Compare the values directly.
- Recording length
- Mean loudness
Benchmarks
We ran ConductSpeech on 290 public recordings and compared its transcripts with the words each person read, and its pitch with Praat's.
Pitch across 6 seconds of one recording. The gray dots are pitch the old version reported where Praat detects no voicing.
Pitch
Praat is the free program voice researchers use to measure pitch. Before version 2.31.0, ConductSpeech also measured pitch in moments Praat doesn't count as voice, so pitch variation and range came out 34% to 55% higher than Praat's. We fixed that and checked the fix on 205 recordings we hadn't used to build it.
Children91 recordings
Second-language women72 recordings
Second-language men42 recordings
Children91 recordings
Second-language women72 recordings
Second-language men42 recordings
| Measure | Before 2.31.0 | Since 2.31.0 | Praat |
|---|---|---|---|
| Children, 91 recordings | |||
| Mean pitch (Hz) | 232.6 | 226.8 | 224.5 |
| Variation (semitones) | 5.02 | 3.48 | 3.36 |
| Range (semitones) | 14.31 | 8.79 | 9.21 |
| Second-language women, 72 recordings | |||
| Mean pitch (Hz) | 230.1 | 226.5 | 223.2 |
| Variation (semitones) | 4.91 | 3.68 | 3.67 |
| Range (semitones) | 14.08 | 9.47 | 9.71 |
| Second-language men, 42 recordings | |||
| Mean pitch (Hz) | 156.3 | 147.1 | 143.2 |
| Variation (semitones) | 5.68 | 4.21 | 3.91 |
| Range (semitones) | 15.30 | 10.64 | 9.86 |
Median pitch since 2.31.0, gap to Praat: Children 0.14%, Second-language women 0.3%, Second-language men 0.57%.
Samples analyzed before Nov 6, 2026 may still show the older, higher pitch numbers. Upload the recording again to update them.
205 recordings in 3 groups · Nov 5, 2026 · version 2.31.0 · group means · one person reading aloud in each recording
Transcripts
We compared each transcript with the text the speaker read and counted the missed, added and wrong words. The second-language speakers are children and adults whose first language is Mandarin.
Word error rate by group
Native adults
4.9%
Median recording 4.1%
Children
22.2%
Median recording 19.7%
Second-language women
26.0%
Median recording 14.6%
Second-language men
26.1%
Median recording 13.2%
The editor plays the recording one line at a time and underlines the words the transcriber wasn't sure of.
See the transcript editor290 recordings · Nov 6, 2026 · version 2.31.1 · one person reading aloud in each; conversation is harder to transcribe
Voice measures
Bridge2AI-Voice publishes the code behind its voice measures (senselab, run with b2aiprep's settings). We ran that code and ConductSpeech on the same 290 recordings and compared each measure.
290 public recordings, none from Bridge2AI · Nov 5, 2026 · version 2.31.0
Limits
Every child here was a Mandarin speaker reading English.
Speaker labels from voices stay off until we check them against clinicians' corrections.
MLU and the other language measures are checked against software counts on 20 samples, and against counts that follow SALT's written conventions on made-up samples. Checks against clinicians' hand scoring of real samples come next.
290 public recordings of people reading aloud, free to use under CC BY 4.0. The children and adults whose first language is Mandarin come from SpeechOcean762 (Zhang and colleagues, 2021; OpenSLR 101). The adults reading English audiobooks come from LibriSpeech dev-clean (Panayotov and colleagues, 2015; OpenSLR 12).
By the correlation (r) between our numbers and Bridge2AI's, in the group where they agree least:
| Test | Results | Date | Version |
|---|---|---|---|
| Pitch | 1.1.0 | Nov 5, 2026 | 2.31.0 |
| Voice measures | 1.1.0 | Nov 5, 2026 | 2.31.0 |
| Transcripts | 1.0.0 | Nov 6, 2026 | 2.31.1 |
Page data 1.2.0.
ConductScience ran these tests. ConductSpeech has not been independently validated in a peer-reviewed study.
Check the transcript against what you hear, then read the measures.