Frequency response and human hearing

[LinearMan]

There have been many threads of late discussing the acoustic response of rooms, bass response, et al. All this seems to me like discussing the number of angels that can dance on a pinhead ~ a complete waste of time!

For example, there have been suggestions on using in room response modelling to try to achieve a flat frequency response. Well, human hearing is non linear. Not only that, it varies from one individual to another. For those that are interested, the following is quite a good text:

http://hyperphysics.phy-astr.gsu.edu/hbase/ph4060/p406i.html


I suggest we all use the best instruments available on the planet, at no cost. Your very own ears! The human response to music is entirely subjective, so attempting to introduce objective comparisons is essentially meanlingless.

To all those that look to this forum for advice and guidance, I suggest that the best advice is to use resources such as this to put together a shortlist and then listen to the equipment in your own listening room. The rest is all audiophoolery.

 

[bottleneck]

Are you suggesting that the room has no influence on the final sound, and that any alterations we may make in terms of decoration, furnishings, room dimension, structure etc are irrelevant?

 

[LinearMan]

Of course I'm not! Obviously these things will affect the sound. But how it affects the perceived sound will vary from person to person, room to room. Thus there can never be any meaningful objective measurement that has any relationship to the individual's perception of the sound.

 

[MartinC]

That human hearing response varies with frequency is I'm afraid irrelevant LinearMan. The point is that a system with a linear response will reproduce the full spectrum as recorded, and assuming it is well recorded thus as it was at the time of the actual performance. The frequency response of the listener's auditory system will then filter this in the same way when hearing either the live event or the recording played back.

Differences in frequency response from person to person are of course significant, and to some extent account for apparent variations in people's taste of musical presentation. But again, a system with a flat frequency response should ideally give the particular listener the same experience at either the live event or listening to a recording (but this experience would to some extent be different for another listener). Two people sitting next to each other at the same concert will not hear the same thing if their hearing response is different, but both would probably like to reproduce what they heard at that concert in their home.

 

[wolfgang]

For example, there have been suggestions on using in room response modelling to try to achieve a flat frequency response. Well, human hearing is non linear. Not only that, it varies from one individual to another.

I think many audiophiles misunderstood what it means by these none linear FR and biological variation in human ears. The simple reason we try to achieve a flat loudspeakers frequency response in our room are based on previous experience that was found by asking lots of people what their ears prefer in carefully arranged listening comparison to many types of loudspeakers when listening to nothing else other then music. Since almost everyone suggest they find speakers with flat FR sound better to their ears, most people also come to accept perhaps that is indeed a good thing to recreate from now on.

 

[bottleneck]

wolfy, I dont think thats a fair assumption.

lots of people prefer speakers without a flat response. Many people prefer speakers without a flat response without even realizing thats what they have and enjoy listening to every day. Indeed, I'd gamble that some people would even argue in favour of a flat response, but in actuality may listen to and prefer speakers without a flat respose.

My last paragraph is irrelevant to the subject of room acoustics however. Paying lots of attention to room acoustics, speaker placement within the room etc will pay far more dividends than a box change.

To put the original post on its thread - to all those visiting the forum looking for answers - spending time on setting up your room to sound at its very best is the very best and most significant upgrade you can make.

Using myself as an example, I recently got some advice (about speaker height and placement) that has completely changed the quality of the bass in my room. Its just a lot lot better now. Its sad that not everybody has had that kind of road to damascus moment.

 

[MartinC]

The simple reason we try to achieve a flat loudspeakers frequency response in our room are based on previous experience that was found by asking lots of people what their ears prefer in carefully arranged listening comparison to many types of loudspeakers when listening to nothing else other then music. Since almost everyone suggest they find speakers with flat FR sound better to their ears, most people also come to accept perhaps that is indeed a good thing to recreate from now on.

I'm not convinced by that tbh. Surely the reason at least the starting point of speaker design is to give a flat frequency response is exactly as I said above? It's certainly the logical starting point, as with all other pieces of HiFi kit, if you want an accurate reproduction of the recording. Suggesting that the idea of a flat frequency response being good comes from extensive listening tests seems somewhat odd.

In practise I think Bottleneck is right about what people preferring not being a perfectly flat response. Certainly from what I've heard of people using TACT room correction, very few seem to find a flat response their setting of choice?

 

[oedipus]

I'm not convinced by that tbh. Surely the reason at least the starting point of speaker design is to give a flat frequency response is exactly as I said above? It's certainly the logical starting point, as with all other pieces of HiFi kit, if you want an accurate reproduction of the recording.

The problem with a "flat frequency" response is that it's difficult to acheive because in practice there are two things to worry about: the direct "on axis" response and the power response (essentially how the speaker radiates in all directions). That later one is very hard to measure and it interacts with the room. Then, there's the ear's perception of the direct and reflected sound.

Suggesting that the idea of a flat frequency response being good comes from extensive listening tests seems somewhat odd.

Actually it's not obvious what people like. It's been the source of much argument, given the difficulty with measuring power response. It was the work by Floud Toole on listener preferences that sorted this area out.

It turns out that people like a smooth on axis response with smooth off axis roll off (this means that reflected power is somewhat even too.)

Certainly from what I've heard of people using TACT room correction, very few seem to find a flat response their setting of choice?

Well, the default Tact target curves all have a falling response (-2/3dB) in the top octave. The rationale for this is that, in general, without a falling top octave, speakers in rooms, sound overly bright.

 

[wolfgang]

All people that I know who use Tag processor with TMERQ or the D series Velodyne Sub reported back they do indeed prefer the sound of their loudspeakers after they managed to 'flatten' the frequency response. The effect seems to be improving clarity and intelligibility of the vocals and speech. As Oepidus suggested there are some groups like those that are carried out by Floyd E. Toole already who tried to get some empirical evidence as to what constitute nicer loudspeakers characteristics by asking lots of people to do nothing other then listen to them and many more variable types of designs. Since they asked lots of people repeatedly they claimed they do see a pattern that build up as to what the majority of people prefer as loudspeakers that appear to sound nicer especially when the listeners couldn't see the brand names. Incidentally those who seem to deviate from the normal preference appear to have hearing that are a bit more different to the biological median type (another way to say defective).

 

[titian]

I would like to know when are you talking about "flat frequency" response. There is always a certain tolerance in dbs and this is sometimes very different according to people who talk about "flat frequency" response.

Also it would be interesting to know where in the room you want to archieve this. Only at the sweet point, everywhere? In some places you just need to change the position of your head (microfon) to get other values.

I wonder if sometimes the reverberation time is more important than the flat frequency response at home.

 

[Tenson]

I would like to know when are you talking about "flat frequency"

I wonder if sometimes the reverberation time is more important than the flat frequency response at home.

Even in really top end studios where they have pretty much all the money in the world to spend on it, a 'flat in room response' is considered achieved when 20Hz-20KHz varies by no more than +/-6dB and has a smooth response.

Your comment about reverb time I completely agree with. In my experience the effect of 'boomyness' is more down to the rooms RT60 (time it takes for the reverb to die to -60dB) than frequency amplitude. It has also been my experience that before it is worth playing with digital room correction methods it is always worth getting the rooms reverb time below 0.5 seconds across the board. This ensures a clean pallet for working on the remaining frequency amplitude problems via digital room correction. Some of the frequency amplitude problems will have been fixed by getting a good reverb time anyway because it is essentially the reverberations that cause them; standing waves etc..

 

[titian]

My acoustician talks of a difference of 5 dbs to be the best results for studios built especially for acoustic purposes without an excessive usage of special electronic units.
The first time he cam to measure my room, which is just under the roof of my house, he noticed that the figures weren't that bad:
- between 50Hz to 15KHz a difference of 15 db
- considering the 30Hz about 25 dbs
- there were two or three falls which created this differences due probable bad positioning of the LS or electronic problems
- the reverb time was above 100Hz less than 0.35 sec.; above 8K under 0.2 secs! There was though a peak of 0.8 secs at about 80Hz and 0.9 at 30Hz.

The main 'problem' was not to put absorbers for mid and high frequencies.
After I developed the acousticians recommendations the measurement came to:
- 7 db difference from 35Hz to 12K. I could achieve this difference up to 20KHz if I would turn in my LS directly to the sweet point but I prefer having a wider sweet point and I believe the ratio direct vs diffusion sound is ideal like it is now.
- at 30Hz there is 7db more difference
- the falls were flattened with slight other positioning of the speakers
- the reverb time was less than 0.3 secs from 65 Hz. At 50Hz there was still 0.7 secs.

I wasn't happy with these results even if the acoustician considered his job finished. He gave me then recommendations (7 bass absorbers) for flattening the 30hz peak and I made them done. End of this month I will have equipment to do the last measurements. It was very interesting to go through this work bzw field.

regards

 

[oedipus]

I would like to know when are you talking about "flat frequency" response. There is always a certain tolerance in dbs and this is sometimes very different according to people who talk about "flat frequency" response.

Well, that's the "audiophile" approach to music reproduction, people have a "tolerance" for crap sound, which is why there are so many poor speakers are on the market.

You can divide the spectrum as follows: from about 400Hz up you are listening to the speaker combined with room reflections. below 400Hz, what's you're hearing is the room being driven by the speakers.

Above 400Hz, treating the room will kill reflections, but if the speaker has a poor direct ("on axis") frequency response you are screwed.

Below 400Hz in small spaces (and in acoutical terms, all home listening rooms are small spaces), the room modes dominate, and you get these huge peaks and troughs in response - particularly below 100Hz.

Also it would be interesting to know where in the room you want to archieve this. Only at the sweet point, everywhere? In some places you just need to change the position of your head (microfon) to get other values.

Killing the bass peaks, improves the sound everywhere.

My acoustician talks of a difference of 5 dbs

Is that +/-5dB, or +/-2.5dB - because there are plenty of good speakers with a +/-2.5dB frequency response over a broad range. I'm listening to a pair of $1.5K active monitors which are have a better than +/-2dB window from 400Hz-14KHz right now..

excessive usage of special electronic units.

What a load of bunk! You need a parametric equalizer to control the bass. You should ask your "acoustician" this question "are room modes minimum phase?" If he doesn't understand the question, or it's relevance to electronic equalization, then you should stop paying him and show him the door.

You might also ask hgim what he thinks of equipment made by Weiss and GML as chances are your recordings passed through one or other of those parametric equalizers during mastering.

The first time he cam to measure my room, which is just under the roof of my house, he noticed that the figures weren't that bad:
- between 50Hz to 15KHz a difference of 15 db

If your speakers have 15dB (or even 7dB) swings above 400Hz they are CRAP and should be disposed of immediately. Tell your acoustician you want an windowed impulse response taken at 1m from each speaker. By windowing this you can generate the high frequency response of the speaker (without an anechoic chamber). If your acosutician doesn't how to do this: show him the door.

If he makes some excuse for not doing it, he's being polite and already knows that your speakers are CRAP.

- considering the 30Hz about 25 dbs

Well, if you will have two sub's what do expect?

- the reverb time was above 100Hz less than 0.35 sec.; above 8K under 0.2 secs! There was though a peak of 0.8 secs at about 80Hz and 0.9 at 30Hz.

That's nothing particularly special - reverb is a somewhat pointless measurement in these situations. I expect you paid for it though

I wasn't happy with these results even if the acoustician considered his job finished. He gave me then recommendations (7 bass absorbers) for flattening the 30hz peak and I made them done.

Which could easily have been killed with appropriate equalization.

Your acoustician probably took one look at your cables and knew immediately that that solution wouldn't fly though

 

[titian]

Well, that's the "audiophile" approach to music reproduction, people have a "tolerance" for crap sound, which is why there are so many poor speakers are on the market.

You can divide the spectrum as follows: from about 400Hz up you are listening to the speaker combined with room reflections. below 400Hz, what's you're hearing is the room being driven by the speakers.

So it's not only speakers : you can have also absolutetly flat frequency response speakers (with no tolerance) and have nevertheless high tolerance. This is also in concerts halls so and it has nothing to do with crap sound. But there you are: you love to push down everything you don't like. I really don't envy you and the persons near you. .

Above 400Hz, treating the room will kill reflections, but if the speaker has a poor direct ("on axis") frequency response you are screwed.
Below 400Hz in small spaces (and in acoutical terms, all home listening rooms are small spaces), the room modes dominate, and you get these huge peaks and troughs in response - particularly below 100Hz.
Killing the bass peaks, improves the sound everywhere.

agree

Is that +/-5dB, or +/-2.5dB

You're not telling me that you cannot divide 5 by 2?

- because there are plenty of good speakers with a +/-2.5dB frequency response over a broad range. I'm listening to a pair of $1.5K active monitors which are have a better than +/-2dB window from 400Hz-14KHz right now..

The 5 db tolerance is at the listening point which is nearly 3 meters from the speakers. Considering what you wrote before and if you understand what you wrote, this means that the 5 dbs include all the influences of the room. Or not?
Today it is quite easy to build speakers with nearly no tolerance: if all of them also sound good, I let your majesty to decide that.
But this is not the concern of my posting.

What a load of bunk! You need a parametric equalizer to control the bass.

why not, that is a way to solve the problem. If I have a flat frequency speaker system and my room causes a probem, I prefer solving it where the cause comes from.

You should ask your "acoustician" this question "are room modes minimum phase?" If he doesn't understand the question, or it's relevance to electronic equalization, then you should stop paying him and show him the door.

..and again I feel your negative attitude towards people who you don't even know. Again I don't envy your personality.
Anyway I will ask him that.

You might also ask hgim what he thinks of equipment made by Weiss and GML as chances are your recordings passed through one or other of those parametric equalizers during mastering.

I will

If your speakers have 15dB (or even 7dB) swings..

IF. You don't need to be in fantasialand...

If your speakers have 15dB (or even 7dB) swings above 400Hz they are CRAP and should be disposed of immediately.

Even if I might agree to what you wrote, again your best known word appears for another time and again your point of view is pointed out to be the truth. I respect though those milions of persons who prefer other stuff even if I don't always like it.

Tell your acoustician you want an windowed impulse response taken at 1m from each speaker. By windowing this you can generate the high frequency response of the speaker (without an anechoic chamber). If your acosutician doesn't how to do this: show him the door.

was done ages ago.

If he makes some excuse for not doing it, he's being polite and already knows that your speakers are CRAP.

Well that is what you expect. You poor guy.

Well, if you will have two sub's what do expect?

great music, and enjoy music

That's nothing particularly special - reverb is a somewhat pointless measurement in these situations. I expect you paid for it though

I paid not per hour. With that measurement or not it would have cost the same. But I thik it is wonderful of you to think and to be so concerned about my money. I appreciate that very much. Very kind of you.

Which could easily have been killed with appropriate equalization.

I agree

Your acoustician probably took one look at your cables and knew immediately that that solution wouldn't fly though

probably but I think those cables are more of a problem for your mind. And again I don't envy you poor guy.
I'm very sorry that I gave you such problems.

 

[wolfgang]

Interesting stuff. Titian you looks like you have made further progress. Oedipus agrees with you and I think I am beginning to understand why some people suggest it might be no bad thing buying studio loudspeakers and I might just go and listen to a few when I have the opportunity to start all over again.

When trying to understand basic room frequency response of my loudspeakers by doing some (my own amatuer measurements using laptop & SPL meter) in my own living room it does raise more questions then answers. The speakers FR appear flattish as long as it is near field (less then 2m) but as soon as one step further back, peaks and tough more then 10dB started to appear. And since no 2 spots are the same it does makes me wonder why bother at all.

By the way, anyone could care to suggest a good for this subject for a novice?

Ah.... just notice your response Titian. I think you have to let Oedipus style of expression pass. Otherwise we could end up have a 'hot' debate that goes in circles......... still interesting but it wouldn't give any useful suggestions to anyone.

 

[titian]

Oedipus agrees with you

Not sure about that but I agree to a lot what he writes.

..and I think I am beginning to understand why some people suggest it might be no bad thing buying studio loudspeakers and I might just go and listen to a few when I have the opportunity to start all over again.

What are your goals? To listen to flat response frequency music or to something that you like to listen to? Maybe in your case it is the same maybe not. In any case I suggest you to at least test some studio monitors.

When trying to understand basic room frequency response of my loudspeakers by doing some (my own amatuer measurements using laptop & SPL meter) in my own living room it does raise more questions then answers. The speakers FR appear flattish as long as it is near field (less then 2m) but as soon as one step further back, peaks and tough more then 10dB started to appear. And since no 2 spots are the same it does makes me wonder why bother at all.

For a lot of people is 2m not anymore near field.
Now tell me what is the use of flat response speakers if at the listening point the room causes so big peaks? When you go to listen to speakers just immagine how many peaks you have there which also differs to each show room. Without speaking about the electronics which are driving the speakers!
The interesting point about me is that I have choosen the speakers without looking at their technical specs. Only years after talking to the developer he told me that they had (of course) flat frequency response and a few satellites are used by some recording engineers (at least in San Francisco) as monitors. But again today there are so many speakers having these characteristics, even the very cheap Nubert, I believe, for a few hundreds €. If they all sound the same, I doubt, so there must be other things which have to be considered. If they sound all the same for you, then buy the cheapest or those which you, or your wife, like (look) the best.

Ah.... just notice your response Titian. I think you have to let Oedipus style of expression pass. Otherwise we could end up have a 'hot' debate that goes in circles......... still interesting but it wouldn't give any useful suggestions to anyone.

don't worry, no debate will go in circles from my side. I would just retire.

 

[LinearMan]

The interesting point about me is that I have choosen the speakers without looking at their technical specs.

Quite right too, my point exactly!

As I alluded to earlier, I belive that many 'enthusiasts' are in danger of missing the point, the point being to listen and have an emotional response to music, not to tryto find objective measurements of a subjective experience. OK, all this response stuff is interesting, but if your goal is music then it's largely irrelevant. In business there is a description that could well apply here ~ analysis paralysis.

 

[Tenson]

I have to say in my case I like a system more if it has a flat response and especially if it has good acoustics in terms of reverb time and diffusion. The bass has to go low if it has a flat response though or it sounds thin.

When I say the 'system' has a flat response I mean from the listening position, not just the speakers.

 

[oedipus]

So it's not only speakers

The room only makes matters worse. If you start with bad speakers, the room will - almost without exception - make matters worse.

This is also in concerts halls so and it has nothing to do with crap sound.

Concert venues are designed to reverberate for music production. Music reproduction is a different matter,

You're not telling me that you cannot divide 5 by 2?

It's patently obvious that I can. What I was trying to find out was whether you were failing to multiply by two in order to make your speakers sound better than they actually are

why not, that is a way to solve the problem. If I have a flat frequency speaker system and my room causes a probem, I prefer solving it where the cause comes from.

Well then just build a proper room

I really don't envy you and the persons near you

Again I don't envy your personality.

Well that is what you expect. You poor guy.

And again I don't envy you poor guy.

Envy and poverty (ie being poor); that's what your system seems to be about. You want people to envy you because they can't indulge the money you've (foolishly) spent? Well, I don't envy what appears to be (by your own account) a severe acoustical problem with a system that cost you so much money.

 

[Tenson]

Even with a purpose designed room, treatment is still necessary. The shape, size and materials of construction only go so far (which is quite a way!). The treatment is just designed into the room at the best places so it isn't obvious and works as well as possible.

Titan do I understand from your post that your system has a in-room response that varies no more than 7dB? Or is it -/+7dB? Either way thats very good! Though +7dB on top of that at 30Hz does sound pretty bad but at such a low frequency and high Q it could be fixed easily with digital correction.