Ohm's acoustic law and stuff

Mr Insest, is fact the one and only Datty 19, he just morphs now and again on different forums to stamp his own brand of 'Numbers & reference' book' ideals on the poor unsuspecting public about to be subjected to blind test terror.
He is in fact highly amusing in a limited way, in fact after being subjected to several old copies of 'The Flat responce' this weekend. I can see where the arragance of the dungoens and dragons forum comes from. Along with their absolution/aversion to change that doesn't involve an lp12 or briks
Still lets get it on Round 3, both fighters have retired to there corners, parent molestor is a jab & run merchant, who punches initialy sting but have little or no substance (unless the net is taking longer than usual ), The wizard of time domain errors, is more an upfront, pugilist favouring the ''ere you ugly spud, cop some this" (insert favourite insult here).
Shaping up to be a classic this bout, more later from your Master of ringside ceremonies on what promises to be a 'corker'

 

I will be more that happy to post my answers and elaborate on why I think that way, once you have posted your answers. We can then compare notes and see how much at odds we really are.

Well, your posts have a very poor signal to noise ratio, but from what I recall you have a very high opinion of the Tact RCS and also the Velodyne DD series subs. Interestingly, both of these products have sophisticated equalization capability and more importantly a model of what is objectively the right way a system should sound. Both systems promote the idea that a flat frequency response is a good thing. Both of these products rely on measurement (the tricky stuff we do in science) to correct the systems response.

That you like systems that rely on measurement suggests that you don't "trust your ears". Or alternatively, that your "subjective experience" can improved by science, and more importantly, that your "experiences" can be explained by science. That you can't provide an appropriate scientific explanation comes as no surprise given your evasion of some simple questions.

Why do you feel the urge to play "pretend scientist" by asking questions including phrases like "time alignment", when you can't understand the answer (which I've already given in post #39 )?

 

In order to stop you being somewhat childish ( although you have stopped short of shooting up your mother ), time alignment AFAIK is the process of ensuring that the overlapping wave fronts of your subwoofer and satellites are in phase.

Now it is also my understanding that unless your sub is positioned between your satellites, it is very difficult to match phase as the problems are different depending on the frequency. Let's assume for sake of arguement we don't have sophisticated EQ and RTA facilities as I am sure most people are not eqipped with them.

So we have a problem. Most subs do not have continually variable phase. They do not allow for more than adjustment of the crossover frequency's phase AFAIK. Consequently they are very hard to integrate without careful positioning. The hard fact is that the majority of commercial subwoofers offer no alternative for correct alignment other than to experiment with positioning.

Maybe you could shed some more light on the subject?

FWIW, undoubtedly subjectively the best bass I have heard is from my JBL subs, placed inside the satellites in the same plane, and without any form of digital processing or EQ. Note I said subjectively, i.e. the ability to enjoy the music rather than the reproduction. So no use of science applied there I'm afraid Datty

 

So, here are the answers to my questions:

Close to a wall is good, and in my experience, corners are the best. In this position, the sub does a good job of exciting the room modes - I'll come back to why this is a good thing. 2 subs is a more interesting configuration which I'll return to at a later date.

Well, the main speakers are usually away from the wall, and out from the corners, so the net result is that they are closer to the listening position than the subs.

So the signal to the main speakers gets delayed (because they are nearer) and the delay is implemented digitally.

And when I was faced with that question I concluded that, yes, an expensive analog source was a waste of money, because its signal was going to be digitized (for delay) on the way to the main speakers.

It's probably best to dispense with "time alignment" and think of the delay and phase dials as a way to get the amplitude response right. You might deliberately have the sat/sub signals arriving somewhat out of sync(/delay/phase) in order to get a flat(ter) response.

Now, getting back to Mr Ohm and his "phase"/acoustic law, that discussion was about the inaudibilty of the different relative delays of different frequencies (in this case, the bass being "behind" everything else) even if the amplitude response is flat... So, getting all the frequencies "in phase" isn't a big deal.

Even if subs had continously variable phase, the phase cannot "advance" the signal to catch up with that from the sats (even the mighty velodyne DD series can't do this) so there will always be a time lag of the bass behind the satellites. Subs don't need much in the way of delay (phase) control because the delay is expected to be supplied by the processor to the main speakers. (Because of the snafu over bass management in SACD and DVDA, it's been pretty difficult to make those formats work in a previously installed system, until the players started doing cross-over/delay..)

To be fair to the subwoofer manufacturers, they expect the delay function to be implemented elsewhere.

Even with the right tools, subs are not the easiest things to make work properly, I suspect many people have given up in frustration and "declared subs don't work"; moreover adding a sub to an analog system seems to require a substantial amount of luck, which explains why there's a great deal of variability in the experiences people report..

Sure, but they are giant subs They would happily drive the room from wherever you put them. However, they'd work less hard in a corner where they'd couple to the room properly and you'd get some "free" gain and that's important for smaller subs that are being expected to work miracles, and even for big subs at high volumes, you'll get lower distortion if the sub isn't working so hard.

Returning to the point about exciting the room modes; with EQ, you can have your cake and eat it: you can put the sub in the corner, and in most rooms excite all the room modes, have the fewest valleys, and get the most volume for the least cone excursion. Then you can flatten the peaks with EQ.

The other nice touch is that if you've got small sat speakers you can considerably ease their load (and hence lower their distortion) by crossing them over further up. And as the sats are putting out less bass, they are less likely to cause room boom resulting in more freedom in placement. Of course to do this, you have to go down the AV processor/Tact/behringer route and give up on analog sources - unless that source is one of the new breed of DVDA/SACD players with bass management rolled in.

Right, I'm off to "experience" my actual (not hypothetical ) DD15...

 

Oedipus,

Your observations over the slight time delay on the sub is interesting and one that I have already been thinking about in my own setup. As I dont have any processing on the satellites, there must be a slight delay due to positioning and the latency in the DD12 signal processing.

Subjectively however, it is absolutely unnoticeable, and the benefits in sound quality that the sub brings are huge. Therefore I am not too worried about it.

Interestingly, WM tells me that he used to use a sat/sub system where he implemented a time delay on the satellites and fired them 5ms behind the sub. Looks like you are behind the curve somewhat

I'd be interested to hear your experiences with the DD15.

 

Well there's a surprise.

Those answers were exactly what I have given before on many occasions, but was unhappy to repeat for fear of being told it was wrong, given your assertion that you could not hear differences in the time domain.

Why you did not answer the orginal question I do not know, but I'm sure you have your reasons. Looks like you've been doing a lot of reading of Tact's promotional material Oedipus

Sadly, many of the posts from earlier days on the forum seem to have been lost which is a shame. There was however an interesting discussion aboyt Titian's Nestarovic setup. This employs two subwoofers located 50cm from the rear wall, with the satellites some four feet into the room. Being entirely analogue, there is no provision for delay to the satellites. And yet those who have heard it say it is extraordinary.

Might I suggest you play with the distance settings on the Tact in the interests of research? Come back and let us know what you think of the subjective effects of reducing the delay to the satellites. It would be interesting to know if they correlate with my own findings from way back.

Neil,

I really would not worry too much. In you case, if you pull the Velodyne away from the rear wall, the back wave will likely be out of phase and cause cancellation which will be hard to EQ. This will certainly be the case if you end up crossing over at 80hz or above.

 

merlin - all the threads that were ever started on ZeroGain (as opposed to GrooveHandle) should still be available. Here's one thread about titian's system:

https://www.audio-forums.com/as-rediect/showthread.php?t=4232

Michael.

 

Michael, that's not the one unfortunately.

I don't seem to be able to reference any threads at all from last year

Ah got it now Michael, it seems to limit the search to 500.

Here's the thread, back in the days when I could afford a decent hifi

http://zerogain.com/forum/showthread.php?t=367

 

Quite a lot of your experiences match my own. The root cause of the argument here is that the explanation of your experience is incorrect. In this case getting wrapped up in phase/time rather than frequency response; if you mess with the relative delay of the sub sat, you cause a difference interference pattern at the listening position, and those changes show up in the frequency response. It's the change in frequency response that explains the differences you heard when fiddling with the time delay, you're not hearing the delay itself (Mr Ohm and his law again)...

I'm not sure which question you're refering too - was this: why do dac's sound different?

As for me reading Tact's promotional literature, well, that's not how I got started on measurement for speaker placement. I was doing that years ago. What really started the ball rolling for me was an article by Floyd Toole in The Audio Critic. [The Audio Critic is an audio review magazine which is the complete antithesis of Stereophile.] Anyway, I read the rest of Harman's whitepapers and got started on EQ. It's how I came to get my hands on the Behringer gear that I have, the Tact is something of a recent arrival..

The Tact is an odd bag of tricks: it's not really clear that fiddling with some of the controls results in changes to the calculated correction "algorithm". Do you know a way of using the Tact itself to measure the corrected room response? [And here I mean engaging a previously created correction and running the measurements to confirm that the tact is "working".]

My experience is that you can dial in a target curve, but the Tact doesn't necessarily generate a matching correction algorithm (it uses its own intiative so to speak). I know what I've asked for (with the software) and, by using another room measurement tool, I get to see what its actually doing - and they don't match.

 

That one puzzled me too. The only method of performing the check is to use a second Tact AFAIK. Tact's explanation for this is that the algorythmn used was deveoped following a great deal of research not only into room acoustics (remember Peter & Boz were heavily involved in the Eureka Archemedes project), but also taking on board corresponding research into psycho acoustics. So in effect, it is measuring the room in the normal fashion and then correcting for what the software writer believes you perceive.

Now whether that's hokus I have no idea. I have not studied psychoacoustics but it is the reason I took issue with your assertion that the Behringer did the same thing ages ago and it appears that they are performing differently. You can be the judge. The results are impressive IMO, so maybe it's more than marketing hype. I believe Boz's field of expertise was psychoacoustics and mathematics. There is an alternative algorythmn available and a new version of the original in the pipe line.

 

FWIW you've never actually addressed my concerns regarding arrival time sensitivity... Specifically this:

Now highly likely I'm being dumb in some way, but I've yet to see a counter argument.

 

That was my thinking as well Martin. From my experience we are susceptable to time smear, where an omni directional bass note is spread over time by the delayed arrival of the back wave..

Given that state of affairs, it would seem logical that time smear at the crossover frequency was also undesirable. Even if you EQ perfectly, surely without the application of a delay or careful alignment of the drive units will result in a band of notes that are artificially long in duration at the listening position. The test of altering the delay to the satellites subjectively suggests that result.

I'd also be interested to know from Oedipus how this can be compensated for by adjusting the frequency response.

 

Your conclusion does not follow from the premise; ITD is "phase-difference" and it does not follow that the ear uses some part of a cycle, after all, over the duration of a cycle the phase difference is constant.

If we take a sound, it's a collection of frequencies; each frequency has a phase; the brain/ears may use the interaural time delay (phase difference) of a frequency in spatial location; it may use several phase differences, each at different frequencies. Given the phase difference and frequency information, there's no need to cross-correlate the relative phases of multiple frequencies for localization.

If, for whatever reason, the phases of the frequencies are shifted around in the hifi, it leaves the phase difference of each frequency between the ears the same. [In more technical terms the interaural delay is independent the signal group delay.]

Moreover, the ear/brain is not cross-correlating the phase information of multiple frequencies to determine the "character" of a collection of frequencies. (ie is it a violin or a piano etc). So, if you mess with the group delay of the signal, the ear hears the same thing, this is Ohm's law.

 

God this time difference is a pain

Here's another one to try. Whilst playing a solo vocal recording and using the Tact, adjust the level of one of the speakers only, upwards in 1db increments. This is effectively a traditional balance control so anyone can do it.

Then reset the amplitude and adjust the delay to the same speaker instead. Hopefuly you will note how much more this affects the perception of a solid central image.

As you say, we are using interaural time delay information to assess the spatial information. This relies on the sound waves from each channel having an identical arrival time for true accuracy, so arrival time is essential for correct image reconstruction.

I'm sure you are aware that the ear is less sensitive to these changes at lower frequencies, hence the ability to place a single subwoofer in the corner of a room.

But IME bass possesses a measure of spatial information that in a typical living room is disguised by the the varying arrival times of the boundary reflections. Theoretically, if you can take them away, the bottom end becomes locatable, although with less ease than at higher frequencies.

Martin's concerns about arrival times are IMO well founded and you don't seem to have adressed them as yet. I am not sure if we are all confusing the terms "phase" and "time alignment" here. Might I suggest we discuss arrival times (both reflected and direct if you prefer). Are you trying to tell us that you can correct for any errors in this field, including time smear, simply by adjusting the frequency response?

 

Here's my thinking. The maximum interaural delay is going to be that of the time it takes sound to propagate the width of the head in air. Let's say we have a 20cm diameter head, then taking the speed of sound in air as 344m/s this corresponds to ~0.6ms time delay. A single cycle of a 100Hz wave lasts for 10ms in comparison. So in order to calculate the ITD at 100Hz the first ear needs to detect a particular part of the waveform, and then wait for this same part to reach the other ear*. As you say above it doesn't matter which bit of the waveform, but the crucial point is that it is information on a time scale of less than a single cycle duration.

*This I assume must be where I'm making a mistake in this assumption, but I still can't see what it is I'm afraid .

That I understand (provided the hifi doesn't introduce a phase difference between channels I guess).

What is really the issue is the extent to which we are sensitive to arrival times, since phase delays introduced will make these frequency dependent. My point above about apparent sub cycle sensitivity leads me to believe we could be sensitive to these differences (not that they would confuse spatial info).

Merlin's point about sub placement and delayed back wall reflection must surely be true though, Ohm's acoustic law or not. Move a sub out from a back wall and surely the total length of time that the note lasts will increase slightly, and this will mean that the time varying power spectrum that this law says we are senstive to will have signal for slightly longer. This must give rise to some sort of time smear surely.

 

ITD is subject to "phase ambiguity" (where the arrival time differences are greater than 1 cycle, where people fail to correctly localize on ITD alone (using simple tones). Therefore ITD isn't using the "absolute time of arrival" in calculating the phase difference. Luckily, ITD isn't the only mechanism used by the ear for localization, the other being the amplitude response as modified by the head (transfer function.) [It's worth knowing that any signal processed by the ear is broken down into component frequency "bands", so that the particular "shape" of the waveform isn't matched between the ears for ITD either.]

If we were sensitive to the arrival time of different frequencies, then Ohm would be wrong.

Yes, and that is why rooms cause so many problems.

 

Oh, I don't know, it gives me time to think. You ask some fairly hard questions

I'll try that later.

No, that's MartinC's claim. Earlier I proposed that the amplitude response and diffraction around the head were what is used.

Discussing ITD is all well and good for real sounds.

However, things are more tricky when we consider what is happening in stereo systems. There isn't a single source of sound generating the image, instead there are two speakers each sending out a signal. The signal from each speaker is heard by both ears, with a phase difference! The ear is obviously missing "something" because it isn't telling us "oh no this isn't a centrally located image, it's two images coming from over there and over there.."

We could go into crosstalk cancellation here, which makes quite a difference, but AFAIK it's not something the Tact does. (My Lexicon does it though).

I don't beleive that last part to be true.

The "problem" I've had with locating subs is that the pressure wave that they send out shakes the floor or the wall and in the process generates a lot of higher frequencies which draw attention to where the sub is..

What we have done is converted what might rightly be called delay into phase - this is fair game some of the time, but can also be confusing when the context isn't clear.

However, now we are stepping out of frequency and phase relationships (Ohm again!) in electronics and speakers (and how the time alignment can be improved with anechoic measurements) and we are heading into the area of room acoustics, which make these other problems seem like childs play

Let me rephrase that as "Can we fix room acoustics by simply flattening the frequency response?" Yes, but only the bass. It turns out that at the lowest frequencies in the room are minimum phase (which I believe was pointed out by Gerzon of Meridian and verified by a bunch of people) which means we can EQ it flat and simultaneously tidy up the time domain. For everything above 400Hz (or so) opinions on whether correction should be done, and how it should be done, vary.

 

Oedipus,

surely if the phase and amplitude is identical fron the two speakers, as well as the frequency response, the brain is able to position that sound centrally. Should any of those parameters be skewed at the listening position, the solidity of that image is reduced. That is why the arrival time is so important IMHO.

With regards to bass, my point is not aimed at VLF produced by subwoofers and movie effects, more at real music. It is feasible IMO that we are able to locate say a bass drum in three dimensions by hearing it's harmonics. These harmonics are usually buried by the time smear caused to the fundemental by the listening room and positioning. Remove this and the spatial clues become obvious. In other words it is the listening room that clouds the issue in most systems.

For me, the importance of preserving transient attack is fundemental to correctly portraying music as opposed to test tones, as music is effectively a series of transients. Surely time smear will have anegative impact on the initial transient? In which case, how can we ignore it?

 

But the arrival time is "messed up" because what comes from each speaker is heard by both ears. So, the right ear hears the right speaker and the left speaker, and the left speaker path is slightly longer. This smears the signal in itself.

The impulse response for many speakers on the market shows that the speaker doesn't launch all the components of the signal simulataneously, however these speakers are still acceptable and "image".

Well, luckily, the initial wavefront from the speaker typically arrives a couple of milliseconds before any reverberation..

Spatial cues come from a broad range of frequencies, mostly in the higher frequencies. On the other hand, the room tends to make itself known in the lower frequencies - by supporting the room modes which show up as the peaks and troughs in a frequency sweep. The HF shows a much smoother character in a sweep.

Well, if you can't get the test tones right, you've got no chance with a transient. An impulse is essentially every test tone played simultaneously for a short period of time

Some folks don't: they've bought into the importance of the impulse response and go with single driver speakers or 2 way designs with the simplest crossover; by going this route, they don't screw up the phase response of the speaker..

Folks like me, stick with Ohm, say that the phase response is irrelevant, opt for steep crossovers and generally have a flat(ter) amplitude response, beleiving that to be more important.

I'd wager your d100's have a less than perfect impulse response (just like my ATC's); I can't seem to find any useful details abou the crossover alignment on the proac site..

 

Sorry guys, I've been mad busy at work the last week and so haven't had chance to check ZG. Not sure whether to let this thread die, but I'm still interested so let me respond a little...

Hmmm... not with you here. I was specifically restricting what I was saying to low frequencies where the time differences are less than 1 cycle, and therefore a frequency range with no phase ambiguity as far as I can see. I appreciate over most of the frequency range IID is used for spatial info.

That point seems much more relevant to me. The question I guess I still need to know the answer to, in regards to this and arrival time sensitivity, is the time resolution of hearing. I mean that banded frequency info must be that corresponding to a particular time window, but how long is this?

Let's consider a tone burst signal from a single speaker. Before the sound reaches us, we here nothing, and when it does we hear it. That is arrival time sensitivity of a sort. Now Ohm's law says we are sensitive to the power spectrum only. That's fine, before the signal gets to us the power spectrum is just noise, and once the tone reaches us, it gives us the power of the tone. But the exact point at which the power spectrum gets signal will be dependent on the exact arrival time of the tone. What may well be the case though I guess is that the time resolution of hearing is such that this difference is irrelevant?

Are you sure about that oedipus? Like Merlin I did think that the only thing that placed a low frequency limit on the ability of us to localise sound by ITD was room reflection issues. I've certainly read as much, but only on the web rather than reliable text books or papers I'll admit. As you said above you'll have a phase ambiguity issue for higher frequencies (plus a better way of doing it via IID), but what would place a low frequency limit on using it?

Out of interest, does anyone know a good text book that would cover the sort of issues raised in this thread?