Verified purchaser
Good Idea, but horrible output (Dubbing).
[update: Edited the tacos from 1 to 3, as Terry did an excellent job in answering in a nice and understandable way. I hope to update the review another and final time after the promised update in two weeks]
I bought this tool hoping it would fulfill my exact use case: translating live webinars into other languages. Unfortunately, the current output is not suitable for professional environments.Here are the main issues:
Voice Cloning Quality:
The cloned voice sounds nothing like me. While the translation from German to English is technically correct, the accent and pronunciation sound highly unnatural and completely artificial. For professional webinars, this quality is unfortunately unacceptable.
Speed Adjustment & Pacing:
The system forces the translated audio to strictly fit the original timeline, leading to terrible pacing. The speed fluctuates constantly—switching from normal speed to an extremely accelerated, distorted rush, followed by abrupt 2–3 second silences before the next section starts. Competitors like HeyGen handle this much better by dynamically stretching the timeline to keep the voice sounding natural.I really wanted to like this tool, but in its current state, I can only give it 1 star. I am happy to update my review as soon as these technical issues are resolved.
Terry_SubEasy.ai
Edited Sep 1, 2026Hi Marco,
Thanks for laying it out this precisely — both points are actionable, so let me
answer them separately.
Voice cloning: you're right, and it's the part we're not happy with either. The
current clone doesn't hold up for professional use, and German-to-English is
exactly where the accent and pronunciation break down. One other German user has
mentioned something similar, so that language pair is on our list. An upgraded
cloning model is already in development and ships within two weeks. It'll be
announced in the SubEasy product updates, so you'll know when to re-run the same
webinar.
Pacing: this one comes down to a setting, and our default is what tripped you up.
In the translation step, under Advanced options, there's "Optimize for" with two
choices: Subtitles or AI Dubbing. Subtitles keeps each line short enough to fit
the original timeline, which is exactly what produced the rushed sections and the
2-3 second silences once it was voiced. Switch it to AI Dubbing and the
translation is written for speech instead, so the pacing stays natural.
One thing worth knowing: this is set at translation time, not at dubbing time. So
you'll need to re-run the translation with AI Dubbing selected, then generate the
voice again. Re-dubbing the existing translation on its own won't change the
pacing.
And thank you for surfacing that second point — when a translation optimised for
subtitles is sent to dubbing, we should be telling you about it right there
instead of letting you find it in the output. That prompt goes into the next
version.
For anything else, write to [email protected]. We have a dedicated support team
on it, and you'll get a reply within 24 hours on business days.