Listening Room Intelligibility Test

[mep]

So Frank, how did your system do on the test?

 

[garylkoh]

Frank, it would be great if you could post your results here. Once I get a mic preamp, I'll do some recording.

 

[fas42]

Frank, it would be great if you could post your results here. Once I get a mic preamp, I'll do some recording.

You people caught me on the hop! I had to quickly go away and burn a disc to try it out. Also, it's raining at the moment with a metal roof at the far end of the house, so it was played at maximum volume to pick it up clearly at this point.

Couple of interesting points: track is very noisy, lots of hiss, worst I've come across, perhaps from the MP3 encoding; there is no sense of garbling irrespective where I go, at the furthest point of the house, where the rain is making things slightly more difficult at the moment, or if I go right up to the speakers. Sounds clean and unchanging in quality.

The tapping is very clear and distinct irrespective of where I go, though in the actual room it varies in intensity depending on how I cock my head and move it around with respective to the side walls. I think I understand what Art means about the yodelling; the pulsing of the sound creates something like an overtone or sensation which you could apply that term to.

Finally, at the end of the track, there are some very distinctive artifacts of sound as the oscillators got switched off, or something like that.

Frank

 

[mep]

No garbling, no matter where you go in the house with the Phillips volume maxed out?

 

[Mark Seaton]

Hello all

I’m Art Noxon and I’m very pleased to be invited to work in the Expert position here and represent the audiophile version of room acoustics.

I believe I need to begin by apologizing in advance should I display anything but absolute neutrality towards all acoustic products and techniques. I will do my best to explain and support the appropriate application of any product or technique we are covering, but still, I do have my favorites…..

Where do we begin this adventure? I have selected a listening experience as the opening salvo. Audiophiles tend to want to listen first and talk second. It is the nature of the golden ear crowd. I will always assume I am addressing golden ears, even though there may be some technical people and collectors out there as well.

High end audio started for me in 1984, having just invented the TubeTrap. In the beginning I spent a lot of time trying to figure out why audiophiles got so excited about putting a pair of TubeTraps in the front corners of their listening room. Yes. I was pleased and proud, but being the hard core engineer that I am, I needed to measure, to quantify, what caused their thrill. I tried all the standard tests known then, which are the same ones known today; sine sweep, pink noise, RT-60s, narrow spectrum analysis, TEF waterfalls, and even Q changes of resonant modes.

What did I get? The best was about 1 dB adjustment in anything, usually less. Certainly not enough to warrant the pleasure those trapped corners gave the audiophile crowd. I published my findings (failure to find) in the AES and didn’t know what to do after that. In the mean time we had gotten so bored with reverb chamber testing of TubeTraps we tried to speed the process up and were getting very illuminating results using rapid short tone bursts. Not only in the chamber but in real rooms we were documenting changes upwards of 10 to 15 dB for rooms with and without TubeTraps in the corners.

Coincidently at this time, 1986, I was going to the SynAudCon meetings. Victor Peutz, from the Netherlands, showed up and presented the new/next wave of audio performance testing: Intelligibility. I couldn’t believe my ears. I realized right then and confirmed with him that our fast tone burst testing was actually a tonal intelligibility test. Much more has come to light since that first peek into the wonderful world of audio playback intelligibility, which turns out to be what really matters the most in hifi playback.

And so, here is the link to the MATT TEST for you to click on. Included here is a description of the MATT test, a training demo and a downloadable MP3. MATT means Musical Articulation Test Tones. This test is also on TRACK 19 of THE STEREOPHILE TEST CD 2 and many audiophiles have that reference CD. Do your own A/B test. Listen to the signal first on headphones and then listen to your audio system try to play it. We’ll talk about what you hear soon enough.

For now, report into the forum about what you heard……………………………… Arthur Noxon

Hi Art,

Great addition and ironic how it has taken much greater awareness of the benefit of listening room acoustic treatments for more listeners to be more receptive to the ideas.

I haven't had time to do much testing with the MATT signal, but have some later, but related experience with STI measurement. I had just been laid off from an engineering position back in 2001 and had started helping a pro audio company get into the home market and better support specifying consultants. The weekend after I had just turned my long time hobby & moonlight work into a full time endeavor, I was directed to a Gold-Line/TEF workshop focused on the topic of intelligibility measurement (particularly STI-PA) which was local to us in Chicago. At that time having mostly a home audio background, the concept of quantification of intelligibility was rather new, as this just wasn't really discussed in the home audio world. After quite a few "Ahah!" moments in the course of a few hours, being the younger local of the group, I was asked to direct a dinner excursion. Through the course of a dinner at a quaint and very unique little pizza joint down the street I started to realize the crowd I had stumbled into... Aside from many other well known names from the pro audio world, I was having some great conversation with Dr. Steeneken of the TNO sitting next to me. That rather inexpensive but fun dinner led to many connections and let to attending a handful of further Syn-Aud-Con and later TEF workshops where modern Speech Transmission Index was always an important topic. I heard many a joke about the looney audiophile & home theater market, as it was clear most "didn't get it." It has been nice to gradually see increasing acceptance of more scientific efforts in the home audio world, which brings me back to MATT test and STI...

Have you looked into correlation between the different numeric STI scores and the results or subjective valuations of the MATT test? While plenty of devices are just concerned with an overall pass-fail score for public address and emergency alert systems, a variety of packages allow further dissection of the scores into scores per bandwidth like the package for the TEF 25. For those following along, it should be noted that the original development by TNO was focused on speech intelligibility in classrooms, not for reproduction systems. I have generally found listening rooms to fall well into the acceptable range of overall intelligibility (>0.90), but differences of a few points (0.01) correlate to significant improvements in clarity. Breaking down things to specific bandwidths can reveal weaker ranges where more attention is warranted.

The various STI measurements primarily look at modulation of different frequency ranges depending on the focus/interest of the specific test. While obviously not the only parameter of interest, it might be that a narrower focused range of such tests could be extracted based on the knowledge that no useful room will exhibit a score lower than 0.75-0.85 on an STI-PA scale. At least there is much better understanding of some of the various mechanisms and characteristics at work different sounding rooms & systems.

 

[Mark Seaton]

No garbling, no matter where you go in the house with the Phillips volume maxed out?

I thought it was obvious that Frank was walking around the house with the small diameter speakers strapped on as headphones. It's a great way to really boost the direct to reverberant levels.

 

[garylkoh]

 

[mep]

That would 'splain it.

 

[vinylphilemag]

Looks like Stax's Earspeakers, before they refined their prototype!

 

[fas42]

I thought it was obvious that Frank was walking around the house with the small diameter speakers strapped on as headphones. It's a great way to really boost the direct to reverberant levels.

Beautiful, boys!! I love it! The only trouble with that poor fellow's efforts is that his speakers are too big, mine are about half that size, I mean, I don't want my ear lobes resting on my shoulders ...

Yes, Mark, there was no garbling, as far as I was concerned. Somebody else with fussier ears may beg to differ, of course. In terms of maximum volume, looking at the waveform and having a reasonable guesstimate of the Philips output being limited to about 105dB at 1 metre, the MATT signal in the room would have been about 100dB at that 1 metre, not that extreme ...

Slightly veering off track, the friend with the analogue setup invited us over out of the blue yesterday, and it has been the experience every time that in between visits he does a round of tweaking, which in his view has moved things forward, but which to me always end up making his system sound a bit out of whack. We listen, I comment, suggest things to try, and lo and behold, by a bit of trial and error we manage to pinpoint little subtleties that are pulling the quality down; by the time we left it was humming along very nicely indeed. The direction his system always goes wrong in is that it starts to sound harsh, edgy on difficult recordings, loses musicality; he had a so-called audiophile LP of Tapestry, it was so off-colour initially his wife had to go off into another room because it disturbed her. But a bit of fiddling brought the system back on song, at the end of the session she was listening to an early 50's LP of Heifetz playing Mendelsohn's Violin Concerto and she was zoning very happily. Sometimes it doesn't take much to get things right, or at least much better ...

Frank

 

[Phelonious Ponk]

it was so off-colour initially his wife had to go off into another room because it disturbed her. But a bit of fiddling brought the system back on song,

Don't be shy, Frank. What did you fiddle?

Tim

 

[mep]

It's probably Top Secret.

 

[fas42]

Don't be shy, Frank. What did you fiddle?

Tim

Key things were mechanically related: as people are well aware vibration control is very important, so an experiment he had tried of damping the catridge in the arm was doing the wrong thing: it was adding almost a reverb to the sound, if the recording had minimal echo it seemed to enhance the sound, but if the recording had plenty of such then it clashed with that intrinsic to the production: it was a distortion, pure and simple. So he removed that material, and luckily he had some quite sticky viscoelastic goo, better quality Blu-Tak, so to speak, so we smeared a very fine layer of that on so it filled the gap between the top of the cartridge and arm, tightened it down as firmly as we dared with the screws. In this case, better mechanical coupling of the cartridge to the arm, with the goo working to dissipate some of the vibration resonance. Much, much better sound.

The second item he used was roller blocks under the Naim amp, he's keen on using those, but I wasn't happy. Took them out of operation, problem solved! Again the vibration was the problem, the blocks were attempting to control it but weren't doing the job correctly for the amp. I believe strongly that you have to damp down vibration, shock absorb it so to speak, rather than move it around ...

Frank

 

[mep]

Did he put the roller blocks back under the Naim as soon as you left?

 

[Phelonious Ponk]

Did he put the roller blocks back under the Naim as soon as you left?

If they were successfully driving his wife from the room, I'd guess yes.

Tim

 

[fas42]

If they were successfully driving his wife from the room, I'd guess yes.

Tim

Now, now, no naughtiness: these are good friends, so no unpleasantness, please.

When four people out of four decide that something sounds better one way than another that's pretty good consensus ...

Frank

 

[mep]

Now, now, no naughtiness: these are good friends, so no unpleasantness, please.

When four people out of four decide that something sounds better one way than another that's pretty good consensus ...

Frank

Unless it's appeasement.

 

[Phelonious Ponk]

Now, now, no naughtiness: these are good friends, so no unpleasantness, please.

When four people out of four decide that something sounds better one way than another that's pretty good consensus ...

Frank

No unpleasantness, Frank, just a bit of humor. I could bring four people into the speaker studio at the store, tell them what they're going to hear and get the exact consensus I'm looking for more often than not. Four out of four, sighted and persuaded, isn't even a cold start.

Tim

 

[Art Noxon]

Hello Frank et al,

I really appreciate your finding value in my work. In the end, that is why I do it. What good would it be to explain one of nature's secrets, for example, to a tree?

I'm curious about your statement about electronic blurring. My frame of reference with respect to room acoustics has always been that the sound systems we have in the audiophile kingdom have infinitely more instantaneous power handling capability compared to the very sluggish way rooms we have. Here, you have suggested in effect that most “room acoustic” problems go away if we use great amps. At first this doesn’t intuitively make sense. But, it’s always best to assume the audiophile is right. So let’s try and see where it might leads us.

1) We do like to use a very loud MATT to run a buzz check on a loudspeaker. Using a hand swept sine doesn’t work because the speaker gets blown more often than not when there is a problem frequency. A fast sine sweep doesn’t stay at any one frequency long enough to get things going. Using a very loud MATT delivers power for 1/16 sec and turns off. That time is long enough to get vibrations moving but not long enough to burn the drivers out due to some surprise feedback problem with the box. So, internal box problems would cause blurring. Since speakers are part of the electronic part of the audio chain, that’s electronic blurring.

2) I can well imagine that a rapid succession of high level instantaneous power bursts could tax the dynamic headroom of a sound system. What would happen is that initially the tone bursts would be loud, but over time, the sound level of the bursts would diminish and finally distort because the rapid succession of high energy bursts drained more energy out of the amps caps then the power supply could replenish and the over all system supply voltage drops into a brown out. We’ve never used MATT in this matter.

The primary measurement in the MATT test is the sound level difference between when the tone is on and when the tone is off. If the amp is starting to brown out, my guess is that it is still delivering sound, just not as loud of sound. This does not affect the result of the MATT test, because how quiet a room gets between bursts depends on the room acoustics, not on how loud the burst is. So, low headroom in a quiet room does not cause the loss of intelligibility, just a loss of loudness.

Now, it’s true, a very loud burst can get surfaces moving that otherwise, with more quiet bursts, would not get moving. But let’s skip this aspect of things for now.

3) The human can hear the full dynamic range, zero to well over 100 db. However, when listening to very short bursts of noise or tone, the human can only perceive sound level changes between the ambience and 20dB above the ambience. The MATT test may have a 60 dB electronic signal difference between the rapidly alternating sound level when the burst is on and when it is off. The human listener can’t tell the difference between 60 dB bursts and 20 dB bursts. However, below 20 dB bursts we do hear differences. A 20 dB burst pattern sounds louder than a 15 dB burst pattern and much louder than a 5 dB burst pattern.

It’s interesting to add electronic headroom to this discussion, If a human can detect a difference between an instantaneous short lived tone burst that is 10 dB compared to 15 dB above mean playback level, then the sound system probably should be able to play that + 15dB burst above mean playback level. Since we humans do not report any difference between bursts that are louder than 20 dB there is likely little to be gained by having a sound system with more than 20 dB dynamic headroom.

I think this might be a good thing to kick around for a while. Who has experience with auditioning low and high headroom amps and is there a practical limit to the amount of audible headroom? If so, from the psychoacoustic side of things, the limit should be around 20 dB. …………………………………Art Noxon

 

[fas42]

I really appreciate your finding value in my work. In the end, that is why I do it. What good would it be to explain one of nature's secrets, for example, to a tree?

The trouble, Art, is that some people here believe I have the intelligence of that tree!

Some interesting new terms were mentioned by you: "electronic blurring" and "supply voltage brown out". System blurring or lack of dynamic headroom in reproduction is something I became aware of quite some ago, and this is an area I have put a lot of energy into addressing. Simply expressed, I wanted the system to sound exactly the same, not change in sound characteristics as the volume was changed, up to the limits of the system's physical capabilities; the same as the sound of the real event, in other words. Being an EE, I naturally approached this from the side of the electronics behaviour, and concentrated on power supply behaviour to largely eliminate the problem.

You mentioned "low and high headroom amps and is there a practical limit to the amount of audible headroom". A low headroom amp seems to be fairly typical amongst audio people, in the sense that you're using the phrase, my experience is that a high headroom unit does make a remarkable difference; the last aspect I quoted from you, limits of "audible headroom", again my experiences tell me that they are high, very high, the human ear/brain is quite remarkable in its ability to handle intense pressure differences, in fact this is part of the mechanism that immediately distinguishes that a sound sample is real as distinct from normal reproduction.

Frank