Why You Can't Hear the Dialogue — And What Actually Fixes It

"I can't hear what they're saying" is probably the single most common complaint in domestic audio, and it is also the one most likely to be answered badly. The usual responses — turn it up, buy a soundbar, switch on subtitles — treat the symptom. Two of them don't even do that reliably, because the thing that makes speech hard to follow is not how loud the television is. It is how loud the speech is relative to everything else in the mix. Turn the volume up and you raise both together. The ratio you were struggling with is exactly where it was, only now the explosions are painful too.

That ratio — the loudness difference between dialogue and background — is the lever with actual research behind it, and it is the frame worth holding onto for the rest of this piece.

What the research actually measured

A 2020 bachelor's thesis from Ansbach University of Applied Sciences, supervised in part by the Fraunhofer Institute for Integrated Circuits IIS, ran an Adjustment/Satisfaction Test with eleven participants over 65. Rather than asking whether they could hear the dialogue, it let them set their own preferred loudness difference between speech and background. The test material was the German crime series Tatort, picked precisely because it attracts speech-intelligibility complaints and is full of scenes where dialogue competes with music and noise.

Three findings matter here. First, the preferred loudness differences participants chose were larger than those in the original broadcast mix — the mix as delivered was not what these listeners wanted. Second, being able to customise the dialogue level increased their satisfaction. Third, and most usefully for anyone setting up a system at home, preferred levels varied enormously: between participants, within the same participant on repeat trials, and from one programme to the next. There is no single correct number.

The same thesis's literature review gathers earlier work pointing the same way. A BBC study from 1991, run with organisations for the hard of hearing, tested roughly 336 participants of average age just over 60 against versions with background loudness at −6 dB, 0 dB and +6 dB. Torcoli et al. in 2017 — the source of the test method — worked with younger listeners, median age around 25 and without significant hearing loss, and found preferred loudness differences spanning 0 to 13 LU, with satisfaction rising significantly when listeners could adjust the mix themselves.

An Oldenburg sentence test (OLSA) is the result to sit with. Ten listeners with age-related hearing loss (mean age 73.3) were compared against ten normal-hearing listeners (mean age 23.3). Attenuating the background by 6 dB and by 12 dB significantly increased speech intelligibility for both groups. The benefit is not confined to people with hearing loss, and — note the direction — it came from lowering the background, not raising the speech. A MUSHRA test with ten experienced participants at speech amplification levels of +6 dB and +12 dB returned results in the "Good" band.

Then there is the BBC's Casualty "Accessible and Enhanced" trial, which gave viewers a single slider running between the full mix and one with narratively important audio enhanced. Of 299 survey respondents, 47% had some form of hearing loss, and 73% said they could follow the content more easily or enjoyed it more because of the personalisation.

You are not imagining the effort

A separate multimodal study is worth knowing about because it answers the sceptic in the room. Twenty-eight normal-hearing participants were tested with listening effort measured three ways: pupil dilation (mean and peak), self-reported effort, and word recall. The trend was consistent — listening effort decreased as the background level decreased. Dialogue enhancement, including when it was achieved by source separation rather than access to separate speech and background stems, significantly reduced both pupil size and self-reported effort.

In other words, a mix with a poor speech-to-background ratio costs you measurable cognitive work even if your hearing is fine. The exhaustion you feel after an hour of straining is real and instrumented. It is not only about your ears.

Separating the four causes

The reason most advice on this subject misfires is that it collapses four genuinely different problems into one.

• The mix. Wide dynamic range, dialogue distributed through an object-based bed, and mastering decisions made for a calibrated room rather than a living room. You cannot fix this at home; you can only adjust the ratio you're handed.
• The reproduction chain. A centre channel that is mis-placed, mis-levelled, firing into the back of a cabinet, or simply absent; multichannel content downmixed to two channels.
• The room. Reflections and reverberation smearing consonants, hard surfaces, a seating position that hears more room than speaker.
• The listener. Age-related high-frequency loss is extremely common and changes what "unclear" even means. It is also the cause most likely to make every room-side fix feel like a disappointment.

The free fixes, in order

Work through these before spending anything. Each one is an adjustment to the speech-to-background ratio or to what's standing in its way.

1. Find the centre-channel level. In a system with a dedicated centre, dialogue lives there. Its level relative to left, right and the surrounds is the ratio all the research above is about. Raise it a few decibels and listen across several programmes, not one scene.
2. Lower the background instead. Per the OLSA result, attenuating the surrounds and effects is a documented lever in its own right. If your processor lets you trim surround levels independently, try that direction too — it helped normal-hearing listeners as well as hearing-impaired ones.
3. Check where the centre speaker actually points. A centre pushed deep into a cabinet, angled at the floor, or blocked by a soundbar is throwing away intelligibility before the ratio matters.
4. Check what the source is sending. If a 5.1 or Atmos stream is being downmixed to two channels somewhere in the chain, the discrete dialogue channel you were counting on has already been folded into everything else.
5. Try the dialogue-enhancement, speech-boost or clarity control. These are the consumer-facing implementation of exactly what the studies tested — they change the speech-to-background ratio. They appear on televisions, soundbars and increasingly in streaming apps. How much of any given implementation is genuine separation-based processing versus level and EQ varies by brand, and is worth treating with mild scepticism, but the control is pointed at the right variable.
6. Turn off everything else. Virtual surround, bass enhancement, "clear voice" stacked on top of a room correction curve — processing layered on processing makes the ratio harder to reason about, not easier.

Two framing points the research insists on. Preferred settings vary widely between people and between programmes, so treat this as finding your setting rather than hunting a correct number — and expect to change it when the programme changes. And satisfaction in these studies rose when listeners could adjust the mix themselves, which is an argument for a control within easy reach rather than a setting buried three menus deep and left alone.

Does a soundbar fix it?

It depends entirely on which of the four causes you have. A television's built-in speakers are frequently the weakest link in the chain — flat panels gave up a great deal of enclosure volume and driver area compared with the sets they replaced — so anything with real drivers can improve matters. But a soundbar that folds a multichannel mix down without a genuine centre channel has changed the loudspeaker, not the ratio. If your problem is a mis-levelled centre or an aggressive downmix, the bar inherits it.

When you've exhausted the room

Here is the honest part. If you have set the centre level, trimmed the background, disabled the processing stack and worked through several programmes, and speech is still hard work while everyone else in the room is comfortable, the variable left is your hearing. Age-related high-frequency loss is common and it is not something a speaker layout addresses. That question belongs with an audiologist, not with an audio retailer, and we are not going to pretend otherwise.

There is a category of personal television listening system that sits awkwardly between hi-fi and hearing assistance, and it exists because it solves the ratio problem outright for one viewer — they get their own speech-to-background balance without imposing it on the room. The trade-off is equally plain: it removes that person from shared listening, which for some households is a genuine cost and for others is no cost at all. That is a judgement about how you watch television, not a technical one, and it is yours to make.

What all of this comes back to is the same finding. The dial that moves speech intelligibility is the difference between dialogue and everything underneath it — and unusually for audio, that is a claim with measurements behind it rather than folklore.

Frequently asked questions

Why does turning up the TV not make dialogue clearer?

Raising the television volume lifts speech and background together, leaving the loudness difference between them unchanged. Research on speech intelligibility points at that ratio, not overall level, as the variable that matters. An Oldenburg sentence test found that attenuating the background by 6 dB and 12 dB significantly improved intelligibility for both hearing-impaired and normal-hearing listeners.

What does a dialogue-boost or speech-clarity mode on a TV actually do?

Dialogue-boost, speech-boost and clarity modes are the consumer implementation of what dialogue-enhancement research tests: they alter the loudness difference between speech and background. How much of any given mode is genuine separation-based processing versus simple level and EQ varies considerably by brand, so results differ between manufacturers and between programmes.

Is it only people with hearing loss who struggle with TV dialogue?

No. A multimodal study of 28 normal-hearing participants measured listening effort via pupil dilation, self-reported effort and word recall, and found effort decreased consistently as background level decreased. Even listeners without hearing loss spend measurable cognitive effort on mixes with a low speech-to-background ratio.

Is there a correct centre-channel level for dialogue clarity?

The research says no single number applies. In studies where listeners set their own preferred speech-to-background difference, preferred levels varied greatly between participants, within the same participant on repeat trials, and across different programme items. Treating centre level as a setting you find and revisit is better supported than chasing a fixed figure.

What should I do if setup changes don't fix TV dialogue?

If centre-channel level, background attenuation, speaker placement and disabling extra processing have all been tried across several programmes and speech remains hard work, age-related high-frequency hearing loss is a common remaining cause. That question belongs with an audiologist rather than an audio retailer, and cannot be resolved by speaker layout.

This article was written by an AI system from Into The HiFi AI Desk, generated from the following sources, with no human editing pass before publication: Straninger, "Dialogue Enhancement in Object-based Audio: Evaluating the Benefit on People above 65" (Ansbach University of Applied Sciences, 2020), "Dialogue Enhancement and Listening Effort in Broadcast Audio: A Multimodal Evaluation".