Hugging Face's logo

Datasets:
simpra
/
xh-tts-samples

Modalities:
Audio
Text
Formats:
soundfolder
Languages:
Xhosa
Size:
< 1K
Tags:
text-to-speech
vits
isixhosa
Libraries:
Datasets
License:
Dataset card Data Studio Files
xet
 Community
1
Dataset Viewer
Auto-converted to Parquet Duplicate
audio
audio
kind
string
run
string
step
float64
speaker_id
float64
speaker
string
note
string
text
string
REAL HUMAN reference
null
null
null
xho_reader_006
male reader excerpt
(excerpt of a Vuk'uzenzele article, not the probe)
REAL HUMAN reference
null
null
null
xho_reader_008
male reader excerpt
(excerpt of a Vuk'uzenzele article, not the probe)
model output
run 1: ViXSD raw, batch 16 (stopped at 16k)
0
4
xho_reader_005
18 dB SNR (noisy)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 1: ViXSD raw, batch 16 (stopped at 16k)
2,000
4
xho_reader_005
18 dB SNR (noisy)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 1: ViXSD raw, batch 16 (stopped at 16k)
4,000
4
xho_reader_005
18 dB SNR (noisy)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 1: ViXSD raw, batch 16 (stopped at 16k)
6,000
4
xho_reader_005
18 dB SNR (noisy)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 1: ViXSD raw, batch 16 (stopped at 16k)
8,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 1: ViXSD raw, batch 16 (stopped at 16k)
10,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 1: ViXSD raw, batch 16 (stopped at 16k)
12,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 1: ViXSD raw, batch 16 (stopped at 16k)
14,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 1: ViXSD raw, batch 16 (stopped at 16k)
16,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
0
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
1,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
2,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
3,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
4,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
5,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
6,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
7,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
8,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
9,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
10,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
11,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
12,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
13,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
14,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
15,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
16,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
17,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
18,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
19,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
20,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
21,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
22,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
23,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
24,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
25,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
26,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
27,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
28,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
29,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
30,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
31,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
32,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
33,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
34,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
35,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
36,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
37,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
38,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
39,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
40,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
41,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
42,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
43,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
44,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
45,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
46,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?
model output
run 2: ViXSD + SLR32 normalised, batch 32
47,000
7
xho_reader_008
66 dB SNR (cleanest)
Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?

isiXhosa TTS — reference audio and training samples

Two very different kinds of audio live here. Check the folder before judging anything.

folder what it is source
speakers/ REAL HUMAN speech — 20 s excerpts of ViXSD readers, for choosing a voice ViXSD recordings
samples_vixsd/ MODEL OUTPUT — what the VITS model generates at a given training step generated

speakers/ — ground truth

male_xho_reader_008_20s.wav etc. Excerpts taken from the middle of each reader's longest recording, not the start, so you hear steady-state delivery. These are the target the model is trying to imitate — not model output.

samples_vixsd/ — training progress

step_NNNNNNN.wav is the model's rendering of one fixed sentence at that training step:

Molo, ndicofa phi xa ndirhalela iqanda eliqhekekileyo?

It exercises all three isiXhosa click letters — c (ndicofa), q (iqanda, eliqhekekileyo) and x (xa).

Every file currently in this folder is speaker id 4 (xho_reader_005). Files written from now on carry the speaker in the name (step_NNNNNNN_spkN.wav), because id 4 turned out to be the wrong default: it has the most audio (108.9 min) but the worst signal-to-noise in the corpus, about 18 dB against 43-66 dB for the others. It is quiet and noisy, which is why step_0008000.wav sounds distorted.

The default is now id 7 (xho_reader_008) — male, 38.7 min, 66 dB SNR, the cleanest voice available.

Expected progression

step expect
0 noise (random init)
2,000 speech rhythm, no words
~8,000 words forming, clicks audible, still rough
20,000+ clearer, less robotic

Judge quality against speakers/, not against commercial TTS: the model trains on 7.4 h, where VITS's reference corpus LJSpeech has ~24 h.

Related

Downloads last month
534
Number of rows:
59
Total file size:
11.8 MB