# Text-to-Speech on Balalaika corpus (test)

Source: [Text-to-Speech on Balalaika corpus (test)](<https://www.sota2.com/research/sota/text-to-speech-on-balalaika-corpus-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-03T07:14:04.209Z

Comprehensive evaluation of Text-to-Speech synthesis quality on the test split of the Balalaika corpus, covering subjective quality (MOS variants), intelligibility (WER/CER), and acoustic fidelity (NOI, DIS, COL, LOU).

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Text-to-Speech |
| Dataset (including any stated split) | Balalaika corpus (test) |
| Metric | MOS (Mean Opinion Score) |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [MOS (Mean Opinion Score)](<https://www.sota2.com/research/sota/text-to-speech-on-balalaika-corpus-test.md?metric=MOS%20(Mean%20Opinion%20Score)>): Higher is better; unit: SCALAR.

- [IntMOS (Intelligibility MOS)](<https://www.sota2.com/research/sota/text-to-speech-on-balalaika-corpus-test.md?metric=IntMOS%20(Intelligibility%20MOS)>): Higher is better; unit: SCALAR.

- [UTMOS](<https://www.sota2.com/research/sota/text-to-speech-on-balalaika-corpus-test.md?metric=UTMOS>): Higher is better; unit: SCALAR.

- [NTMOS (Naturalness MOS)](<https://www.sota2.com/research/sota/text-to-speech-on-balalaika-corpus-test.md?metric=NTMOS%20(Naturalness%20MOS)>): Higher is better; unit: SCALAR.

- [NOI (Acoustic Fidelity)](<https://www.sota2.com/research/sota/text-to-speech-on-balalaika-corpus-test.md?metric=NOI%20(Acoustic%20Fidelity)>): Higher is better; unit: SCALAR.

- [DIS (Acoustic Fidelity)](<https://www.sota2.com/research/sota/text-to-speech-on-balalaika-corpus-test.md?metric=DIS%20(Acoustic%20Fidelity)>): Higher is better; unit: SCALAR.

- [COL (Acoustic Fidelity)](<https://www.sota2.com/research/sota/text-to-speech-on-balalaika-corpus-test.md?metric=COL%20(Acoustic%20Fidelity)>): Higher is better; unit: SCALAR.

- [LOU (Acoustic Fidelity)](<https://www.sota2.com/research/sota/text-to-speech-on-balalaika-corpus-test.md?metric=LOU%20(Acoustic%20Fidelity)>): Higher is better; unit: SCALAR.

- [WER (Word Error Rate)](<https://www.sota2.com/research/sota/text-to-speech-on-balalaika-corpus-test.md?metric=WER%20(Word%20Error%20Rate)>): Lower is better; unit: PERCENT.

- [CER (Character Error Rate)](<https://www.sota2.com/research/sota/text-to-speech-on-balalaika-corpus-test.md?metric=CER%20(Character%20Error%20Rate)>): Lower is better; unit: PERCENT.

- [SIM (Similarity Score)](<https://www.sota2.com/research/sota/text-to-speech-on-balalaika-corpus-test.md?metric=SIM%20(Similarity%20Score)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | MOS (Mean Opinion Score) | Result date | Report paper |
| --- | --- | --- | --- |
| F5 | 2.938 | Not recorded | 69cf49980636e5bf4afcda63 |
| XTTS | 2.905 | Not recorded | 69cf49980636e5bf4afcda63 |
| DiTTo-TTS | 2.521 | Not recorded | 69cf49980636e5bf4afcda63 |
| Tortoise | 2.51 | Not recorded | 69cf49980636e5bf4afcda63 |

### F5

MOS (Mean Opinion Score): 2.938. Result ID: 69cf684747f34ce924b8c832.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49980636e5bf4afcda63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78a71e209131cea79a4f.md>)

[Implementation](<https://github.com/SWivid/F5-TTS>)

Recorded evaluation configuration: {"model":"F5"}. This record does not assert independent reproduction.

### XTTS

MOS (Mean Opinion Score): 2.905. Result ID: 69cf684747f34ce924b8c834.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49980636e5bf4afcda63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73831e209131cea5cd91.md>)

Recorded evaluation configuration: {"model":"XTTS"}. This record does not assert independent reproduction.

### DiTTo-TTS

MOS (Mean Opinion Score): 2.521. Result ID: 69cf684747f34ce924b8c831.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49980636e5bf4afcda63.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cf49980636e5bf4afcda63.md>)

Recorded evaluation configuration: {"configuration":"1+2 encoder"}. This record does not assert independent reproduction.

### Tortoise

MOS (Mean Opinion Score): 2.51. Result ID: 69cf684747f34ce924b8c833.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49980636e5bf4afcda63.md>)

Recorded evaluation configuration: {"model":"Tortoise"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
