Auphonic alternative

Auphonic alternative with fewer word errors

Cleanvoice Studio Sound 4.1 is an AI audio enhancer you can use instead of Auphonic Studio Voice, Auphonic's speech enhancement model. On 100 noisy clips from Dolby's public UNIVERSE set, it made fewer word errors (15.5% vs 20.3%), scored higher on speech quality (SCOREQ 4.54 vs 4.16) and restored more bandwidth (17.3 vs 16.5 kHz). It also dropped fewer words: 3.0 per 100, against 6.2.

Last updated: 2 November 2026 · 100 clips from Dolby’s public UNIVERSE set · Full benchmark

Try Cleanvoice Studio Sound free See every metric

Word error rate (lower is better)
15.5%
Cleanvoice, vs 20.3% for Auphonic
Dropped words (lower is better)
3.0%
Cleanvoice, vs 6.2% for Auphonic
Speech quality (SCOREQ) (higher is better)
4.54
Cleanvoice, vs 4.16 for Auphonic

Head to head on every metric

Auphonic Studio Voice against Cleanvoice Studio Sound 4.1, with the unprocessed input for context. Same 100 clips, run on 1 November 2026: Cleanvoice processed each clip separately; the other tools processed the 100 clips joined into one file. Method and all eight tools.

Cleanvoice Studio Sound 4.1 vs Auphonic Studio Voice on 100 noisy clips from Dolby’s UNIVERSE validation set, 1 November 2026
Metric	Unprocessed input	Auphonic	Cleanvoice Studio Sound 4.1	Better
Word error rate ↓Omnilingual ASR on the output vs on the clean original	17.0%	20.3%	15.5%	Cleanvoice
Dropped words ↓Deletions per 100 reference words	5.9%	6.2%	3.0%	Cleanvoice
Changed words ↓Substitutions per 100 reference words	10.8%	13.8%	11.2%	Cleanvoice
Speech quality (SCOREQ) ↑Natural speech quality, 1 to 5	2.64	4.16	4.54	Cleanvoice
Overall quality (DNSMOS OVRL) ↑DNSMOS P.835 overall score, 1 to 5	2.27	3.21	3.35	Cleanvoice
Speaker similarity ↑How close the voice stays to the clean original	0.916	0.949	0.955	Cleanvoice
Intelligibility (ESTOI) ↑Against the clean reference	0.714	0.793	0.840	Cleanvoice
PESQ ↑Against the clean reference	1.55	1.60	1.88	Cleanvoice
Bandwidth ↑Median output bandwidth; the input stops at 8 kHz	8.0 kHz	16.5 kHz	17.3 kHz	Cleanvoice

Cleanvoice has the better number on every metric. The biggest gaps are word errors, dropped words and speech quality.

↓ lower is better, ↑ higher is better. Dropped and changed words are per 100 reference words. Word error rate also counts words the transcriber added, so those two don’t add up to it.

Clip by clip

Of the 100 clips, Cleanvoice had fewer word errors than Auphonic on 41, the same number on 45 and more on 14.

41Cleanvoice betterfewer word errors
45Sameequal word errors
14Auphonic betterCleanvoice had more word errors

Listen: clip 0014

A damaged clip: echo, a band-reject filter and AC-3 codec damage. The speaker says: “move for the next couple of days it is not too late to get on board some have accepted it as a”. Listen for the end of the sentence, “some have accepted it as a”.

Clip 0014

Selected: Original (unprocessed)

0:00

Spectrograms of UNIVERSE clip 0014 (echo, a band-reject filter and AC-3 codec damage) through each tool, with what Omnilingual ASR heard. Cleanvoice: 5% of words wrong, keeps the end of the sentence, "some have accepted it as a". Unprocessed input: 14% wrong, the end comes out as "asam". Adobe Podcast: 27% wrong, stops at "some". Auphonic: 32% wrong. ElevenLabs and Descript Studio Sound: 36% wrong. Dolby UNIVERSE: 55% wrong. Each of them drops or garbles the end of the sentence. — Clip 0014 through six of the tools. The dashed line is 8 kHz, where the 16 kHz input ends; anything above it was restored by the tool. Transcripts by Meta’s Omnilingual ASR (CTC 7B v2).

Auphonic alternatives, ranked

Every tool in our benchmark, fewest word errors first. Same 100 clips, run on 1 November 2026. Cleanvoice and DeepFilterNet3 processed each clip on its own; the other products processed the 100 clips joined into one file. See all metrics.

Speech enhancement tools ranked by word error rate on 100 noisy clips from Dolby’s UNIVERSE set, 1 November 2026
#	Tool	Word error rate ↓	Dropped words ↓	Speech quality (SCOREQ) ↑	Bandwidth ↑	Note
1	Cleanvoice Studio Sound 4.1	15.5%	3.0%	4.54	17.3 kHz	Fewest word errors and dropped words; best SCOREQ, DNSMOS and speaker similarity; 17.3 kHz bandwidth, tied with Descript for the widest.
2	Cleanvoice Studio Sound 4.0	16.7%	3.2%	4.51	14.0 kHz	Our previous model. 8% more word errors than 4.1, and 14.0 kHz bandwidth.
3	Adobe Podcast Enhance Speech	17.1%	6.8%	3.57	9.5 kHz	Fewest word errors of the other companies (17.1%), close to the unprocessed input (17.0%). Fewest changed words and best PESQ of the products; 9.5 kHz bandwidth.
4	Auphonic Studio Voice	20.3%	6.2%	4.16	16.5 kHz	Best speech quality of the other products; 16.5 kHz bandwidth.
5	Dolby UNIVERSE (research outputs)	22.3%	5.1%	4.33	8.0 kHz	Research outputs from the paper, not a product you can use. Best ESTOI and PESQ.
6	DeepFilterNet3	22.6%	9.6%	3.18	7.5 kHz	Open-source noise remover. Most dropped words (9.6 per 100) and lowest speech quality; no added bandwidth.
7	ElevenLabs Voice Isolator	23.8%	7.1%	3.72	11.8 kHz	Lowest speaker similarity; 7.1 dropped words per 100.
8	Descript Studio Sound	25.4%	8.1%	3.99	17.3 kHz	Most word errors and changed words; 17.3 kHz bandwidth, tied with Cleanvoice for the widest.
–	Unprocessed input	17.0%	5.9%	2.64	8.0 kHz	The noisy file every tool received, for context.

The unprocessed input scored 17.0%. Cleanvoice (15.5%) was the only tool clearly below it; Adobe Podcast (17.1%) was about level with it, and the other four tools made more word errors than the input.

Which one to choose

When to choose Auphonic

You want one tool for the whole post-production chain. Auphonic is a full post-production tool, with levelling and loudness targets as well as noise reduction. This page only compares Studio Voice with Studio Sound.

When to choose Cleanvoice

You want higher quality scores. Cleanvoice scored higher on speech quality (SCOREQ 4.54 vs 4.16) and overall quality (DNSMOS 3.35 vs 3.21), and restored 17.3 kHz of bandwidth against 16.5 kHz.
You want fewer dropped words. Cleanvoice dropped 3.0 words per 100, against 6.2 for Auphonic and 5.9 in the unprocessed input.
You want enhancement that only works on the sound. Cleanvoice Studio Sound removes background noise, echo and other voices and restores bandwidth. It has no step that cuts or rewrites words.
You may also want editing later. Filler-word, stutter and silence removal are separate, optional Cleanvoice edits. You choose whether to use them, and Studio Sound works without them.
You want the same model in the web app and the API, so a manual workflow can become an automated one without changing the sound.

More on what it fixes: Cleanvoice Studio Sound, the AI audio enhancer.

How we tested

Data: the 100 noisy clips of Dolby’s public UNIVERSE validation set (Serrà et al., CC BY 4.0; noises include DEMAND, CC BY-SA).
Processing: each tool’s own product or API on 1 November 2026, at default enhancement settings. Cleanvoice processed each clip separately; the other tools processed the 100 clips joined into one 8.4-minute, 16 kHz file, which we cut back into clips. Dolby UNIVERSE is the paper’s released research outputs.
Word metrics: Meta’s Omnilingual ASR (CTC 7B v2) transcripts of each output against the same model on the clean originals (1,423 words). CTC models generally don’t hallucinate the way models with a language model do. Word error rate also counts words the transcriber added, so dropped and changed words don’t add up to it.
Quality: SCOREQ and DNSMOS P.835 OVRL. Speaker similarity, ESTOI and PESQ are measured against the clean original.
Limits: one dataset, 100 clips, 16 kHz input only, an automatic transcriber, and a test run by Cleanvoice itself.

The full method, every tool’s numbers and more audio are on the speech enhancement benchmark page.

Auphonic comparison questions

Still have a question?

Send it over and a real person will answer.

support@cleanvoice.ai

Ask Now

For speech enhancement, Cleanvoice Studio Sound 4.1 scored best in our 1 November 2026 test on 100 noisy clips from Dolby's UNIVERSE set. It made fewer word errors than Auphonic Studio Voice (15.5% vs 20.3%) and was ahead on SCOREQ (4.54 vs 4.16), DNSMOS (3.35 vs 3.21), bandwidth (17.3 vs 16.5 kHz) and dropped words (3.0 vs 6.2 per 100). Among tools from other companies, Adobe Podcast had the fewest word errors (17.1%), then Auphonic.

In our test, yes. Cleanvoice made 15.5% word errors and Auphonic 20.3%, and Cleanvoice dropped fewer words (3.0 vs 6.2 per 100) and changed fewer (11.2 vs 13.8). Clip by clip, Cleanvoice had fewer errors on 41 clips, the same on 45 and more on 14.

On the quality metrics in our test, yes: SCOREQ 4.54 vs 4.16, DNSMOS OVRL 3.35 vs 3.21, and a median 17.3 kHz of bandwidth against 16.5 kHz from the same 8 kHz input. Listen to clip 0014 on this page and judge for yourself.

It has no step that cuts or rewrites words: it removes noise, echo and background voices and restores bandwidth. It can still get a word wrong on badly damaged audio, as every tool in this test did. The transcriber found 15.5 word errors per 100 on Cleanvoice's output, against 17.0 on the unprocessed input. Filler-word, stutter and silence removal are separate Cleanvoice edits you turn on yourself.

We ran the 100 noisy, 16 kHz clips of Dolby's public UNIVERSE validation set through each tool on 1 November 2026, at default settings. Cleanvoice and DeepFilterNet3 processed each clip on its own. The other products processed the 100 clips joined into one 8.4-minute file, which we cut back into clips. Dolby's row uses the outputs published with its paper. We transcribed every output with Meta's Omnilingual ASR (CTC 7B v2), a CTC model, and compared it with the same model on the clean originals (1,423 words). CTC models generally don't hallucinate the way models with a language model do. Word error rate also counts words the transcriber added, so dropped and changed words don't add up to it. SCOREQ and DNSMOS rate the sound without a reference; ESTOI and PESQ compare it with the clean original.

Cleanvoice AI Research ran it. It covers one dataset of 100 clips, mostly English VCTK speech, and 16 kHz input only. Tools ran on 1 November 2026 at default settings, and other audio or settings could give other results. The method, every number and the audio are on the speech enhancement benchmark page.

Compare

More comparisons and the full benchmark

Adobe, Auphonic, Descript, ElevenLabs and Dolby are trademarks of their owners. Tests run by Cleanvoice on 1 November 2026; results may differ on other material.