Media2Text

Experiment run August 7, 2026 · Published September 2, 2026 · Media2Text research

2026 ASR benchmark: text, timing and speaker trade-offs

Short answer: no model won every category on this sample. The strongest observed text result did not include speaker labels, while integrated speaker-capable systems made different segmentation or transcription errors. Native word timing also varied by engine, so timing and diarization must be evaluated separately from text quality.

Important: this was a screening experiment on one 15.093-second Chinese animated-dialogue clip. It is not a universal accuracy ranking, and this page does not publish a formal word error rate, character error rate or diarization error rate.

Question and scope

The experiment asked a practical product question: when the same short recording is sent to different speech-recognition systems, how do the returned text, processing time, word or character timing and automatic speaker output differ?

The clip contains rapid character and voice changes. Visual review suggested four characters and at least three plausible speaking voices. There was no independently human-labelled speaker timeline, so speaker behavior was assessed qualitatively by turn consistency rather than by a formal diarization score.

Method

  1. Use one unchanged source. Every result in the primary table used the same 15.093-second source clip.
  2. Record wall time. Cloud figures include upload, network, scheduling and polling. Local figures identify warm inference where the model was already loaded.
  3. Review text against a consensus transcript. The consensus was assembled through repeated listening, but it was not a separately commissioned professional ground truth.
  4. Inspect timing and speaker output independently. A model could return strong text but no speaker labels, or return detailed timing with weaker text.
  5. Keep caveats with each result. A selected observation applies to this clip and configuration, not to every recording in the model's supported languages.

Selected same-sample observations

Model or pipelineObserved processing timeText observation on this clipTiming returnedSpeaker output
Qwen3-ASR-Flash-Filetrans4.751 sNo material word error found in the qualitative review; punctuation was clean.Native word timingNo
Fun-ASR-MTL5.739 sClose to the consensus, with some turn and punctuation merging.Native word timingAutomatic labels, but not fully reliable on this short clip
Paraformer-v24.827 sCore words were mostly correct, with a visibly incorrect sentence split.Word timingAutomatic labels returned
iFlytek LFASRAbout 5 sNear-consensus text in this run.Word and sentence timingAutomatic mode returned only one speaker
Speechmatics Standard6.351 sMultiple material Chinese transcription errors were observed.Word timingAutomatic mode returned one speaker
Whisper large-v3 INT8, local CPU8.184 s warmOne clear lexical difference was observed in the reviewed transcript.Word timingNo integrated speaker labels
Qwen3-ASR 0.6B INT8 ONNX, local CPU1.602 s warmVery clean text on this short sample.No native workspace timing in this runNo
SenseVoiceSmall INT8 ONNX, local CPU0.200 s warmOne material lexical difference was observed.Character timing pointsNo integrated speaker labels

Times are observations from the recorded environment, not vendor service-level commitments. Network, queue load, audio length and hardware can change end-to-end latency.

What the experiment shows

Text quality and workspace completeness are different questions

The cleanest text result on this sample did not include integrated speaker labels. Systems that returned text, timing and speakers in one response were more feature-complete, but some merged turns or returned an implausible speaker count. Selecting a production route therefore requires separate acceptance thresholds for text, timing and speakers.

Word timing is not a universal capability

Some engines returned native word or character timing; another strong local text result did not. A transcript product should not manufacture precise word timing when the engine only returns segment timing. Media2Text uses word-level timing when reliable data is returned and otherwise keeps playback and subtitle generation at segment level.

Automatic speaker count is an estimate

A separate local diarization screening on the same clip produced two detected speakers at one clustering threshold and three at another. That sensitivity is why Media2Text exposes generic, editable speaker labels instead of presenting automatic labels as verified identities.

Fast local inference is not the same as a complete service

The fastest local runs were model-ready inference measurements. They did not include all of the work needed for a production service, such as file transfer, queuing, timing alignment, speaker processing, retries and durable result storage.

Limitations

What a stronger follow-up requires

A production-grade benchmark should use rights-cleared, human-labelled recordings across interviews, meetings, overlapping speech, noisy mobile audio, multiple accents and every launch language. It should report CER or WER, diarization error rate, speaker-count error, timestamp drift, wall time, queue time and failure rate. Results should be versioned with model identifiers and the test date.

Primary model references

For the product behavior built around these findings, see Audio to Text, Video to Text and the security and data-flow page.