[MRG1]

When I bought a mid-range (analog) HiFi around 1982, I was told to pay attention to two things: frequency range (e.g., 20-20000 Hz, +/- 3 dB, though I admit the speakers I bought only went up to about 16,000 Hz) and THD (I was told to stay below 0.01%).

Virtually any modern AVR supposedly meets the 20-20,000 Hz criteria - if anything, I suppose the higher frequency capabilities and ditigization might introduce digital aliasing effects.

I hadn't heard yet of enharmonic distortions, such as are frequently produced by intermodulation effects.

I had heard of phase distortions - though of course that is only a subset of timing distortions.

Many modern consumer market amplifiers and AVRs have THDs in the 1% range, though when I bought my (used) AVR, I decided to keep THD fairly low - after all, fairly old used AV equipment is pretty cheap. I got a Yamaha RX-V863. Here are the specs from the owner manual:

I'm not sure about the enharmonic and intermodulation distorition figures. For that matter, I understand that THD is only measured at one frequency - sometimes the frequency at which the amplifier or AVR is best at, and at the volume level the amplifier or AVR is best at.

An interesting question is whether any of these distortions - including THD in the 1% range of many consumer market AVRs - significantly affect dialog comprehension, in TV shows and movies. And whether an even better AVR would have improved my dialog comprehension.

Many sources say most of the sounds that affect phoneme (and microphoneme) discrimination of American are roughly in the 40-4000 Hz range (with some variation in the stated range). E.g., see

Generating and understanding speech | Ecophon

https://www.linguisticsnetwork.com/the-basics-about-acoustic-phonetics/

https://www.linguisticsnetwork.com/phonetics-the-basics-about-acoustic-features-of-consonants-in-standard-english/

Sounds of Speech | MESHGuides

www.meshguides.org

Some sources say people create varying amounts of those frequencies by altering the shapes of their throats, mouths, and lips. (And in fact, different people do this somewhat different ways - as they must, since the typical fundamental frequency range of their voices are different, and the shapes and acoustic of the vocal tracts are different.)

Many modern speakers don't reach 20-20,000 Hz - but many do include the 40-4000 (8000?) Hz frequencies, so maybe that doesn't matter to dialog comprehension. I confess I bought my speakers as an afterthought, and they aren't all that great - just the speakers that came with an even older Onkyo SKC-540C Home Theater (Front: SKF-540F, Center: SKC-540C, Surround: SKM-540S, Surround back: SKB-540, Powered Subwoofer: SKW-540 - which I don't use). They don't include the lower frequencies. Here are the specs for a somewhat similar system:

Note that the center channel speakers only go down to 55 Hz. I did not find distortion figures for them.

The (used) headphones I bought are somewhat better: Sennheiser HDR 175, rated at 17 - 22,000 Hz, but still have fairly high distortion figures: THD<0.5 % at 1 kHz, 100 dB SPL. I am using them instead of the speakers, because because I am sharing a home with other people, and because they sound better than those speakers.

What I'm really wondering is whether any of these distortions could produce sufficient contributions in the 40-4000 (or 8000?) Hz frequency range to affect comprehension.

In addition, timing distortions in a typical home theater system might be significant, since the different speakers and speaker cones are in different locations. That should mean that the time of arrival at your ears of different sound components is different, which might make it hard to understand many consonants. Perhaps many phonemes and microphonemes might arrive mixed together, as in a room with substantial echoes. Maybe that is part of why I understand dialog better using my headphones than my speakers...

So - are noise figures like THD, EHD, IMD, and timing distortions important to dialog comprehension? Is it possible that people buying consumer market home theater equipment - not just me - sometimes lose dialog comprehension in part because of that?

(Of course, the restriction of dialog to the center sound channel must create another comprehension problem. Many sources say that most people obtain a lot of information for dialog comprehension in normal human conversation by lip reading - i.e., looking at the mouths and lips of the speakers. We also obtain some information by looking at their bodies for gestures. They talk about "gaze behavior" - that we look at the person speaking - which means that in the ideal case, the sound should be heard by us to the location of the speaker on the screen (not in real 3D space). But I guess there isn't much we can't do much about that - that's the way TV shows and movies are made, and even if some of the dialog occurs in other speaker channels, it is unlikely to be mapped to the screen positions of their mouths and bodies - so a lot of that information must sometimes be lost. Is there anything we could do about that?)

 

[Ryan Statz]

I think you're overthinking things.

 

[m. zillch]

Typically modern day, solid state AVRs like the Yamaha mentioned have perfectly fine distortion IF played at volume levels where the loudest peaks in the signal are kept below the amplifier's "clipping" point. Unfortunately hardly any provide a visual indication of when this red-line level is exceeded, using for example clipping indicator lights or power level meters found on upscale gear, so you can only go by your ears unless you invest in expensive test instruments. There is also no way of describing a specific volume knob setting as one's "safe maximum limit" because the power used at any given moment is not just dependent on that but also the particular connected speaker's varying load per frequency , phase angle, and the incoming musical signal content's level and frequency.

Moral of the story: if there 's something you don't like about the sound and this occurs even at low to modest playback levels (so you are confident there is no clipping going on) then its not amp distortion.

Speakers, on the other hand, even expensive ones, have mind-boggling higher levels of distortion, especially in the bass, but it is so embarrassing the manufacturers never talk about it in their specs.

When distortion does first start to occur audibly it happens on the loud peaks first but then as you continue to turn it up it adds a buzzy type of sound, easiest to first detect on pure tones or sustained flute or piano notes.

 

[MRG1]

Are you saying that even the AVRs with 1% THD are good enough for good dialog comprehension, if the source material is reasonably well done?

 

[18Hurts]

The one that creates the most distortion creates the largest problem,

If you want to know what an amp does, look at the tests and pay particular attention to the range of power you will use. This is typically a graph and you will note that distortion is very, very low over most of it's operating range. The random number thrown up is marketing, pick a point on the graph and go. Since distortion radpidly rises as you start to overload the amp, then pick whatever number of distortion you like, look at the amount of output and publish that.

The big one is the speaker itself. Commonly, distortion easily goes past 1 percent and commonly hovers at 3 to 10 percent or more at higher outputs. If a AVR generate 0.08 percent distortion at XX watts and the speaker generates 3 percent distortion at XX watts then the speaker wins!

Vocal clarity then goes past distortion once the speaker moves, it then has to deal with the room, reflections and the dispersion of the speaker. If you are listening 30 degrees off axis and the speaker only measures to 20 degrees off axis at -3 or -6dB down, then whatever it measures where you are sitting is what you get. The amplifier distortion or speaker distortion has nothing to do with it if you are not seated in it's operating range. Then throw in the sound will reflect off the floors, ceilings, walls and the dog--and remix with the direct sound which depends on your room and placement--then ponder how much 0.08 percent distortion means at the end of the chain.

So, if your center channel is a mess--changing the amp will achieve nothing unless you are clipping the amp (of course) Look at the measurements of your center, determine if the distortion is low enough, the frequency response is accurate enough across the vocal passband, if the dispersion is correct for the nujmber of seats AND check your placement and first reflection points (you can find those with a mirror, look it up)

Judging by the sheer number of posts by people that can't hear the voices--this is a common problem with the up and coming HT crew. I really wish amps were magic and could fix room acoustics problems, placement problems, speaker design and performance issues but they can't. They are just dumb boxes that make little signals into larger signals.

Sometimes, the HT person gets weird so always a good to have a second opinion. Don't want to get all wrapped around the axle and waste time, expense and space on random shots in the dark. My wife happily points out any flaws in any project I work on--she is cool like that! My goal was to have a few people watch a movie and not comment on the sound--it just is in the background, is done right and nothing to worry about. Normally, I get comments on the subwoofers--cuz subs are cool which tells me they can hear the vocals clearly.

As far as worrying about amps, if you need more power to prevent clipping--get more power! If the amps are noisy and you can hear hum--need to fix that. If you want to gain something, then get speakers that have far lower distortion than the amplifiers do so it will actually solve a problem. My speakers put out far more distortion than my amplfiers, by at least 10 times higher at higher volumes so I'm working on the speaker distortion problem first. I caught a lead on a new and improved speaker from Unicorn Audio. It is there new unobtanium model which claims 0.05 percent distortion at 110dB at one meter--seems there is a supply disruption because the pixies want profit sharing. The classic pixies VS the elves are in play here but they should start shipping by the end of the first quarter.

Hope that helps you, I find that fixing the largest real problems first tend to give me the best end results. Have a great weekend!

 

[m. zillch]

Are you saying that even the AVRs with 1% THD are good enough for good dialog comprehension, if the source material is reasonably well done?

Essentially, yes.

 

[RayGuy]

Go listen to some new speakers. Pick ones you like and bring them home. Listen again. If you like, keep. If not, return and start again. The speakers and the room have far more to do with what you hear than the amp.

 

[Ryan Statz]

Are you saying that even the AVRs with 1% THD are good enough for good dialog comprehension, if the source material is reasonably well done?

It hinges more on your speakers than the amp, but the mix of the content is also important. Just don't use Tenet as a test 🤣

 

[CharlesJ]

Are you saying that even the AVRs with 1% THD are good enough for good dialog comprehension, if the source material is reasonably well done?

Dialogue volume and power needed should never be in the range to create that 1% distortion that usually happens at high power output close to the start of clipping perhaps.

If you have dialogue issue, there are threads out there discussing why. Have not seen any mention that it is due to that 1% distortion or any distortion.

 

[a_ok2me]

I hear distortion is what makes tube amps sound good.

 

[CharlesJ]

I hear distortion is what makes tube amps sound good.

So then dialogue should be excellent even when recording is terrible.

 

[a_ok2me]

So then dialogue should be excellent even when recording is terrible.

No that's not how it works. When an actor says, "FU", it comes out as "Forget You".

 

[Worf]

Amplifier distortion starts low then as you get to the limits of the amp distortion increases. So one way they boost the numbers is increase the distortion.

So if you have an amp capable of say 80W at 0.01% THD, you can probably say it's a 90W amp 0.1% and maybe 95W at 1%. Distortion of solid state amps increases rapidly because the amp starts to hit the rails. But hey, by playing with the numbers we boosted the power by nearly 20% from 80W to 95W. Consumers like big numbers.

No, what affects dialog are the speakers, placement, room, and of course, the content. Speakers and room are the biggest variables that affect the sound - you can take the best equipment but stick it in say, a bathroom and it'll sound terrible because the acoustics of most bathrooms are terrible with flat reflective surfaces. Content is important because if you watch a Nolan movie, you will have problems hearing the dialog no matter how perfect your room and speakers are.

 

[MRG1]

I share a home. I never listen to anything on a loud setting. So amplifier power is almost certainly irrelevant. I'm fairly certain I'm nowhere near saturation.

I have so far dealt with dialog comprehension problems by turning off the TV's own sound (I guess most thin TVs have lousy speakers), playing with a 31 frequency graphic equalizer on each channel, sometimes using center channel boosting, using headphones, and using a "standard" (no reverb) sound program (the latter two eliminate echoes). Sometimes I rewind and listen again. If there are closed captions or subtitles, I turn them on. That's all fairly effective with some sources. But isn't enough for others.

Part of my problem is that my interest is in the dialog and storyline, not the music, background sounds, and special effects. I want to understand every word. I don't think many modern show and movie makers think that is important. They have other priorities.

I could give you any number of extra reasons why dialog has become harder for me to understand than a few decades ago - mostly having to do with how dialog and background sounds are balanced now, and how actors are (not) trained to speak as clearly now. Maybe that I don't use carpets anymore. Plus, I'm older.

My AVR probably isn't capable of anywhere near as much as 1% THD from external sources. But I've noticed that a lot of the AVRs sold in big box stores are rated at that. As though they don't think it matters much. That's part of why I wondered if it does, for dialog.

 

[nathan_h]

So - are noise figures like THD, EHD, IMD, and timing distortions important to dialog comprehension? Is it possible that people buying consumer market home theater equipment - not just me - sometimes lose dialog comprehension in part because of that?

Nope.

The culprits are room elements (too reverberant), sound level (playing below reference level), background noise (ie, other sound in the room which "masks" nuances in the playback), crappy speakers and or poor speaker placement, and (for many of us) deteriorating hearing with age.

Solutions are seldom found in the gear itself, though a good center speaker with wide even consistent dispersion and relatively flat frequency response can help. And using dynamic range compression (Dolby Volume, Night Mode, Audyssey Dynamic Volume, etc) can help, too. But after that, things like acoustic treatment, better speaker placement, better base management, etc begin to become necessary steps.

 

[a_ok2me]

Nope.

The culprits are room elements (too reverberant), sound level (playing below reference level), background noise (ie, other sound in the room which "masks" nuances in the playback), crappy speakers and or poor speaker placement, and (for many of us) deteriorating hearing with age.

Solutions are seldom found in the gear itself, though a good center speaker with wide even consistent dispersion and relatively flat frequency response can help. And using dynamic range compression (Dolby Volume, Night Mode, Audyssey Dynamic Volume, etc) can help, too. But after that, things like acoustic treatment, better speaker placement, better base management, etc begin to become necessary steps.

I'm not sure about that. There was a time I watched something in English and then i hear, "Si papi."

 

[Tiga]

I'll admit to never considering amp distortion as it relates to dialogue comprehension. Seems a bit involved. I've typically had good success with focusing on a decent speaker in the CC position and using acoustic treatments, especially behind the listening area.

 

[MRG1]

Why would I want a wide dispersion center channel speaker? I would have thought a narrow dispersion speaker, aimed directly at me, would create fewer echoes.

 

[nathan_h]

Why would I want a wide dispersion center channel speaker? I would have thought a narrow dispersion speaker, aimed directly at me, would create fewer echoes.

The main key is consistent dispersion (where the off axis frequency response mimics the on axis response).

But since most people in the room (especially if there is more than one person in the room) are off axis, it is often helpful if the dispersion is also wide, otherwise the volume will be lower (and masked by the Left or Right speaker) for people the further off axis they sit.

 

[DeepPurpleFan]

What I'm really wondering is whether any of these distortions could produce sufficient contributions in the 40-4000 (or 8000?) Hz frequency range to affect comprehension.

The variation or 'distortion' of a "real" loudspeaker set up and playing in a real room is more than an order of magnitude greater than the measured distortion of any amplifier benchmarking spec. That's not saying those specs are unimportant or inaudible, but even under a worst case scenario they only make a small contribution to degrading the sound.