Supertone Clear and Xcentize DX Revive claim to rebuild speech, not just denoise. Is regenerated dialogue ethical for a documentary interview?
ASKED BY · Paul Vahus · AUSTIN, TEXAS · 18 SEP 2026 · ▶ 00:11:24
DaVinci Resolve has a similar feature that works fairly well, but ADR (automatic dialogue replacement) is the traditional and preferred approach for fixing dialogue issues.
The ethics and feasibility of AI speech manipulation is evolving rapidly; it's increasingly difficult to distinguish real from generated speech, raising significant societal concerns about authenticity.
These tools use neural networks to separate speech from noise rather than regenerate voice; the ethical concern is less about changing someone's voice and more about taste and whether over-processing sounds artificial.
Summarised by AI from what the panel said. Watch the moment for the full answer.
From the transcript.
Machine-transcribed, so it may contain errors. Every timestamp opens the episode at that second.
11:11Not just denoise is regenerated dialogue ethical for a documentary interview. Alex Knight. Well, I can't make a statement that says that it is or it isn't because, you know, I can only comment what I would or would not do. DaVinci Resolve has had this feature for a while now, and it's surprisingly good. It's not a while. It's a fairly new feature in DaVinci Resolve Studio.
11:39I think you've got to pay for it. But it seems to work decently well, and this technology is evolving rapidly, and it keeps getting better and better. I was it's able to take audio from something that I've done and just relatively convincingly recreate that person's voice. I'm struggling to think of a scenario where I would need this. If there's bad dialogue, that's what ADR is. You get the actor or get someone to come back in and
12:06redo it. If it's a voiceover, redo the voiceover. If there's a problem, you know, if it's location sound based, you would just redo it. There are tools inside the software programs that we use to match the ambience. There are techniques you can use to, you know, when you do ADR so that it doesn't sound like it was just done in some ISO booth and it doesn't match the environment that
12:28the actor was actually in on set. So would I use it? No. Is it ethical? That's for other people to decide. Bill? I'm going to talk about the ethical part of this because this has been a problem. Well, not just the problem. It's kind of been a universal worldwide difficulty since generative speech became so easy to do and the chips became so sophisticated. It is now very hard to tell the
12:50difference between something that was generative, at least in short spurts. Usually if I'm listening to something for five minutes, I'll find a telltale clue that what I'm listening to was never touched a human vocal cord. But so much of it is so easy to produce. And the economics of using generative speech in all of its forms, whether it's a benign little thing like,
13:16this should have been plural, and it was singular, and the actor misspoke, and I just want to stick an S on the end of that so I don't have to reshoot it. Almost everybody would say, yeah, that's probably fine. But we're now getting these things where we're starting to creep into the world of forced emotion in generative speech and things like that. So I think the ethics
13:39discussion needs to continue to move as fast as possible because, boy, it's getting hard when you're just watching content to figure out whether something is real or 100% manufactured for a purpose of fooling you. And that's not good for society, in my opinion. Yeah. And the thoughts about, or the debate, rather, on AI versus human-generated speech aside, it should note that neither of those plugins are actually doing AI speech. If I look
14:13in the chat, we've got a lot of comments that are, Mickey and others are mentioning that these don't actually use AI to generate the speech. So the question about these plugins is a little separate from the actual debate on how do we feel about the use of it. Courtney? Yeah, I don't really like the sound that it generates. It is not generative AI.
14:40It's basically, it separates the actual speech that was recorded from the background sounds, kind of like the filter that I'm using right now on Zoom, So they have a version of this called personalized audio. And I use it for removing the sound of the garbage trucks outside of the phone ringing or the air conditioner behind me. And it manages to find my voice and send it through in real time.
15:06This Supertone also does the same thing. They do have a real-time version of it. It can either process or do it real time. So it's doing a similar thing. I think they use something like this. If you watch the Apple keynote for the iPhone Duo intro, all of those people sounded exactly the same. Regardless of their environment, whether they were in that curved glass parabolic mirror hallway or on a green screen stage, we don't know, or outside, they all sounded close mic'd with an open condenser microphone.
Similar questions
from other episodes.
About the same products and topics, most related first.