Does Nugent audio's dialogue check give a usefully objective measure of speech intelligibility or is a human listener still the only meter that matters?
ASKED BY · Paul Walhus · AUSTIN, TEXAS · 14 OCT 2026 · ▶ 00:35:28
Data measurements (integrated, short-term, peak loudness, crest factor) help unearth issues; Nugent DialogCheck uses psychoacoustic technology; measurements can flag cleanup, noise reduction, or ADR needs but human review essential.
Human ears remain essential; no single standard applies universally; intelligibility varies by context (commercials vs. audiobooks); human judgment crucial to account for speaker variability and creative intent.
Objective measurement addresses subjective listening test variability; based on Fraunhofer IDMT research on hearing aids; provides standard achievable regardless of playback system or room acoustics.
Based on Fraunhofer listening effort standard; doesn't use it himself; confidence in proper mixing technique; contention arises when creative intent demands intentionally low intelligibility; similar to loudness specs—may need to contest.
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.
35:28usefully objective measure of speech intelligibility or is a human listener still the only meter that matters? Bill Davis. So it's all eventually going to come down to human ears. But I wish I could say that that was a single standard, but it certainly in my experience over the course of my career has not been. Let me just say this. I have worked in my career predominantly in two areas,
35:52and they're at the two far ends of intelligibility as an idea. Commercials, which I did in my early life, intelligibility is everything. You are in the position where you are trying to get a commercial message of 30 seconds out into a world where people had background noise, they were in in cars driving around. They were on little transistor radios in the early days of my career.
36:18So pumping up the clarity parts of a signal so that intelligibility survives all that background junk was everything. And we used to do all sorts of tricks to make things as clear and punchy as possible. Here later in my career, I find myself working in audio books where you can expect somebody working at an assembly line to be listening for 10 hours a day to somebody talk
36:46in their head. And that same profile of intelligibility might exhaust them and make them not want to continue to listen. So it's not that there is an objective intelligibility standard. There are standards, certainly. And Mickey and others can speak to how to make things not muddy or indistinct such that the person listening has to struggle to hear what someone's saying you also have people that have unusual speech patterns
37:16and sometimes extra boosts in certain frequencies can help that person's voice cut through and be distinct but what I'm describing is it's never as simple as if If you see this on anyone's meter and it says it's good, that means it's always good for everything. That takes out the human variability that is always going to be a part of this. And we just have to deal with that if you're dealing with audio.
37:43At the end of things, you really do need somebody who understands what they're listening for and can judge this file at this time to make sure its quality is the best it can be so that you reach the most people with the message you're trying to get out. I'm sorry, I forgot who's next. Alexander, is it? It's me. All right, so I'm going to limit what I say here because I don't work in the post industry.
38:07Although I have some very little experience mixing in terms of, like, you know, I've done some, like, audio book style stuff, radio dramas, that type of stuff where there's a lot of sound beds and effects and you've got dialogue and stuff like that. So I've used other Nugent tools, not specifically DialogCheck, but I agree with Bill, still going to come down to a human being.
38:31But I think we do have the data now that at least can show you because we can now measure, you know, we've previously on the show just this morning, we talked about, you know, program integrated loudness. There are other aspects, too, of this, like short term loudness, short term max, momentary, momentary max. there's also PLR peak to loudness which is the measurement the difference between the peak
38:54the peak and the integrated loudness I think it's sometimes referred to as crest factor too but there's there's all sorts of other data so this is important data to have as a dialogue mixer because it can unearth other information about what what is going on in your mix and even Nugent says now this particular dialogue check application it has some other technology to do with psychoacoustics
39:17I haven't really done a deep dive on that too, but you know, it may unearth some things about like Problems with the dialogue to like not just the mix but like does your dialogue need more cleanup? Do you need more noise reduction? Does it do you need ADR? Do you actually need to record it was it properly recorded in the first place and like Bill said to there are also issues too