and the proposed source-filter GOLF vocoder (GOLF).

- +

p361

@@ -99,135 +108,373 @@

Listening Samples

- + + + + + + + + + + + + + + + + + - + + + + + + + + + +
Clip 1 + p361_004 + + + + + + + +
p361_006 + + + + + +
p361_008 + + +
Clip 2 + p361_011 + + + + + + + +
p361_024 +
+

p360

+ - + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
Clip 3 + Test clipReferenceLPCNHV▽WorldGOLF
p360_003 +
p360_005 + + + + + + + +
p360_007 + + + + + + + + + +
p360_009 + + + + + + + + + +
p360_019 + + + + + + + + + +
+
-
-

Additional Materials

-
- - +

GOLF variants

+
+

Citation


-@inproceedings{ycy2023golf,
-    title={Singing Voice Synthesis Using Differentiable LPC and Glottal-Flow-Inspired Wavetables},
+@inproceedings{ycy2024golf,
+    title={Differentiable Time-Varying Linear Prediction in the Context of End-to-End Analysis-by-Synthesis},
     author={Chin-Yun Yu and Gy{\"o}rgy Fazekas},
-    booktitle={International Society for Music Information Retrieval (ISMIR)},
-    year={2023}
+    booktitle={INTERSPEECH},
+    year={2024}
 }
@@ -254,5 +501,21 @@

Citation

crossorigin="anonymous"> + \ No newline at end of file