Acoustic Measurements: Understanding Time and Frequency

[microstrip]

Let me put it this way. I say you are hearing the reduced amplitude of the peak that is visible in the SPL graph at 56 Hz or whatever it was. The fact that its ringing also went down was a consequence of the same thing so while seeing it in the waterfall may be reassuring, it didn't provide more insight there. As to 34 Hz reduction in time domain only, I remain dubious that anything good was done since the amplitude of the resonance is there. And if there is one thing we know is that we absolutely hear such frequency response variations.

Yes. I have not seen your displays for short and long time window. But in my testing there is a huge difference in the tail (see my examples earlier). BTW, in one of the waterfall displays in Dr. Toole's book he is using 40 msec. I don't recall now which one it was but did notice the short time. BTW, to see the difference I am talking about, you need to calibrate the noise floor. If that is set wrong, then it might look like it is not doing what I said.

Amir,

Thanks for your patience.
The 56 Hz attenuation is not directly due to the tuned bass trap - curiously it is an interaction between the bass trap and the abbfusors placed behind them. If I mode the bass traps, the 56 Hz level stays in place, the decay at 37 Hz still reduces and the big subjective improvements remain.

I also have in my room two Martin Logan Descent subwoofers, that have some equalization facilities at 25 Hz and 50Hz. Even changing the FR at these frequencies by a much larger amount than the referred values I can not get anything similar to the variation in sound quality of the bass due to the bass trap.

We should consider that my room is very long (9.35m), with heavy and solid thick stone walls and the 37 Hz is the second axial model. Perhaps an atypical case.

Just one additional question - how should we usually know if bass traps are needed? Just by simulation?

 

[JonFo]

Let me put it this way. I say you are hearing the reduced amplitude of the peak that is visible in the SPL graph at 56 Hz or whatever it was. The fact that its ringing also went down was a consequence of the same thing so while seeing it in the waterfall may be reassuring, it didn't provide more insight there.

Not sure I follow. Reduction in amplitude at a given frequency will of course show up in the waterfall of the same. However, insight may be gained by seeing the amplitude and length of the decay in the waterfall plot.
I use the waterfalls to help my isolate problem frequencies, determine the Q of the fix and then evaluate fixes (whether electronic or acoustical). Often the resonance is severe enough in terms of duration that a combination of approaches is required to address it. The frequency response alone is not enough (at least for me) to determine whether I'm improving things enough or not. Part of that is because of the windowing used in FR plots is much shorter (like half) than typically used in waterfalls, right?

 

[amirm]

Just take a look at my system's first page, where I have the response graph at the listening position. It's even smoother now than what's shown there - no EQ or even room treatments, just clever positioning; room treatments to come...

The system shown here was designed as an experiment to be as bad as it could be so that I could show the correction effect! I put the sub in the corner to excite all the modes and made the measurements. It is not meant to show anything other than that. Here is an optimized placement and optimization:

Note the faint lines showing pre-EQ. The EQ nicely smoothed the response. See the correction filters at the bottom. 10 db variations are now about 3 db.

The graph for your system seems to be smoothed since it doesn't show comb filtering at high frequencies. Do you have an unsmoothed or 1/24 smoothing graph?

 

[amirm]

Not sure I follow. Reduction in amplitude at a given frequency will of course show up in the waterfall of the same. However, insight may be gained by seeing the amplitude and length of the decay in the waterfall plot.

Sorry. I should have been more specific. I meant that the correction shows up in the time axis (Z) of waterfall. Since that is the only value of waterfall over SPL graph, I said that it is not necessary in this case other than if one wants to understand more what is going on. For minimum phase systems, that correction mathematically follows the amplitude reduction.

I use the waterfalls to help my isolate problem frequencies, determine the Q of the fix and then evaluate fixes (whether electronic or acoustical). Often the resonance is severe enough in terms of duration that a combination of approaches is required to address it. The frequency response alone is not enough (at least for me) to determine whether I'm improving things enough or not. Part of that is because of the windowing used in FR plots is much shorter (like half) than typically used in waterfalls, right?

I don't understand the last sentence . In the case of REW, it only makes one sweep that it then uses for everything. It does not re-run something different to get waterfalls, Impulse, etc. Back to your point, as you long as you know the severe limitations that time vs frequency brings here, and do check what happens if you change the time window, and have considered the effect of noise floor, etc., then it is fine to do as you say. For me, I find that it is a hard balance. To make any correction, I need to know the exact frequency down to 1 Hz. But if I am down to 1 Hz, then my time domain display is shot. As I tried to do the exercise for this article, I kept running into this problem in that the frequencies that I was picking in waterfall were off relative to non-smoothed SPL display.

 

[jkeny]

Can I ask, yet another, naive question?
Your original graph showed a filter reduction of the resonance at 53Hz frequency from an amplitude of 80dB (where the graph showed ringing) to about 65dB where the ringing was greatly reduced. I'm probably missing something but what is the amplitude of the signal (swept sine wave?) at this frequency? I take it that the room mode resonance has caused an increase in amplitude at this frequency? What happens if in a music signal the amplitude at 53Hz is actually approaching 80dB, do we have the ringing back?

Sorry if this is a question from the back of the class in the room correction 101 course!

 

[amirm]

Can I ask, yet another, naive question?
Your original graph showed a filter reduction of the resonance at 53Hz frequency from an amplitude of 80dB (where the graph showed ringing) to about 65dB where the ringing was greatly reduced. I'm probably missing something but what is the amplitude of the signal (swept sine wave?) at this frequency?

It is the first slice in the back of the waterfall. See the first green lines in the back:

Unfortunately my laptop rebooted and seems like I did not save those measurements . Or I would show it on play SPL graph. Using the above graph for now, you can see how the green line is now flatter and smoother relative to the rest of the response. Since the next notch is around 100 Hz, if I were using this with a crossover, I would almost be done with that one filter!

I take it that the room mode resonance has caused an increase in amplitude at this frequency? What happens if in a music signal the amplitude at 53Hz is actually approaching 80dB, do we have the ringing back?

Correct on the first question. And no, once I back out the effect, it is gone. Again think of an amp that has 2 db peak at 1 Khz. If I put in an inverse notch for the same amount in the pre-amp I would get a flat response irrespective of the input signal.

Sorry if this is a question from the back of the class in the room correction 101 course!

No problem. This is such a complex mix of signal processing, math and acoustic soup that it is hard to find one's way through it .

 

[jkeny]

It is the first slice in the back of the waterfall. See the first green lines in the back:

Ok, so the amplitude of the signal at 53Hz is 65dB & therefore this is also the amplitude at each frequency represented in the graph?

Unfortunately my laptop rebooted and seems like I did not save those measurements . Or I would show it on play SPL graph. Using the above graph for now, you can see how the green line is now flatter and smoother relative to the rest of the response. Since the next notch is around 100 Hz, if I were using this with a crossover, I would almost be done with that one filter!

Yes, I understand & therefore the null @100Hz is 50dB - down 15dB from the signal level.

Correct on the first question. And no, once I back out the effect, it is gone. Again think of an amp that has 2 db peak at 1 Khz. If I put in an inverse notch for the same amount in the pre-amp I would get a flat response irrespective of the input signal.

But this is the bit I don't get - the resonance occurs in the room at this node for a frequency of 53Hz signal. The lower the amplitude of this 53Hz the less the ringing will be audible. Conversely, the higher the amplitude, the more audible the ringing. By reducing the amplitude we haven't solved the ringing, per se, just it's audibility. So what happens if a 80dB signal @ 53Hz occurs - without the filter it would be boosted by 15dB & the ringing would be very obvious but with the filter it will be reduced to the correct 80dB amplitude. However, according to your graph this will ring audibly, no?

No problem. This is such a complex mix of signal processing, math and acoustic soup that it is hard to find one's way through it .

Yep

 

[terryj]

No, I have just been busy getting my next article out.

No worries, twas a joke.

The graph is generated using computer modelling.

Ok, let's be sure here (I mucked up which graph I posted, I was sure you would be able to work out the one I meant but your answer confuses me a little now).

In the graphs below, the graphs to the right, below, are generated by computer modelling? In other words, these are not true measured graphs??

that can't be right, so awaiting confirmation before going on.

 

[amirm]

Terry, you are confusing the heck out of me as I am lost as to what picture you are referring to . The above graphs are real measurements. They say so on the graph . The RPG graph is a simulation.

 

[terryj]

Terry, you are confusing the heck out of me as I am lost as to what picture you are referring to . The above graphs are real measurements. They say so on the graph . The RPG graph is a simulation.

Ah, thank god for that! Yeah I made a mistake in my last post you responded to, said I had included the wrong graph (by mistake the RPG one) so when you answered...well anyway, glad to have it sorted out.

Anyways, now the correct graph has been replanted in all our minds , do you know how the right hand graphs were generated?? As I said, I have not seen anything quite like it in MY travels, so do not know how to generate my own.

If I could, the following would be my comparisons. Generate as shown, ie apply a filter and show the FR changes and then the time graph, just as they have done there. Noting that to my eyes nothing seems to have changed *much* in the right hand graph EXCEPT magnitude which is very different, (all the peaks and troughs still coincide in time), AND noting that the filter applied has basically reduced the magnitude over the entire spectrum measured (particularly evident in the second set of measurements) then my comparison measurement would be to NOT apply any filter but simply lower the volume of the subs.

If my interpretation holds some water, I would expect to see much the same changes as in these right hand graphs. Along the 'time' axis not too many changes, but in the relative level axis the reduced output showing, just as it is here. In both the peaks and troughs would remain the same in time (as the 'ringing' or varying time responses of the room have not changed), just at a changed level that reflects the reduction in spl.

At these different frequencies with the different lengths of the pendulum, the length and mass of these different pendulums have not been changed at all. Sure we can swing them to have the same amount of starting swing (equal FR) but then the properties of decay of the individual pendulum determine how long it takes to decay to nothing, which can vary (ie ringing).

It COULD very well be that in some rooms, 'most rooms' (?), the differences remaining are swamped and not important. But as always we gotta be wary of extrapolating a particular case to the general.

 

[jkeny]

Yep, Terry & I are saying the same thing - the ringing hasn't been removed, just reduced in audibility by the reduction in amplitude at the frequency targeted by the filter.

 

[terryj]

Yep, Terry & I are saying the same thing - the ringing hasn't been removed, just reduced in audibility by the reduction in amplitude at the frequency targeted by the filter.

hi john, as good as it is to find something we can both agree about! I want to make clear that 'I am simply not convinced by the graphs provided', that's all. A subtle but important distinction I think. Don't worry, we WILL find something!!

"why" I am concentrating on that graph is exactly the type of thing micro has been asking (I think). If amir's point about how misleading looking at a WF can be (due to the intertwined nature of the resolutions) is accurate, and I don't particularly doubt that bit, then it is a reasonable question to ask 'ok then, what graph/measurement of time can we obtain that is not tainted by these problems?' I think that is what micro has been on about. It is a question that occurred to me as well.

In that case, the graphs shown HAVE been out forward as the 'proper' way to see what is going on. It is for that reason I am trying to get the interpretation clarified, and for now remain unconvinced with them.

 

[jkeny]

hi john, as good as it is to find something we can both agree about! I want to make clear that 'I am simply not convinced by the graphs provided', that's all. A subtle but important distinction I think. Don't worry, we WILL find something!!
......

So, you are unsure how to interpret the impulse graph on the right. Your interpretation is exactly what I see also - the post-ringing is reduced in amplitude simply because the amplitude of the initial signal (at this frequency) has been reduced in amplitude. Your point that the same post-ringing reduction would be achieved by just turning down the volume across the board seems correct. However, it seems to me that there is a room resonance at this frequency no matter what volume it is played at & therefore the amplitude of the signal at this frequency is perceived to be boosted by this resonance. I can understand how a filter is needed to reduce this frequency boost. It follows that the post-ringing will also be reduced relative to the full bandwidth signal.

My question: this may or may not result in the post-ringing no longer being audible - it depends on a number of issues but the main one under consideration here being the original amplitude of that frequency in the signal.

So Terry, my question is the corollary of yours but comes from the same logical conclusion. So I am happy that we agree

 

[terryj]

So Terry, my question is the corollary of yours but comes from the same logical conclusion. So I am happy that we agree

right ON soul brother!

 

[jkeny]

right ON soul brother!

 

[bblue]

Well, let's take a look at this graph where it shows the frequency response of the identical speaker but from different seating positions:

Which line is most faithful the recording below transition frequency? The blue, red or green? What if I changed the room? Would the lines change?

I think the obvious answer is that the room that we put the speaker in completely changes the character of the sound that is coming out of the speaker in low frequencies. It can create swings of up to 30 db. That is 6X different level perceptually (and far more as far as measurement is concerned).

That's no different than if you were sitting in a room listening to a live instrument. You'll hear a different version of it depending on your position. In the listening room, there really can't be much of a notion as to exactly what is right or wrong, since what you're listening to isn't the live instrument, but a recording of it with only two channels. Your only goal if there was one would be to determine which sound is most accurate to the recording, not the instrument! It's an lose lose situation to expect to hear the real thing.

Now, the point you are making is that during recording there were also such peaks. That is very correct. Alas, we don't know what that curve looked like. Not from one venue, or the other million that is used to produce music. So in no way can we ever match what was heard during the recording. All we can do is be faithful to the groves or the bits on the source. That faithful reproduction says don't have your room make up its on "EQ" and change the tones. Because when it does, it doesn't sound right. Somehow as humans we are able to tell that. When I was testing the sub and I played guitar music into it, I could easily tell it was "wrong" when the peaks were there. When I took them out, it sounded much more like a guitar string without the overhang and boominess and seemed to come and go. From Dr. Toole:

"Since the true nature of the original sound cannot be known to listeners one cannot say “it sounds as it should.” But listeners routinely volunteer opinions on scales that are variations of like-dislike, which frequently have a component of emotion. Descriptors like pleasantness and preference must therefore be considered as ranking in importance with accuracy and fidelity. This may seem like a dangerous path to take, risking the corruption of all that is revered in the purity of an original live performance. Fortunately, it turns out that when given the opportunity to judge without bias, human listeners are excellent detectors of artifacts and distortions; they are remarkably trustworthy guardians of what is good. Having only a vague concept of what might be correct, listeners recognize what is wrong. An absence of problems becomes a measure of excellence. By the end of this book, we will see that technical excellence turns out to be a high correlate of both perceived accuracy and emotional gratification, and most of us can recognize it when we hear it."

As to whether EQ is a natural or unnatural fix, I can't agree . There is no magic in an acoustic material that makes it more right than anything else. Acoustic material can also shape the sound just like an EQ. Here is a simulated absorber:

Clearly the absorption rate varies with frequency and there is even some ringing/vestiges of "comb filtering." Who says that is a good way to get rid of the energy in the room? Maybe it is more natural that we don't energize the room to start than to try to absorb it later!

Why on earth would you want energy in the room created randomly (as far as the music is concerned) generated from reflections, sums of reflections, cancellation of reflections, etc. etc.? ALL that can do is significantly degrade the listening experience of speakers and room combined.

To be honest, yes, an EQ can be poorly implemented and cause its own set of problems. Graphic EQs are one such example. But DSP is pretty cheap these days and we can do a very good job.

Hope I didn't miss your point .

I think you did. Big time.

There is just no way you can fix a room suffering from reflections, resonances and cancellations by simple EQ. Even if traps are not ideal or perfect, the goal is to minimize those reflections, resonances and the resulting cancellations. Since those are ADDITIVE (yes, even the cancellations-nulls) to what the speaker is producing, they don't contribute anything but distortion to the perceived sound. Why tolerate them in the first place? Yes, you might not be able to reduce all artificially generated frequencies to zero, but you can get them down below a destructive level acoustically.

Here's another example, one I'm particularly familiar with. Suppose you have a room that among others, has a very broad resonance at around 91 Hz. It's low Q and more than 1/2 an octave wide. That means that not only will a 91 Hz signal excite it, but so will a 90,92,89, and 93 Hz signal. But regardless of what excites it, it resonates at 91 Hz. What does that mean? It means that any frequency in that excitation range will trigger a 91Hz response which will mix with the original, changing its apparent tuning and overall distinctiveness. That whole frequency range is trashed, basically.

Now suppose you EQ a 1/2 octave dip at 91 Hz? That just reduces the speaker's output in that range, but it doesn't do anything to the character of the room in that range. It still behaves in the same manner, obscuring detail. It'll be a little less obvious because the triggering level is lower, but it's still there. You have to fix it at the source (room modes usually) and then if needed, fine tune with eq.

It seems incredibly obvious to me. EQ alone will never produce a room that is truly a listening reference. This IS WBF, right??

If a listening area is already pretty decent, with even decays across the necessary bandwidth (below the room transition frequency) and no major nodes interfering, then sure, EQ might help fine tune the curve a bit, but that's all it can do.

--Bill

 

[amirm]

Anyways, now the correct graph has been replanted in all our minds , do you know how the right hand graphs were generated?? As I said, I have not seen anything quite like it in MY travels, so do not know how to generate my own.

That impulse response, is well, an impulse response. The simplest measurement around and you can easily make it yourself. See below.

If I could, the following would be my comparisons. Generate as shown, ie apply a filter and show the FR changes and then the time graph, just as they have done there. Noting that to my eyes nothing seems to have changed *much* in the right hand graph EXCEPT magnitude which is very different, (all the peaks and troughs still coincide in time), AND noting that the filter applied has basically reduced the magnitude over the entire spectrum measured (particularly evident in the second set of measurements) then my comparison measurement would be to NOT apply any filter but simply lower the volume of the subs.

Here is an impulse response with and without a single filter (same sub arrangement as before)

We see similar results to Dr. Toole's measurement. The amplitude of the ringing is reduced and with it, its audibility. At the cursor position for example, there is a 40% reduction in negative swing there. For the sake of discussion, if the blue cursor also represented the noise floor or minimum average SPL of music playing by other speakers, the audible ringing would have terminated for the filtered version at that time (18 msec) whereas for the original it would still be audible and last another swing.

I think you have an unrealistic expectation that with insertion of one filter the room becomes an ideal anechoic chamber with all the resonances disappearing. That is just not so and will never be so in typical listening rooms. What we are trying to do is dial out gross amounts of it. We will never achieve flat response no matter what you do. It is just a fact of life that we are going to have variations in both time and frequency. Those resonances support the low amplitude tail and hence will become harder to dial out. But fortunately their levels at some point get low enough that we don't worry about them.

Note that acoustic products have the same issue. Even expensive anechoic chambers often are not flat down to 20 Hz. And there we are talking about multi-feet deep treatment. Fortunately no one listened to the music that way when it was produced either so we are not poorly situated in that regard.

If my interpretation holds some water, I would expect to see much the same changes as in these right hand graphs. Along the 'time' axis not too many changes, but in the relative level axis the reduced output showing, just as it is here. In both the peaks and troughs would remain the same in time (as the 'ringing' or varying time responses of the room have not changed), just at a changed level that reflects the reduction in spl.

As I explained above, the sound is still going to bounce around the room and set off resonances at many frequencies. Run a room mode calculator and you see them all listed. Also note that if I just change the level for the sub, I will get no frequency domain correction so I would accomplish nothing. The high order bit here is frequency response variations. That part is audible completely. Whether you hear a hanging bass note is secondary to that.

At these different frequencies with the different lengths of the pendulum, the length and mass of these different pendulums have not been changed at all. Sure we can swing them to have the same amount of starting swing (equal FR) but then the properties of decay of the individual pendulum determine how long it takes to decay to nothing, which can vary (ie ringing).

I think the analogy is causing more confusion than helping . A resonant system has a natural frequency that starts to help amplify the input signal. If I know that frequency and lower the amount of energy I put in the resonator, then even though it is amplifying, it won't overboost what I put in it. And since it is not overboosting, if I stop, it will also stop sooner. One way I can tell that is that if its rate of decay is faster. You can see this already in my one filter setting. The reason you don't see the trend continue is that weaker resonances are still there after some point and start to dominate and keep ringing going.

It COULD very well be that in some rooms, 'most rooms' (?), the differences remaining are swamped and not important. But as always we gotta be wary of extrapolating a particular case to the general.

Why should we be wary? We took out the biggest contribution to frequency response and time domain errors. You can run the sweep and see if you also have resonant peaks. If they are there and you take them out, you can and will expect an improvement. As I post initially even though I quickly picked those values, the subjective result was very positive with smooth response and definite reduction in time domain. I have also post research papers including two that involve room correction and they all point to the same thing. You can also read the REW manual and it tells you the same thing.

Yes, if by magic your room has no resonances this doesn't apply to you. This is why we don't blindly dial in some random EQ. We run measurements and put in an EQ with the right center frequency and Q (bandwidth). If you can't run this yourself, automatic correction systems will do this for you and the good ones will do a better job than the bad one.

 

[amirm]

That's no different than if you were sitting in a room listening to a live instrument. You'll hear a different version of it depending on your position.

Not just the position. Everything changes in the other room. Further, the entire science of it changes if the room is sufficiently big such as a concert hall. We will then get a diffused soundfield and the bass becomes far better as we get a dense set of modes. The problems we are talking about here are perils of small room acoustics which unfortunately make our life difficult. But yes, if you are talking about a small room, same science will apply. I will not only hear variations due to where I sit, I also hear variations in the identical spot because the room rooms are guaranteed to have different frequency responses due to so many variable contributing to that from speakers to room. For this reason, we need to divorce ourselves from thinking we can ever replicate what the talent heard. Unless both presentations are in anechoic chamber with identical speakers, we are out of luck there.

In the listening room, there really can't be much of a notion as to exactly what is right or wrong, since what you're listening to isn't the live instrument, but a recording of it with only two channels. Your only goal if there was one would be to determine which sound is most accurate to the recording, not the instrument! It's an lose lose situation to expect to hear the real thing.

Which is exactly what I said and repeated above. Not sure why you quoted me and reply as if I disagreed.

Why on earth would you want energy in the room created randomly (as far as the music is concerned) generated from reflections, sums of reflections, cancellation of reflections, etc. etc.? ALL that can do is significantly degrade the listening experience of speakers and room combined.

No, they don't degrade the *experience.* How do we know that? We test people. We test them with or without that reflection and with the exception of recording engineers, the rest of the world tends to like side reflections better than not having them. Not all reflections. But strong side reflections. Yes, the sound is "distorted." I cover this in my other thread/article: http://www.whatsbestforum.com/showthread.php?8226-Psychoacoustics-of-Room-Reflections.

I can cite research paper and listening test after listening tests to demonstrate the above. For brevity and an interesting data point from someone who should think like you say, here is George Augspurger (designer of PRO listening rooms) in his AES paper, LOUDSPEAKERS IN CONTROL ROOMS AND LIVING ROOMS:

”Third, I did a lot of listening with various amounts of absorptive treatment in the comers behind the speakers. When first-order reflections were largely absorbed I noted that locations of individual sound sources were more precise, that the timbre of individual instruments was more natural, and that the overall stereo picture was more tightly focused. These observations agree well with other reported listening tests. Nonetheless, after extensive listening to classical and pop recordings I went back to the hard, untreated wall surfaces. To my ears the more spacious stereo image more than offset the negative side effects. Other listeners, including many recording engineers, would have preferred the flatter, more tightly focused sound picture.”

So if the goal is to have a pretty graph, you are absolutely right. We should create an anechoic chamber and get rid of all reflections. But if the goal is to *enjoy* music, listening tests and research into how we hear, which includes why we don't hear distortions that show up on graphs, tells us that some reflections are beneficial. Others are not. Floor reflections for example bring no benefit since both ears hear it equally. So you would want to absorb them. Likewise front and back walls do not have benefits so while they are not damaging per-se, they can be attenuated if you need absorption. On that point, mid to late reflections can be a negative so all rooms need to have fair amount of absorption in them. The surfaces I just listed make good candidates.

There is just no way you can fix a room suffering from reflections, resonances and cancellations by simple EQ. Even if traps are not ideal or perfect, the goal is to minimize those reflections, resonances and the resulting cancellations. Since those are ADDITIVE (yes, even the cancellations-nulls) to what the speaker is producing, they don't contribute anything but distortion to the perceived sound. Why tolerate them in the first place? Yes, you might not be able to reduce all artificially generated frequencies to zero, but you can get them down below a destructive level acoustically.

I have explained "why tolerate" them above. I also want to be clear as to which part of the room acoustics we are talking about. In this thread we are all about the bass frequencies/sub. In this region, psychoacoustics does not play a role and the soundwaves act differently. It is important to not mix topics and say that I am talking about EQing everything. Not even close. I am talking about bringing down resonant peaks in bass frequencies. I am not telling you to go and apply EQ to 8 Khz and expect to necessarily do good. You could be making the sound worse if you don't know the cause of that. As to doing things acoustically, let's see the low frequency of a room without EQ. Show me how none of the variations in its frequency response matter.

Here's another example, one I'm particularly familiar with. Suppose you have a room that among others, has a very broad resonance at around 91 Hz. It's low Q and more than 1/2 an octave wide. That means that not only will a 91 Hz signal excite it, but so will a 90,92,89, and 93 Hz signal. But regardless of what excites it, it resonates at 91 Hz. What does that mean? It means that any frequency in that excitation range will trigger a 91Hz response which will mix with the original, changing its apparent tuning and overall distinctiveness. That whole frequency range is trashed, basically.

Now suppose you EQ a 1/2 octave dip at 91 Hz? That just reduces the speaker's output in that range, but it doesn't do anything to the character of the room in that range. It still behaves in the same manner, obscuring detail. It'll be a little less obvious because the triggering level is lower, but it's still there. You have to fix it at the source (room modes usually) and then if needed, fine tune with eq.

We always start with minimizing the problem. That solution calls for optimal placement of subs to cancel some of the modes, and use of more than one to cancel some of the others. Once you do this you still have some peaks. Those peaks are then brought down to give us near flat response. I will never have flat response as I mentioned to Terry. It simply is not possible to get lab type results. And you don't get there with acoustic products either. You will absolutely still have peaks and valleys and those sure will have a character.

It seems incredibly obvious to me. EQ alone will never produce a room that is truly a listening reference. This IS WBF, right??

The foundation of this forum is have a mix of solid science and open discussion of audio. All opinions are respected though there is no obligation to accept . What I am describing here is backed by decades of acoustic research. You are hearing 10% of it. Over time, I will post more. The science is all targeted toward bettering the sound in the room. So the goal is the same as what we have here. When Bradley researchers speech intelligibility and finds that room reflections contribute up to 9 db there, we want to know about that even though it goes against one's guts that reflections hurt rather than help. The science tells you that if you take out reflections you take out energy from the room and that has its negative effects. So yes, get ready to hear the best that the acoustic science has to present us. That is what this forum is about. Again, no obligation to accept any of it.

If a listening area is already pretty decent, with even decays across the necessary bandwidth (below the room transition frequency) and no major nodes interfering, then sure, EQ might help fine tune the curve a bit, but that's all it can do.

--Bill

Well, you should repeat my exercise. Put a single sub in a corner and measure. Take out the first massive peak and tell me if made your room sound less natural. I assure you it will not. There is reason most high-end subs come with built-in DSP filters and software that programs them. Why would they do that if it makes the sound worse?

Reading your posts makes me think that you are thinking that it is being said that EQ should be used as the only tool to fix anything that is wrong in the room. That is not remotely what I am saying. Again, we are talking about narrow but important area in small room acoustics which is bass resonances. We know these will be there. You are right that you can't take out floor reflections using EQ. You better have a thick carpet there. Likewise if your room is too live, no EQ will save it although getting rid of bass resonances is an improvement.

I plan to write a simple tutorial on the top level picture as I think that is most of our problems. Here is a quick version of it. To get good sound, you need to do the following:

1. Start with good speakers. This is what makes sound in the room. I am amazed that entire texts are written on room acoustics but ignore speakers. What makes a good speaker is a subject of another thread .

2. Make sure you have sufficient absorption for mid to late reflections. Too much here causes intelligibility problems and makes for bad sound in general. Too little of it and the room becomes uncomfortable (think anechoic chamber).

3. Deal with problematic reflections (e.g. floor).

4. Optimize your bass frequencies. This is what I am talking about in this thread. Studies show that we contribute 30% of sound fidelity to how good the bass is. So if we fix this, we are one third of the way there.

 

[terryj]

Ahh, I use impulse graphs all the time, but never from that 'far a distance'. Yep, once you scale waaay back then et voila, they are the graphs!

And in raising those questions, I also outlined how I would test my hypothesis..

If I could, the following would be my comparisons. Generate as shown, ie apply a filter and show the FR changes and then the time graph, just as they have done there. Noting that to my eyes nothing seems to have changed *much* in the right hand graph EXCEPT magnitude which is very different, (all the peaks and troughs still coincide in time), AND noting that the filter applied has basically reduced the magnitude over the entire spectrum measured (particularly evident in the second set of measurements) then my comparison measurement would be to NOT apply any filter but simply lower the volume of the subs.

If my interpretation holds some water, I would expect to see much the same changes as in these right hand graphs. Along the 'time' axis not too many changes, but in the relative level axis the reduced output showing, just as it is here. In both the peaks and troughs would remain the same in time (as the 'ringing' or varying time responses of the room have not changed), just at a changed level that reflects the reduction in spl.
.

Ran a few quick ones this morning, want to do a bit more digging still, but for now and quickly I can say without a shadow of a doubt my hypothesis is busted. No question, zero maybes, completely and totally wrong.

You can simply take my word for it, or at some time I can post the graphs. I still want to do a bit more anyway so once that is done and I find a usb stick I can post these results. I ran the same sweeps, changing only the levels as suggested above. Not any change anywhere. Completely identical no matter that the measurements differed by ten db.

And yes, also a few quick sweeps with changing a filter (even adding a boost as opposed to cutting) and very much the same differences appeared in the impulses as amir has already posted.

Needless to say my curiosity has been piqued, love a good learning problem to sink the teeth into. Have a few more ideas I want to follow still.

It will be a few days till I can get to it, have a rather pressing family problem atm. You see, I have to catch up with the girls on Breaking Bad before we can all watch the next season together. We are currently running about four episodes a night to catch up, and as some of you might agree with me that takes precedence over all else right now!

Anyway, very interesting stuff amir, thanks.

 

[Soundminded]

Modern computing power has put in our hands incredibly powerful measurement tools. Prime example is Room EQ Wizard (REW), a free program which is an amazing toolbox of acoustic measurements. Alas, while computing power may be free enabling us to run such analysis that used to cost a lot of money in dedicated hardware, the fundamentals of what the tool measures must still be understood. Otherwise, it is exceedingly easy to arrive at the wrong data and worse yet, wrong conclusions about the science.

The purpose of this article is to demonstrate a key element: the relationship between time and frequency resolution. This is an underlying signal processing concept that is core to functionality of everything from audio and video to compression to how we measure room acoustics. Alas, while the concept is rather simple, it is not intuitive nor is it talked about much outside of the circle of industry researchers and professionals. Simply put, time and frequency resolution are enemies of each other. If attempt to get a lot of frequency resolution, you lose time resolution and vice versa. Oh, please stay with me . This will get interesting in a hurry as we apply it to electronic correction/EQ of the sound in our rooms.

When we look at bass response of our room, often we find massive peaks and valleys. These are caused by so called room modes and we could see swings of 20 or 30 db (every 10 db means 2X louder perceptually). We clearly hear the frequency response variations. These room "resonances" have another manifestation. They create "ringing." Ringing is a time domain phenomena where an impulse -- an infinitely short pulse with infinite amplitude -- continues to live on in the room even after it has vanished. The lingering effect of a low frequency note after it has gone away from our source causes the rest of our music/video soundtrack to sound bloated and boomy. Bass notes are not sharp and distinct but seem to last long and overwhelm the higher frequency tones.

An assertion by some in acoustic circles is that you must correct the time response in addition to frequency response. Some will say that this is only possible using acoustic material and not electronic means. As a matter of mathematics, this can't be right. What causes ringing is the room resonance. The very same resonance will also increase the amplitude at that frequency. Fundamentals of signal processing fall apart if what these people say is true. For their proof, they present acoustic measurements that seemingly show that correcting frequency response did not improve the time domain ringing. Let's see if we can duplicate their work and demonstrate where they have may have gone wrong and in the process cover the time and frequency relationship.

For this exercise, I took a subwoofer and stuffed it in the corner. I shut off the crossover and ran a sweep up to 200 Hz. Since the topic of interest here is both frequency smoothness and time domain ringing, we invoke the "waterfall" display in REW. Here is what I got:

The graph shows at time zero the original signal and its frequency response and then in successive "slices" moving forward in time it will perform the same analysis showing that frequency response. It will keep going until the limits of the graph setting are reached which in this case is 300 milliseconds or 1/3 of a second. I suspect other than noticing the non-smooth frequency response in the original signal in the back, you are scratching your head as to what else this graph means. It sure is pretty though . But it is not giving us much insight. The problem is one of settings for that measurement and graph. Let's dig into them.

The Noise Floor
The biggest problem here is that we have not paid attention to the vertical scale. This axis shows the amplitude of the sound in the room. Lowest level is 15 db SPL and highest is 95 db SPL. The max is OK. The problem is the min. You can't just pick a random number here. There is a minimum below which we are just measuring the noise in the room which in the case of low frequencies can be a lot. See my article on room noise: http://www.madronadigital.com/Library/RoomDynamicRange.html.

Let's put this to practice. Here is the same REW waterfall display, this time with me turning off the output so that nothing was playing:

We see noise floor that reaches up to 45 db, a whopping 30 db higher than what we were measuring before! So a bunch of data in the previous graph was simply noise that is independent of what our system is doing. Let's overlay this noise on top of our sub response and see the two together:

Let's correct for that by raising the bottom level of our measurement to top of the noise floor. We do this simply by sliding the scroll bar or setting the limits in REW for our graph. While we are at it, let's perform some other changes. For one, let's re-orient our viewpoint onto the 3-d graph to make it better to see the time response. Hit "Controls" button on top right and change X to 21, Y to 122 and Z to 50. This is what we get now:

I am liking this better already! We have a much clearer picture since we got rid of the noise. We see the frequency response changes and with it, the associated time domain activity. The peaks in the amplitude of the frequency response are accompanied by time domain forward/back continuation of signal.

Time and Frequency Domain Resolution
As I mentioned, the core of this article is to talk about how these two factors are enemies of each other. We can easily demonstrate this by changing the "Window" parameter in REW. The Window sets the time window or put another way, the number of audio samples uses to determined the frequency response. If you set this to 1 second, then you get a resolution of 1 Hz. If you set it to 0.1 seconds, then the resolution jumps to 10 Hz. Let's see what happens when we set this to 40 milliseconds for a 25 Hz frequency resolution:

We see a huge change in the graph. The curves left to right have become very smooth and a lot of the ups and downs have disappeared. The reason? We have lost frequency resolution. As stated, we now are dealing with chunks of 25 Hz at a time. This is way, way too big for analyzing bass frequencies because room resonances are much smaller than this. So let's optimize for frequency response by going the other direction. Here is what it looks like with 300 milliseconds window and 3.3 Hz resolution which by the way, is the default for REW:

This brought back our frequency ups and downs but notice what it did in time domain. Look at how the curve front to back for any one resonance peak has a huge hump. Just like what we had in the other direction before. Here, we know the offending frequencies but no longer know what they are really doing in time domain. I.e. determining ringing is more difficult since time domain response is distorted.

We therefore see that we cannot have high resolution in both time and frequency domain. One comes at the expense of the other and there is no way around it.

Optimized Display and Filtering
For this test, the best settings I found were 400 milliseconds for the total "Time Range" or total display and 200 milliseconds for time window. 100 milliseconds also works well but values above and below this range just cause too much distortion. I eyeballed a frequency response peak and picked 53 Hz as the first frequency to go after. I put in a parametric EQ at that frequency to pull it down and the frequencies around it. Here is what we get with the new settings and filtered response, overlaid on the non-filtered measurement.

What do you know? The math worked! I put the cursor at 53 Hz (blue line) and you can easily see that time domain ringing has heavily subsided at that frequency. The rest remains there because I have not touched them. BTW, a key note. If you change the Time/Window setting in one graph of REW and attempt to overlay another, you have to go to the other graph and apply the same settings. Otherwise, REW will happily mix the two graphs even though they were not analyzed the same.

Let's add another filter at 140 Hz:

The math works again with the both the frequency response peak going down and time domain ringing reduced.

Oh, do not be alarmed if you see some extra bits pop up in the noise floor of the system after you make changes. This is a much more advanced topic but the analysis of the system in slices can cause errors in measurements due to truncation of samples at either end. Focus on the big picture here, pun intended .

Let's go all out and add a few more filters:

We see across the board improvement in time domain and frequency domain. In practice, we would only care about response to ~100 Hz as the rest would start to get filtered by the crossover. So that sharp drop around 100 Hz is a much smaller problem than it seems.

Subjectively the bass response became all that I said at the start. I played a string track with I could now hear every one plucked even though all I was listening to was the sub! Before one pick would result in a lot of boom and the strings would all run into each other as they were played. So clearly we had made improvements in time domain. Overall bass level however seemed low. This was fixed by a boost of the entire sub by a few db. I know had the impact but kept all that was clean about my optimization. Considering that I was just playing to create these set of scenarios, these are pretty encouraging results.

For grins, let me show you how not to do this. Here is the overlay of the 53 Hz correction over no correction but with levels wrong:

Looking at this, it is very easy to conclude that we did not correct anything in time domain. It seems like a jumbled mess before and after. But per above, this is the wrong use of the tool. Set the levels right and you too can be in good hands of science .

The first graph is of limited value with many shortcomings that hide useful information. For example, the response of the subwoofer is almost impossible to separate from the room response. This could be corrected by a) normalizing the response at T=0 taking the woofer out of the equation and showing only the differences and b) rotating the graph to make it orthogonal to the T axis, that is looking straight down the T axis. the output as a function of time would be presented as a series of curves on a two dimensional X-Y plot with different lines signified by different colors., This would show us where the room resonance and anti-resonance (cancellation) points are. The measurement is non directional. Since the worst reinforcements and cancellations are usually the result of parallel surfaces, the front/back wall, the side walls, the floor/ceiling, only directional microphones showing us response in each axis would tell us where the problems lie and where to apply sound absorbing material or other means to smooth the response. I'm not saying that this sort of graph isn't a good start but that's all it is, a start. Until it can provide much more useful information, it's of limited value in understanding and improving room acoustics and room/speaker interactions.