Loudspeaker listening - what I hear

[STHLS5]

Alan wrote (More later but to get you thinking .... record at a nominal level of -18dB on a 16 bit DAT machine and what resolution are you recording at?)

Good to see you back from your busy schedule with your new factory. The thread is closed but I think the original post shouldn't be there in the first place.

My answer as an average music lover with zero knowledge in audio engineering would be 13 bits, but then I am maybe wrong because 1 bit resolution affects the dynamic range and therefore, it may depend on the recording dynamic range.


____________________________

Back to the title,

In the tech talk, you have describe the ideal loudspeakers should project sound as described in the image below.
Attachment 1421

I am not sure if any of us could hear the sound projecting strictly behind the speakers. I am sure what I hear is sound as described by the image, but they do extend to the front of the speakers but as described as in the picture.

Maybe, it is just me because I have never heard the other effect such as sound extending beyond the room wall. My brain just can't accept that's possible.

ST

{Moderator's comment: You're right about the -13dB but do remember what Alan stated was the recording level - nominal? average?. As for image *behind* the speakers are you sure that you've never heard this in other speakers?}

[STHLS5]

I hope this image will put across my point more clearly. My perception is the 3D image of the sound cover both from the front of the speakers to the back wall.

Attachment 1424

I do get this with other speakers but not as good and undistorted as Harbeth.

ST

[A.S.]

...My perception is the 3D image of the sound cover both from the front of the speakers to the back wall...

Ah ok. I hadn't written the TecTalk article with the intention of showing what you call '3D depth'. That's because it wasn't the point I was trying to convey. What I was trying to get across was that certain colorations typical to loudspeakers could bias our perception of where the sound stage appeared. (See article here for the full story).

What I think you are commenting upon is how you interpret the well balanced, coloration-free left-right sound stage to give it some apparent depth. But whatever illusion of depth your brain conjures cannot be anything more than an enjoyable trick. That's because there is absolutely no real depth information collected by simple L-R microphones arranged across the sound stage. In fact, although the sound engineer may have sprinkled spot microphones from the front violin section to the tymps at the back of the orchestra 10 or more metres behind, he will have mixed the whole lot down to one 'flat' left + right mix, with all the performers overlaid on top of each other (in the Z , front to back) plane - unlike what you would have actually experienced at the concert hall. He may have taken an additional step of digitally time aligning the spot microphones relative to some fixed point - say relative to the violin section nearest the conductor - to 'bring forward' those performer's microphones deeper into the orchestra. There are subtle differences in the sound but you would have to be an expert to hear them so it's a technique rarely used - and impossible in the analogue era.

Alternatively you could be saying that the L-R sound stage seems correctly balanced (i.e. no obvious bulge towards you or dished away from you) but that the entire orchestra seems to have stepped towards your listening position (this is what I at first thought your picture indicated). The solution to that is simple. Turn down the volume! As QUAD's Peter Walker said - the volume control gives you the ability to zoom into the recording. In other words, to bring the performers nearer.

If we are able to construct a fantasy 3D sound image in our brain of illusory performers in space behind our loudspeakers the vividness and stability of this image is going to vary from listener to listener. It must to some extent depend upon preconceptions of how musicians arrange themselves in front of microphones, familiarity with the music and our ability to fantasise. I wonder if an Amazonian aborigine, completely isolated from western music were to listen to a pair of loudspeakers, would he hear '3D depth'? I doubt it.

[STHLS5]

..Alternatively you could be saying that the L-R sound stage seems correctly balanced (i.e. no obvious bulge towards you or dished away from you) but that the entire orchestra seems to have stepped towards your listening position (this is what I at first thought your picture indicated). The solution to that is simple. Turn down the volume! As QUAD's Peter Walker said - the volume control gives you the ability to zoom into the recording. In other words, to bring the performers nearer....

Yes, you have described exactly what I intended to convey. However, this brings us to another problem, i.e, what should be the realistic volume? Should I set the volume so that the sound stage 'sticks' to the L-R plane of the speakers (that will be a very low listening level) or should I adjust the level to be more realistic level of a live instruments heard at a comfortable distance but at the expense of the instruments seeming to step closer to the listener i.e. in front of the loudspeakers?

Could it be that for someone who is familiar with the psychoacoustics of listening to loudspeakers may interpret the sound they hear very differently to those of us whose sole point of reference is our diet of reproduced stereo sound, not real live sound? Just like I cannot image sound to be extending laterally beyond the L-R walls (though I have described so, previously) and therefore my brain can not perceive this to be possible although other listeners may willingly allow their brain to accept such tricks.

..That's because there is absolutely no real depth information collected by simple L-R microphones arranged across the sound stage. ....

I have been listening to Erroll Garner (Telarc) for the first time just now, and I observed that the recording was a strict dual mono (stereo). On the left, it was a piano and on the right - bass and drums. Interestingly, in this case the sound is confined to the speakers. I believe I experienced the same in Sonny Rollins (Way out west?) another dual mono stereo recording where the sound seemed to be stuck in the speakers.

My observation is if the sound is mono on one channel (NOT BOTH) the sound seemed not to float but sticks to the speaker's space though occasionally you hear them shifting up and down due to different notes, like more emphasis to the tweeter than the woofer. So if the same note is played in mono on both channels the sound seemed to be floating in the centre (some sense of depth) but when the sound is recorded in stereo, and if it is predominately on the left or right you get a sense of the sound to be out of the speakers.

I believe stereo do create a sense of depth even without the digital trick. Could phase difference do this trick?

.. I wonder if an Amazonian aborigine, completely isolated from western music were to listen to a pair of loudspeakers, would he hear '3D depth'? I doubt it. ....

When I was played the sound of waves CD, I know children (about two or three years old) to perceive the 3D depth when they refuse to put their feet on the floor because they perceive waters rushing to their legs. They were aware the sound was from the speakers by their eye movements which clearly showed they were following where they expect to see the wave crushing and the sound of bubbles from the wave foams sinking into the sand.

ST

[Drdennis]

I listened to a live broadcast of the Boston Symphony Orchestra last night on analog FM (Marantz 10B) and C7ES-3. The soundstage was bigger and deeper than my room! It just seemed bigger than previous broadcasts. I have tinkered with my system, removing an active power conditioner, and replacing it with a passive design. I have also changed out some cables. Of course the broadcasters may have made changes such as mic placement, or upgraded equipment. Give a listen next week, their internet broadcast is fairly high quality, but if you are local analog FM is preferable on a good tuner. The broadcast starts at 7:00 pm EDT, and the performance begins at 8:00. You can find it at "99.5 All Classical".

[A.S.]

There are several points you make. I've lightly edited your post (and my previous one) to draw out the points I believe you want to make. Please check my edits.

...this brings us to another problem, i.e, what should be the realistic volume? Should I set the volume so that the sound stage 'sticks' to the L-R plane of the speakers (that will be a very low listening level) or should I adjust the level to be more realistic level of a live instruments heard at a comfortable distance but at the expense of the instruments seeming to step closer to the listener i.e. in front of the loudspeakers?...

Well this is a really interesting point. How do we know, in Peter Walker's terms, that we have applied just the right amount of 'zoom' to the recording when we were not there.

I suspect that many 'audiophiles' listen to orchestral music at an average reproduced loudness higher than they would experience in the hall. In effect they enjoy the experience of repositioning themselves from say a nice natural perspective in row ten down to a seat in row one or two. That would give a more exciting, 'in your face' sound and crucially, dry-out the sound as we discussed at length here. So, they would here more of the direct sound from the orchestra and less of the hall, simply because the hall reverb would be more effectively swamped by the louder direct sound.

The clue to answering your question about what is the optimum replay loudness (for any given track) is partially concealed in that very point. Unlike the recording engineer who can play around with microphones and by implication change the relationship between the direct and reverberant sound (in other words, move us the listener to whichever seat he wants us to be in) we at home have no such luxury of adjustment. The recording we have is fixed: the listening seat is, in effect determined for us according to the producer's commercial objective. The sound engineer merely gave the producer the sound he wanted, the sound that would hopefully sell. But, as we noted, the amount of 'hall sound' (the reverberation) relative to the direct sound can give us some sonic clues as to which seat the producer seems to have reserved for us.

We know that all good concert halls have a sweet, even and characterless reverberation time, possibly running to several seconds. We know that the closer we sit to the performers the less of that hall acoustic we hear. So we can estimate by careful listening to the relative amount of reverberation in the recording (relative to the direct sound) where we are seated in the hall. And once we know that, there is only one volume setting for that recording, with that amount of 'wetness'. Turn up the volume much beyond that point and we start to swim in reverberation - the reverberation becomes overpowering; we could never experience that amount of reverb even if we were sitting in the front row (very loud) because there, the loud direct sound would completely swamp the reverberation to inaudibility. Conversely turn the volume down too far and the reverberation disappears, and the sound dries out. But note: that again is not at all what we would hear in the hall. If we 'turned down the volume' by sitting in the back row, the entire balance would tip completely in favour of the reverberation and it would sound like listening in a bathroom. So, there is a correct loudness for every recording.

Makes sense?

There are two other very important spin-off points from this:

1) You only need a few watts of power to achieve this loudness level, which is probably far lower than you are used to
2) The speaker designer needs to know right at the start of the design process how loud his customers will listen.

Regarding point 2 - before you buy a hifi speaker be sure you know what loudness spec the designer was working to. If they have designed for very loud listening levels (as many proudly do) they will not sound warm enough when operated at a much lower level. If you take away nothing else from the HUG but this idea of target loudness you will make a better choice of loudspeaker. The answer to this point and the thinking that goes with it is perhaps the most important single issue in loudspeaker design and buying. If you are intending to listen at a long-term safe 85dBA at home, you want a speaker designed to sound natural at that replay level. You should disregard any designs which have a recording studio following as those engineers monitor at insanely high levels at which the loudspeakers perceived balance is completely different to a normal home listening level.

[A.S.]

I'm sure you did enjoy your FM broadcast, but I'd bet that, along with every other FM station, they used an automatic signal 'optimiser'. That works by measuring the audio signal level in various audio bands, then according to the desired 'station sound' that the station manager wants, compressing, cutting and boosting according to some predefined maths. These products are often called 'audio exciters'.

We already showed that type of multi-band optimiser in outline here. Here is a video of an earlier generation optimiser in action, in an FM station. Look at the second meter from the left (Limiting). When the needle drops downwards that shows the amount of limiting being applied. That's how they get that loud station sound without exceeding their legal maximum permitted loudness! Also here.

I do not know if their internet feed is taken before or after the optimiser, but possibly before. In which case, although the internet feed may sound thin and less punchy than the FM feed, it is in fact true to the source. If you like the more punch FM sound then you are not alone - optimising for FM creates a big, fat loud sound that listeners like.

I'm sad to say that here in the UK, there is no justification for buying (or even listening to) classical music on FM. DAB is good (enough, just) and BBC radio 3 over the internet is excellent.

[Drdennis]

Alan,

Unfortunately, the time difference puts the live broadcast at 1:00am in Great Britain. Thank you for the information. If I am able to obtain this information from the station, I will share it with the HUG.

BTW, the full address is, 99.5 allclassical.org

Dennis Girard

{Moderator's comment: Can you confirm the date of the broadcast? Are you able to make a digital recording of your FM analogue signal? Maybe we can record the internet feed.}

[STHLS5]

........

A) I *suspect that many 'audiophiles' listen to orchestral music at an average reproduced loudness higher than they would experience in the hall. ....
....... *
B) So*we can estimate by careful listening to the relative amount of reverberation in the recording (relative to the direct sound) where we are seated in the hall. And once we know that, there is only one volume setting for that recording, with that amount of 'wetness'. Turn up the volume much beyond that point and we start to swim in reverberation - ......... So, there is a correct loudness for every recording.

Makes sense? * * .... *

A) Yes, many of them do listen very loud. But wouldn't *listening much louder would also add more reverberation? One thing for sure, the bass boom would be more prominent. *

B) No. I am lost here. If I do not know anything about the recording then how am I to know the exact wetness that should be in the recording? For an example, in the Erroll Garner CD that I was listening yesterday, the piano tone sounded pure and accurate but little bit metallic and dry. Then on track 11 onwards the sound *was wet and smooth but seemed lacked "resolution". It is only then I read from the booklet of the said CD that it comprised of two albums recorded in two different studios one or two years apart. So what should be my reference volume on track 1 of a very unfamiliar musical instruments and recording? Shouldn't musical enjoyment be much more simpler than that?*

ST

[A.S.]

...B) No. I am lost here. If I do not know anything about the recording then how am I to know the exact wetness that should be in the recording?...

Well, any classical recording that comprises two recordings made months apart cannot really be considered a natural-sound high fidelity reference can it? There is not much that we can really deduce from such a composite performance.

I think you are saying you can hear in the recording the reverberation of the instruments (in the hall) but you are not sure how to interpret that 'wetness'. I assume that you have been in a concert hall and can recall the sonic difference between sitting nearer the front compared with (much) further back. If you can recall those experiences, I think you'd probably say that seated nearer the back of the hall the performers sounded 'smaller', quieter and the sound rather bathroomy. Seated nearer the front, you'd probably recall the sound as louder, more dynamic, and dryer - that is, you may hardly be aware that you are in a hall and even if you are aware, you'd probably be unable to say how big a hall. Whatever seat you are sitting in, the performers are playing the same notes at the same tempo. The point is that your brain is deconstructing what it hears in whatever seat you chose by decoding the relative amount of direct (completely dry) sound versus the same notes* swimming around the hall to arrive at certain deductions. And the primary one is to figure out how far away you are from the performers.

I would think that if you were blindfolded and led to an mystery seat in the hall, with the blindfold on as you listened you'll be able to deduce approximately where you were sitting. And you'd be doing that by evaluating the loudness of what you hear, and the wetness.

*It's important to mention causality. First the performer plays the note. That sound travels the shortest path from the instrument to your ear. That path will be reflection free, completely dry, because any and every other route from the instrument to your ear involving a reflection must have a longer journey time. When that direct, dry sound is registered in your ear/brain, a clock is started: call that Datum X. And when the ear detects the same note via a reflective and therefore longer path it stops the clock, at Datum Z. Your brain then calculates (Datum Z - Datum X = echo time) in seconds, and knowing from experience the sort of acoustic spaces that yields various echo times, sizes the hall. This process must be causal. The direct sound must arrive first and that precise note must arrive later via its circuitous route around the hall. If the echo arrived before the direct sound (if this were possible) causality would not be present (in this universe anyway) and we would be utterly confused as to distance, time and space.

How about that?

[A.S.]

OK, I was driving home and listening to the radio I heard two useful examples of recorded reverberation 'around' musical instruments which can give us some clues about the size of the acoustic space in which they were recorded.

First, an excerpt from Bolero. This is useful because the rhythmic, impulsive nature of the piece allows just enough time for the echo to diminish before the next beat comes along. In fact, the drum and brass blast energy into the hall with much the same intensity as a starting pistol would. Listen out for the echo of the drum and brass between notes. If you listen carefully on headphones you can hear that the echo is distinctly biased to the left, an indication of a large reflective surface somewhere left of stage perhaps?

Loading the player ...

IGNORE THIS MP3 LINK DUE TO BUG ..... HERE IS FLASH VERSION



Unfortunately there is a bug in our new MP3 player app that we're looking at, so I'll have to post the next two clips as separate postings and then edit them together when we've fixed the player bug. Continued ....

[A.S.]

Continued ....

If we slow down the Bolero we can hear the echo even more distinctly ....

Loading the player ...

IGNORE THIS MP3 LINK - Here is the FLASH version:

[A.S.]

And here is another example again from the radio driving home. Once again, it's the impulsive brass sound which excites the reverberation in the hall.

On this recording, listen to how the acoustic space (the reverberation wetness) and image width around the solo trumpet is completely different to that of the piano. It's a little unnatural but even so, I like the intimacy of the piano set against the distance of the trumpet*. The brass sounds as we would expect it to in a large space, far from us and exciting the acoustic. But the piano is very close to us and it occupies a much wider L-R spread. So this recording is sending us somewhat contradictory information about where we are in the hall relative to the performers.

What do you think of this? Where are we relative to these instruments? And how was this effect achieved?



* The sonic balance is chosen by the producer from a number of alternatives the sound engineers are capable of achieving with given performers, available time, microphones, music and hall. It is entirely at the whim of the producer to put us in a particular point in space relative to his artists. What sonic balance he may select for a commercial CD recording is driven by the primary requirement to sell CDs and at least cover the cost of the recording. He is not necessarily trying to achieve the same sound we would hear in the concert hall - if there was a hall. Conversely a BBC producer overseeing a live classical concert broadcast may well want to give us a familiar seat in the hall and has no eye on commercial success.

P.S. I am aware of some clicks on this recording.

[weaver]

The mental image I have of the trumpet and piano is something like an English town hall or possibly school hall recital: a shoe box shaped hall with a raised stage at one end, the trumpet up on stage playing directly to us and the piano down at floor level and somewhat closer to us - though perhaps tucked towards one side as I also find the piano slightly muffled in contrast to the trumpet.

If it were a more usual concert recital I would expect that with both performers up on stage together the balance would be different.

I find the relative volumes of the two instruments slightly disconcerting in relation to the 'wetness' of their sounds, but I think the sort of analysis that Alan talks about above is not something I have been fully conscious of before: the trumpet being louder but further away, the piano both closer and quieter (allowing for the fact that it is being played more quietly as well).

[Note: in my work set-up I don't really get any sense of left-right information due to room layout]

[HUG-1]

... the trumpet being louder but further away, the piano both closer and quieter (allowing for the fact that it is being played more quietly as well).

What was it about the piano and trumpet sound that positioned them nearer you or further away from you?

[weaver]

It was the space around the trumpet that put it at a distance - my very first reaction was of an off-stage sound from a lofty point*, but the relative loudness seemed to contradict that. It would probably be more accurate to say that I thought the trumpet was in an open indoor space - one which allowed the notes to build and hang around in the air - which I guess means that the reflections were suitably separate from the direct sound; had it been distant but close to hard surfaces then the sound would have been different.

Conversely the piano sounds as if it is being played rather softly - at the very beginning I thought it could in fact be a fortepiano. I don't know enough about listening critically to this type of thing to say whether, by playing the piano softly it would excite less reverberation or not, but the space around it seems different to the trumpet; to have less openness, less brilliance. At the same time it doesn't (to me) sound up-close in a reach-out-and-touch-it way as A.S. said

I like the intimacy of the piano

but though I don't actually disagree, I also don't find that there is a particularly tangible 'closeness' to the sound - that may just be down to a definition of terms though.

There do seem to be two different aspects going on though:

The size of the space and the distance from the audience

*you get them in eg the Verdi Requiem, I heard it in an amateur performance at, I think, Ealing Town Hall where it was up on the balcony.

[A.S.]

It was the space around the trumpet that put it at a distance ... Conversely the piano sounds as if it is being played rather softly ... by playing the piano softly it would excite less reverberation ... but the space around it seems different to the trumpet; to have less openness, less brilliance. At the same time it doesn't (to me) sound up-close in a reach-out-and-touch-it way as A.S. said...

Several points here. First, regarding the relative loudness. It's important to appreciate that the way the trumpet shoots sound out into the hall compared with way the piano projects itself is completely different. In fact, it's possible (even probable) that the piano is capable of generating a much higher loudness, but that balloon of sound will radiate (roughly) in 360 degrees around the instrument (assume lid closed). The trumpet bunches up all its energy into an intense beam of sound, exactly as brass instruments were invented to do. So, I don't think the relative loudness of these two instruments can reliably tell you much*. What is much more revealing (to me) is the amount of reverb around them. That's the real tell-tale about their relative positions.

Your comments about the piano surprise me a little. The acoustic space around the piano is totally different to that around the trumpet. As noted above, the piano is widely spread left-to-right where as the trumpet is quite clearly a point source placed at a specific point across the sound stage. There is no reverb or air around the piano plus it spreads across the entire L-R sound stage. Surly that can only mean one thing .... we must be standing right next to it, with the full width of strings running (approximately) left to right.

Are you with me on this point?

* and nothing at all if the trumpet is independently mic'd from the piano.

[STHLS5]

Just a quick note on Listening to echoes -1. I feel that I am seated more than 50 to 100 feet way. I am unable to listen more because it hangs half way with IE9, Firefox10, Chrome, Opera (Window7).

About listening to echoes-3, I agree with Weaver except the part about the distance of trumpet and piano. On the first instance, I thought the trumpet was closer than the piano but than the suggestive words of Mod and Weaver makes me to think the piano is closer but I just don't know.

It is obvious that I listen differently from Alan or Weaver because the moment I heard listening to echoes-3, my mind is fixed around the trumpet sound as the primary instrument and piano was secondary to it and received less attention from myself. I am unable to judge the distance nor I would have paid any attention to it. Strangely, I am very good with vocals that I can judge number of singers and distance but not so much with instruments. It looks like I accept colouration (reverberation) as part of a note though I understand reverberation is important for human survival to judge location and distance but somehow I am on a different wavelength here

ST

[weaver]

Are you with me on this point?

Yes, absolutely.

In live performance though, both the trumpet player and the pianist would be aware of this and tailor their performances accordingly wouldn't they.

If a supplementary question is allowed at this point: in a concert hall is the beam of sound vs the 360 degree sound significant in relation to reverb? eg would the piano sound have more early reflection points than the trumpet?

[A.S.]

If a supplementary question is allowed at this point: in a concert hall is the beam of sound vs the 360 degree sound significant in relation to reverb? eg would the piano sound have more early reflection points than the trumpet?

Good question. First, I picked this piano + trumpet piece with care to highlight just these sort of issues. Yes, in theory and sound source that radiated omidirectionally in a confined space (a hall) would presumably generate plenty of early reflections from obstacles in close proximity to it. If we are talking about a piano, that would include the stage floor around the piano, then the musicians and their stands and instruments near the piano. But we know that early reflections, by definition are from solid surfaces quite near to the source (the piano) and that their sound wave doesn't have much further to travel to our ear than the direct sound. So those early reflection would all fuse together in our brain and be completely inseparable from the direct piano sound. So it is not going to be possible for a listener to gauge the size of the hall from listening to the piano with its early reflections if either we are physically sitting next to the piano or the piano's microphones have, in effect, put us there.

The trumpet, conversely, has no (or let's assume that) early reflections. It sends out its beam of sound like a searchlight, and in the same way that a squash ball exits the point of impact with a wall at the same angle, the reflected trumpet sound beam will make its way to the microphone and arrive slightly after the direct sound. Eureka! If we subconsciously time the difference between the high-intensity direct sound beam and the reflection, we know how far away the walls are, and hence how big the space is. We can't gather any of that spatial awareness information in this recording from the piano. That's because it, unlike the trumpet, is 'dry', with little or no late reflections that would give us the clues we need to estimate the size of the recording venue. So then, the most impulsive instrument (the trumpet) has unambiguously revealed the size of the hall. On the Bolero clip I highlighted the drum as exciting the room echo - once again, a high energy impulsive blast of sound has produced a nice, clean, identifiable late reflection.

What next? Well, let's forget about level differences between the trumpet and the piano. By using multi-microphones those become irrelevant, and we could easily make the piano louder than the trumpet in the mix-down if we wanted to. The real issue I want to get to is that the trumpet has some 'air' around it (i.e. generates echoes) whereas the piano is completely dry and super-wide. What must that mean?