# Style-controlled Speech Synthesis on Seed-TTS-eval en (test) Source: [Style-controlled Speech Synthesis on Seed-TTS-eval en (test)]() Collection index: [llms.txt]() Data updated: 2026-06-05T09:13:16.992Z Evaluation of Style-controlled Speech Synthesis performance on the English test split of the Seed-TTS-eval dataset, assessed using a comprehensive set of metrics including SPS, speaker similarity (SpkSim), and various MOS scores. 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 | Style-controlled Speech Synthesis | | Dataset (including any stated split) | Seed-TTS-eval en (test) | | Metric | SPS (Steps Per Second) | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 9 of 9 | | 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 - [SPS (Steps Per Second)](): Higher is better; unit: SCALAR. - [F0 Similarity (Mean)](): Higher is better; unit: SCALAR. - [F0 Similarity (F)](): Higher is better; unit: SCALAR. - [WER (Word Error Rate)](): Lower is better; unit: PERCENT. - [SpkSim (Speaker Similarity)](): Higher is better; unit: SCALAR. - [UTMOS (Utterance MOS)](): Higher is better; unit: SCALAR. - [S-MOS (Style MOS)](): Higher is better; unit: SCALAR. - [N-MOS (Naturalness MOS)](): Higher is better; unit: SCALAR. ## Recorded results | Method | SPS (Steps Per Second) | Result date | Report paper | | --- | --- | --- | --- | | Fast LoRA | 5.59 | Not recorded | 6a22586e86173ad9c1468964 | | DSP speed-up | 5.48 | Not recorded | 6a22586e86173ad9c1468964 | | Low-pitch LoRA | 3.68 | Not recorded | 6a22586e86173ad9c1468964 | | Baseline (CosyVoice2-0.5B) | 3.65 | Not recorded | 6a22586e86173ad9c1468964 | | DSP pitch-up | 3.65 | Not recorded | 6a22586e86173ad9c1468964 | | DSP pitch-down | 3.65 | Not recorded | 6a22586e86173ad9c1468964 | | High-pitch LoRA | 3.61 | Not recorded | 6a22586e86173ad9c1468964 | | Slow LoRA | 2.3 | Not recorded | 6a22586e86173ad9c1468964 | | DSP slow-down | 2.19 | Not recorded | 6a22586e86173ad9c1468964 | ### Fast LoRA SPS (Steps Per Second): 5.59. Result ID: 6a22932bfbc373592c20ee5d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Style Control Axis":"Speed-axis control","Operation":"speed-up","Zero-shot transfer":"true"}. This record does not assert independent reproduction. ### DSP speed-up SPS (Steps Per Second): 5.48. Result ID: 6a22932bfbc373592c20ee5c. [Reporting paper]() Recorded evaluation configuration: {"Style Control Axis":"Speed-axis control","Operation":"speed-up","Zero-shot transfer":"true"}. This record does not assert independent reproduction. ### Low-pitch LoRA SPS (Steps Per Second): 3.68. Result ID: 6a22932bfbc373592c20ee63. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Style Control Axis":"Pitch-axis control","Operation":"pitch-down","Zero-shot transfer":"true"}. This record does not assert independent reproduction. ### Baseline (CosyVoice2-0.5B) SPS (Steps Per Second): 3.65. Result ID: 6a22932bfbc373592c20ee5b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Style Control Axis":"Baseline","Zero-shot transfer":"true"}. This record does not assert independent reproduction. ### DSP pitch-up SPS (Steps Per Second): 3.65. Result ID: 6a22932bfbc373592c20ee60. [Reporting paper]() Recorded evaluation configuration: {"Style Control Axis":"Pitch-axis control","Operation":"pitch-up","Zero-shot transfer":"true"}. This record does not assert independent reproduction. ### DSP pitch-down SPS (Steps Per Second): 3.65. Result ID: 6a22932bfbc373592c20ee62. [Reporting paper]() Recorded evaluation configuration: {"Style Control Axis":"Pitch-axis control","Operation":"pitch-down","Zero-shot transfer":"true"}. This record does not assert independent reproduction. ### High-pitch LoRA SPS (Steps Per Second): 3.61. Result ID: 6a22932bfbc373592c20ee61. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Style Control Axis":"Pitch-axis control","Operation":"pitch-up","Zero-shot transfer":"true"}. This record does not assert independent reproduction. ### Slow LoRA SPS (Steps Per Second): 2.3. Result ID: 6a22932bfbc373592c20ee5f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Style Control Axis":"Speed-axis control","Operation":"slow-down","Zero-shot transfer":"true"}. This record does not assert independent reproduction. ### DSP slow-down SPS (Steps Per Second): 2.19. Result ID: 6a22932bfbc373592c20ee5e. [Reporting paper]() Recorded evaluation configuration: {"Style Control Axis":"Speed-axis control","Operation":"slow-down","Zero-shot transfer":"true"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.