Adjusting Room sound using material damping methods (not DSP)

[EricW]

I don't know if I'm expressing this properly, but what I'm hearing sounds to me almost like a kind of "standing wave" effect, where the reverb is having the effect of reinforcing certain frequencies and perhaps also filtering out (or drowning out) other frequencies. On top of that, Clip 6 sounds almost (but not quite) as though it's on the verge of the kind of acoustic feedback you'd get through a PA system if the monitor levels were too high, and started to feed back through the microphones.

Does that make any sense?

{Moderator's comment: Bingo. That's the comment we were hoping for. Alan will now be able to advance the discussion tomorrow. Thanks.}

[honmanm]

All that's different is that I've just added a special sort of synthetic reverberation to the original dry clip. Has the PA system gone mad? There was no PA system, but in fact, this is not such a stupid suggestion as it may seem. Why?

Ouch, never mind an orchestra in your living room, that clip sounds like an orchestra in the bathroom (a large military bathroom, mind you). I guess what you mean by PA gone wrong, is the kind of effect where there is accoustic feedback in a large venue - like the shriek of feedback one sometimes hears when a PA is being set up, but lower down the scale and with a longer delay.

BTW Joe Jackson's "Body and Soul" is a rare example of a popular music recording made in a lively hall (the sleeve notes say: "You could say that this one started when... much displeasure was expressed at the contrived sound of too many modern recordings, not to mention the sterile atmosphere of too many modern recording studios").

{Moderator's comment: again, you are on the right track with your PA comment.}

[STHLS5]

........ OK, let's stop playing around and make the quality of the reverberation a more prominent feature of the clip. Yes, there is a lot of reverb (a large quantity) but that's not important. I hope that you can now clearly hear that there is a dominant quality issue with the reverb. It has a pitch to it. It has tonality. ...

We could have gone on forever without me ever mentioning the word PA or standing wave. ;)

OK, what is the dominant quality issue here with the reverb? What I hear in Clip 7 was a superimposed ringing tone that you get when you shout or talk inside an oil barrel (metal). In my opinion, Clip 7 reverberation influence affects the upper frequencies more than the lower, i.e not so much boom but more ringing.

ST

[A.S.]

OK, what is the dominant quality issue here with the reverb? What I hear in Clip 7 was a superimposed ringing tone that you get when you shout or talk inside an oil barrel (metal)...

Great - we're moving towards the point I've been aiming at from the start. Observers have commented that Clip 7 (and to a lesser extent, Clip 6) have a characteristic tonality in the reverberation. I'd like to call that a "twang". So the reverberation has a distinct twang. We're agreed on that now.

Previously I said this of Clip 7 with its oddly colored reverberation ...

But how has [the reverberation taken on this twang]? Are the musicians playing the same notes at the same loudness. Absolutely yes. Have additional musicians been drafted into the recording playing some weird instruments that have a drone in a narrow band of frequencies - some sort of bagpipes perhaps? Definitely not. All that's different is that I've just added a special sort of synthetic reverberation to the original dry clip.

Now the really big leap of imagination. Let's lay out the facts ...

• The recording is now colored because the reverberation has taken on a strange sonic character - a 'twang'
• That twang has some sort of tonality that you can whistle along with
• No additional musicians have been drafted in to the studio to play just that twang note
• There is no PA system 'gone mad'

So, where has this errant note or band of frequencies come from? How can a note be created out of the air?

[STHLS5]

.....So, where has this errant note or band of frequencies come from? How can a note be created out of the air?

They are the residual sound that continues to bounce of the enclosure (room) creating its own sonic signature?

ST

[A.S.]

They are the residual sound that continues to bounce of the enclosure (room) creating its own sonic signature?

So you are saying that a time domain problem becomes a frequency domain problem? That's rather magical isn't it? How an echo, a time issue involving nothing more or less than a bounced around and therefore delayed version of the original mixed back with the original somehow taking on a sonic character, with tonality from what is just a mirrored but delayed version of the original sound? Don't you think that is really remarkable? Can you see how that could be - and is - a potentially serious issue in the recording studio/hall and equally at home in the listening room?

N.B. We are not saying that all echoes (i.e. general reverberation) are necessarily bad. I've demonstrated that even, characterless reverberation is essential to creating a fresh, open sound. The point is that if the reverberation takes on an acoustic character (for whatever technical reason), the overall enjoyment of the music will diminish. It will sound colored. And that is what we are all here to avoid.

Now let's revisit that 1951 paper we started out with on this thread.

[A.S.]

...We are not saying that all echoes (i.e. general reverberation) are necessarily bad. I've demonstrated that even, characterless reverberation is essential to creating a fresh, open sound. The point is that if the reverberation takes on an acoustic character (for whatever technical reason), the overall enjoyment of the music will diminish. It will sound colored. Now let's revisit that 1951 paper ...

In the full paper attached to my earlier post, I highlighted Fig.2.. Attached to this post is page 1 and again Fig.2.. I'd say that Fig. 2 is one of the most useful and pertinent graphs in the entire subject of listening room (or studio) acoustics. If we combine what that shows us with a graph of absorption coefficients (and we looked at that several posts back) we have the tools to attack the worst listening room problems without resorting to complex maths or needless expense. Fig. 2 is the one to remember!

Fig.1 shows an idealised situation, the reverberation level decaying with time. The vertical axis is loudness, and at the back of the graph we see 100% (=0dB) loudness and the Y axis (at a slant) is time advancing towards us. So at zero time, the loudness is 100%. Imagine that as someone firing a starting pistol at the 0,0 coordinate and then the microphone collecting all the sound bouncing around in the room for seconds*. You can see that after one second has passed, the sound of the gun shot has diminished to -60dB or one thousandth. Which must imply that the surfaces of the room have soaked up 60dB of sound. That would be a very long, obvious reverberation, probably about the length you would find in a church or concert hall. The shape or gradient of the decay (in Fig. 1 it's a ruler-straight line) also gives us useful information. In a real church there would be many small rooms and chambers and as the gun-shot permeated through the building and its many rooms the signal received at the microphone would not have such a straight line. Sound travels relatively slowly and the echo arrival time back at the microphone depends entirely on how far it has travelled, in the range of fractions of seconds to full seconds, so as the sound wave probes into the deepest recesses of the cathedral, there will be a series of echoes picked-up. First from the seats in the main hall, then the choir area, then the walls, then high ceiling then all those innumerable rooms and chambers further and further away. And then there would be the echoes of the echoes. And the echoes of the echoes of the echoes and so on until all the acoustic energy from the big bang at 0,0 is dissipated.

Fig. 2 shows us a much more realistic situation. On the right side I have again marked those conical peaks called flutter echoes. Whilst the average decay gradient of the room is still a straight line (as we would expect for any moderately damped acoustic space) this space suffers from a serious echo problem. At regular intervals as time passes, due to some sort of dimensionally-related path lengths, the echoes take the form of a series of pulses. In essence, some time after the original single gun shot is fired, what we hear in the room sounds like a machine gun, and that repetition takes on a pitch. I'll make a clip to demonstrate that later.

(*I've put my glasses on and can read the graph properly now! The timescale is 1.0 secs. not 10 secs.)
>

[Pluto]

Have you considered the differing reverberation needs of various musical genres?

Here's a clue: Antonio Vivaldi wrote most of his music for performance in big, cold, stony Italian renaissance churches. Ditto, J.S. Bach (a few hundred miles to the north but basically, big church music).

Mozart was essentially a court musician, writing for his employers (and the size of the musical forces they were prepared to fund), typically middle European potentates who would entertain courtiers and rivals in the grandest rooms of their fine, beautifully furnished palaces. Even Mozart's best known operas were "built" as relatively small scale music compared, say, to Verdi & Wagner's creations a century later.

The point in these two examples is that different genres tend to require radically different environments, and the great composers are instinctively adept at making their music deliver under the conditions in which it had to work - be they venue, size of band, availability of particular instruments and strong soloists (or otherwise).

With these considerations in mind, what do you think the ideal acoustic characteristics of Overture "Marriage of Figaro" ought to be?

For my own part, the dry example was nostalgically reminiscent of several hours I spent in the orchestra pit of a London opera house during rehearsals, but it is most certainly not the sound I would expect as a punter sitting in the audience.

[A.S.]

At regular intervals as time passes, due to some sort of dimensionally-related path lengths, the echoes take the form of a series of pulses. In essence, some time after the original single gun shot is fired, what we hear in the room sounds like a machine gun, and that repetition takes on a pitch. I'll make a clip to demonstrate that later.

I've made a synthetic digital gunshot - this is approximately what you would hear if you fired the starting pistol outside in a field ie. the direct sound, no reverberation. Four 'shots'. Actually, this is really a digital click, but it is acoustically dry which is the point I want to make. You may need to turn up your replay volume.

(Clip #9) Dry gunshot

Clip #9 above is completely dry. No hint of any tonal character in the impulse. Now if we take the starting pistol inside and fire it four times again in an untreated room we will hear something like this (source file is Clip #9):

(Clip #10) Gunshot in large untreated hall

Now a strange thing has happened: we can hear that the reverberation has a definite tonal character, a 'twang'. Suppose we make the gun fire more rapidly, so that there is less time between the pulses. (Clip to follow)

[STHLS5]

....
Fig. 2 shows us a much more realistic situation. On the right side I have again marked those conical peaks called flutter echoes. Whilst the average decay gradient of the room is still a straight line (as we would expect for any moderately damped acoustic space) this space suffers from a serious echo problem. .....>

Your explanation about the effects of reverbertion refers to flutter echoes which was confined to higher frequencies as shown in Fig-2. Why wasn't bass boom discussed? I believe there are significant low frequencies in gun shots.

Meanwhile, there are some audio samples (by Concert Hall Acoustics) recorded at different reverbertion time which may help our readers to understand how, for an example, speech finally becomes unintelligible. Don't miss out the samples of cello solo played on stage as heard at four different seats at the Amsterdam Concertgebouw.

ST

[A.S.]

Please understand I just do not have the time or the need or desire to open the subject too wide. We are still not at the conclusion of the journey. What point are you making with these clips?

I'd rather you concentrate of the character of the reverberation, not what different degrees of reverberation sound like.

[STHLS5]

Sorry Alan. Audio is not working in clip9 and 10. It is very faint, like a click sound.

[EricW]

I'm not sure exactly what the question is here but one observation I have on Clip 10 is that it's possible to hear the reverb almost as a separate sonic event from original impulse. It has a character of its own. The two together (impulse plus reverb) almost sound like a simple synthesized version of a snare drum.

[A.S.]

... One observation I have on Clip 10 is that it's possible to hear the reverb almost as a separate sonic event from original impulse. It has a character of its own....

I'm not surprised; that's due to the Haas effect which is another slightly off-topic discussion for later.

I think I've made my point (or I sincerely hope that I have because it's supremely critical) that a series of reflections - we call them echoes -can take-on a tonality. In the case of the previous classical music clips, we could hear in Clip 7 a sort of metallic coloration. We could have concluded (and it may or may not necessarily be true) that the reverberation's tonality was somehow related to the pitch and notes in the music, so that if we change the music, the tonality of the reverberation would change with it. But from Clip 9 and 10, the gunshots, I wanted to show that even in the absence of any musical pitch at all in the dry gunshot, fire that gun inside a room and magically a pitch appears in the reverberation. Not in the direct gunshot sound - in the reverberation. In the room the reverberation now has a distinct twang or 'glongk' sound which you can mimic with your mouth. Try it.

Without deviating from this line, is this point completely recognised regardless at this stage of an understanding of how and why echoes in a reflective room somehow mysteriously develop a pitch? And yes, change any detail of the room's dimensions or acoustic absorption and the character of the reverberation most likely would correspondingly change with it.

[STHLS5]

...Without deviating from this line, is this point completely recognised regardless at this stage of an understanding of how and why echoes in a reflective room somehow mysteriously develop a pitch? ...

Yes. It is like two different events - one just the gun shot sound and in another the gun shot sound with a separate instrument playing "twang". But sorry! I did not hear them.

ST

[A.S.]

...and in another [Clip #10] the gun shot sound with a separate instrument playing "twang". But sorry! I did not hear them.ST

I don't want to move forward unless I can demonstrate the 'twang' to you, that is, the tonality in the reverberation which is definitely not present in the dry gunshot clip.

I recommend headphones to listen to any of these clips. Let me increase the loudness. BEWARE! These sounds could damage your speakers, headphones or hearing. The gunshots were previously at about 80% of full scale: now to 100%, maximum loudness.

(Clip #9A) Dry gunshot, gain increased to 100%

Clip #9A above is completely dry. No hint of any tonal character in the impulse. Now if we take the starting pistol inside and fire it four times again in an untreated room we will hear something like this (source file is Clip #9):

(Clip #10A) Gunshot in large untreated hall, gain of gunshot increased to 100%

Now if we remove the direct gunshot from the in-room recording we are left with only the reverberation. And if we turn-up the loudness of the reverberation by about 40dB (100x) the reverberation clearly has a tonality. The reverberation could be 'played' as musical instrument here:

(Clip #11) Gunshot in large untreated hall, gain of gunshot increased to 100%

The essential point is that the dry gunshot had no tonailty, but fire it in the room and now we have a clear pitch. Interesting?

[A.S.]

The starting pistol when fired in the room has (I hope you can hear now) now taken on a sonic signature of the room, a definite pitch or tonality. In fact, there are multiple frequency components in the reverberation. We can isolate them with very sharp narrow-band filters.

Again to remind us:

(Clip #11) Gunshot in large untreated hall, gain of gunshot increased to 100%

Here is one strong frequency component isolated from within the reverberation of Clip 11, via a filter between 500-700Hz, all other frequencies are rejected:

(Clip #12) Filtered echo component of gunshot in room (gain increased)

There are multiple frequences in the reverberation, but that band is particularly strong. Remember: these energetic tones were not at all audible in the dry gunshot. They have been 'generated' by the room itself.

Next step: consider this. We've shown that if we take a gun with a clean, characterless click or crack sound when used outside and we take it inside and fire it, now we hear tones in the gunshot's reverberation tail due to something strange about the acoustics of the room and nothing whatever to do with the gun. The gun is a sonic generator. So is a loudspeaker. So if we take a loudspeaker with a pure, clean characterless sound and place it in the ordinary listening room and start it playing, what we hear will also be polluted by the room. The direct sound from the speaker to our ear sitting in the sweet spot will be the same clean sound. But immediately following that will be the reverberation of the room. What we hear at the sweet spot is therefore a combination of the speaker's clean direct (anechoic) sound + the twang from the room. Regardless of where we fire the gun in the room and regardless of where we place the speakers(s) in the room, the room is going to add its twangy contribution which will spoil out listening pleasure.

The proof of the above would be that if we reproduced the dry gunshot over the speaker(s) in the room, assuming that the speakers were basically of good quality, the speaker-gunshot's reverberation would exhibit exactly the same twangy character as the real gunshot.

Are we ready to move forward to the next step now? Tell me if not please.

[weaver]

Firstly, this is a fascinating thread which I am just catching up with and which has already started to explain something that has puzzled me for years. Thank you.

Secondly - this may well be a pedantic point but can I ask anyway? The gunshot is described as a "clean, characterless click or crack" but that does not mean it is without pitch does it? I realise that 'pitch' is a function of human perception but the sound will nonetheless have a frequency won't it?

I know that this is not the point that was being made but the reason for asking is that I actually found clip 12 rather startling in the sense of being surprised that that component was present in the total echo sound, ie had it not been isolated I would not have 'heard' it. I do take the point that it was not present in the original gunshot.

My current way of thinking about this is that, visually, the characterless sound is grey - an equal balance of colours. The 'tonal character' comes about when we move away from grey.

[STHLS5]

....
Here is one strong frequency component isolated from within the reverberation of Clip 11, via a filter between 500-700Hz, all other frequencies are rejected

Re: Clip # 12 Amazing! How reverberation can generate a different pitch. I am still trying to gather from clip # 11 the influence of 500 -700Hz.

OK, now everything is loud and clear. Thanks

ST

p.s. My audio issues related to poor speakers quality. Now rehearing in a proper speakers all the clips make sense. Including Clip #1.

[EricW]

Entirely clear so far.