A couple of things up front: I'm no audiophile and I'm new to building my own speakers. With those caveats in mind, can someone answer a question that's been bugging me since I started researching DIY speaker building: why is everyone striving for flat freqquency responses, since that isn't how our hearing works?
My wonder got quite a boost last night. I was invited to hear a very expensive, highly tuned system. There were enough Sunfire amps in the room to power a large cruiseliner. The owner had built the speakers himself out of the finest components he could find. Then he spent days and days tweaking them until his response curve was remarkably flat. He even took them to a recording studio in town for testing. I was impressed with his data. Then he played them for me. They were about the worst sounding speakers I've ever heard. Harsh and shrill to an extreme. In fact, they had a treble boost that I feared was going to peel the paint from the walls, if it didn't burn the skin from my body first. I'm glad the guy liked them, but I thought they were hideous.
So I'm asking, do people really like that sound? We don't have flat hearing curves, and it seems odd to me that we would want our speakers to emit sounds that are artifically boosted. It has been a long, long time since I fooled with any audio gear, but back in the early `70s when I was first interested in sound systems, the idea was to match your speakers' response with your hearing. Can't remember the specs exactly, but that required a pretty significant dip in the midrange, as much as 8-10 db at high SPLs.
So what's the deal? After a lifetime of open exhaust race cars and rock 'n roll, I know it's not because I have good high freq hearing.
My wonder got quite a boost last night. I was invited to hear a very expensive, highly tuned system. There were enough Sunfire amps in the room to power a large cruiseliner. The owner had built the speakers himself out of the finest components he could find. Then he spent days and days tweaking them until his response curve was remarkably flat. He even took them to a recording studio in town for testing. I was impressed with his data. Then he played them for me. They were about the worst sounding speakers I've ever heard. Harsh and shrill to an extreme. In fact, they had a treble boost that I feared was going to peel the paint from the walls, if it didn't burn the skin from my body first. I'm glad the guy liked them, but I thought they were hideous.
So I'm asking, do people really like that sound? We don't have flat hearing curves, and it seems odd to me that we would want our speakers to emit sounds that are artifically boosted. It has been a long, long time since I fooled with any audio gear, but back in the early `70s when I was first interested in sound systems, the idea was to match your speakers' response with your hearing. Can't remember the specs exactly, but that required a pretty significant dip in the midrange, as much as 8-10 db at high SPLs.
So what's the deal? After a lifetime of open exhaust race cars and rock 'n roll, I know it's not because I have good high freq hearing.
I think the emphasis will make more sense if you start by recognizing this fact:
What our speakers reproduce and what we hear are two different things.
So, it's important for speaker frequency response to be flat because we want our speakers to reproduce sound exactly the way it was originally produced. That doesn't have anything to do with how sensitive our ears are to certain frequencies, or whether we having high frequency hearing damage.
Think of it this way, if you were to attend a music event live and that same event was also being recorded. You would listen during the event and the sound you are precieving and the sound that is actually occuring is different because, as you stated, our ears do not have a flat frequency response. But the sound that is occuring and the sound that is recorded will be very similar, and this is important. When our speakers reproduce that sound we want them to accurately reproduce what was recorded (and therefore accurately reproduce the sound the actually occured) so that we can precieve the sound the same way as when we heard it live.
I hope this helps clear things up, but if not, just know this. The fact that our ears do not have flat frequency response does not rid use of the need for speakers that have flat frequency response.
What our speakers reproduce and what we hear are two different things.
So, it's important for speaker frequency response to be flat because we want our speakers to reproduce sound exactly the way it was originally produced. That doesn't have anything to do with how sensitive our ears are to certain frequencies, or whether we having high frequency hearing damage.
Think of it this way, if you were to attend a music event live and that same event was also being recorded. You would listen during the event and the sound you are precieving and the sound that is actually occuring is different because, as you stated, our ears do not have a flat frequency response. But the sound that is occuring and the sound that is recorded will be very similar, and this is important. When our speakers reproduce that sound we want them to accurately reproduce what was recorded (and therefore accurately reproduce the sound the actually occured) so that we can precieve the sound the same way as when we heard it live.
I hope this helps clear things up, but if not, just know this. The fact that our ears do not have flat frequency response does not rid use of the need for speakers that have flat frequency response.
Without arguing over whether 'flat response' (what do you mean by this, flat response on axis, off axis, flat power response?) is desireable or not, have a look at the articles below to understand why many promote a flat power response:
http://www.reed-electronics.com/tmworld/article/CA475937
http://www.stereophile.com/interviews/231/index.html
http://www.stereophile.com/news/10705/
http://www.moultonlabs.com/articles/94startover2.htm
Basically as far as I can tell, the focus on flat power response (NOT the same as most people's conception of 'flat response') is all traceable back to Floyd Toole's excellent research at the NRC in Canada.
http://www.reed-electronics.com/tmworld/article/CA475937
http://www.stereophile.com/interviews/231/index.html
http://www.stereophile.com/news/10705/
http://www.moultonlabs.com/articles/94startover2.htm
Basically as far as I can tell, the focus on flat power response (NOT the same as most people's conception of 'flat response') is all traceable back to Floyd Toole's excellent research at the NRC in Canada.
You are assuming that "harsh sound" is a result of "flat response" when in fact the quality of his speakers or lack thereof may be stemming from a totally different issue. I bet you could take a very nice, expensive, flat ribbon tweeter and cross it too low and it will happily peel the paint from your walls.
It sound like your friend needs to treat his room. Even the best speakers will sound bad in a bad room. Also, for measurements to be truly valid they need to be done at the seating position in the room they are used.
The idea of flat response is that it is necessary to perfectly reproduce what is on the recording, which is a valid point. However, just because it's flat doesn't necessarily mean that it sounds good. It just means that any frequency is played with equal SPL. That doesn't mean that flat response equals bad sound. Yes, our ears have different sensitivities to different frequencies, but that doesn't have anything to do with whether or not a speaker's FR should be flat. A flat FR is a perfectionist's goal in designing and building speakers.
Any dipole fan like myself will say that flat response is overrated because with dipoles a flat response is impossible, yet when done right they sound great, at least to us.
The idea of flat response is that it is necessary to perfectly reproduce what is on the recording, which is a valid point. However, just because it's flat doesn't necessarily mean that it sounds good. It just means that any frequency is played with equal SPL. That doesn't mean that flat response equals bad sound. Yes, our ears have different sensitivities to different frequencies, but that doesn't have anything to do with whether or not a speaker's FR should be flat. A flat FR is a perfectionist's goal in designing and building speakers.
Any dipole fan like myself will say that flat response is overrated because with dipoles a flat response is impossible, yet when done right they sound great, at least to us.
Re: Re: Why the emphasis on flat speaker response?
Thanks for your gentlemanly response. But you're wrong. I looked at the results of his in-room testing, after the room had been tweaked. Flat as a pancake. So, something was up: the test equipment might not have been properly calibrated, or maybe my ears are out of whack. Whatever the reason, I couldn't live with his speakers for more than a few seconds.
Bill Fitzpatrick said:
Thanks for your gentlemanly response. But you're wrong. I looked at the results of his in-room testing, after the room had been tweaked. Flat as a pancake. So, something was up: the test equipment might not have been properly calibrated, or maybe my ears are out of whack. Whatever the reason, I couldn't live with his speakers for more than a few seconds.
If you design for flat on axis response and disregard all else, I'm sure your system will sound quite bad.
If as you say you've sustained some hearing damage over the years, your high frequency hearing would likely be serverly compromised. Which is what makes it so puzzling to me that you heard a 'treble boost', which as Bill rather bluntly pointed out, is impossible in a 'flat' (btw, you still haven't defined this in the context of your friends stereo) system. Maybe the treble sounds shrill, harsh, or just plain wrong, but if the system measures flat, there is by definition no 'treble boost', at least not on the measurement axis.
If as you say you've sustained some hearing damage over the years, your high frequency hearing would likely be serverly compromised. Which is what makes it so puzzling to me that you heard a 'treble boost', which as Bill rather bluntly pointed out, is impossible in a 'flat' (btw, you still haven't defined this in the context of your friends stereo) system. Maybe the treble sounds shrill, harsh, or just plain wrong, but if the system measures flat, there is by definition no 'treble boost', at least not on the measurement axis.
Bah.
I bet the speakers and room and all that were measured with pink noise or warble tones or something.
Were you listening to pink noise when your ears wanted to crawl into a corner and hide? Or were you perhaps listening to some type of music from some source which I might guess in this case was a CD (perhaps in the "Rock" genre? - certainly, probably "modern" music of some type) through all solid state components. Many CD's are very poorly mastered and are excessively hot - accurate speakers will reproduce this very faithfully! All solid state does nothing to help this fact.
C
I bet the speakers and room and all that were measured with pink noise or warble tones or something.
Were you listening to pink noise when your ears wanted to crawl into a corner and hide? Or were you perhaps listening to some type of music from some source which I might guess in this case was a CD (perhaps in the "Rock" genre? - certainly, probably "modern" music of some type) through all solid state components. Many CD's are very poorly mastered and are excessively hot - accurate speakers will reproduce this very faithfully! All solid state does nothing to help this fact.
C
As would I! I am a young guy, but still curious how I'd do. I did convince the Occupational Health department here to let me try their test, but its intended for something else, and only goes up to 16khz. So I know I can hear up to there, but then I already knew that since the flyback transformer 'whine' from old TVs drivers me nuts. I'd LOVE to have a fletcher-munson type curve of my own hearing plotted.
BTW, if you look at some older Yamaha recievers, they had a variable loudness feature that would apply a boost to the lows and highs very similar to the fletcher-munson curve. This boost was variable, almost nil at high volume levels, and greatly attenuated when the volume was low. I have one of these at home, and it works quite well IMO, if a bit exaggerated. I always thought such a setup would work very well if properly implemented.
I think he's talking about this situation as described by Debertin:
"Overwhelming Highs. Of course, there was a lot of emphasis on projects that when tested using computer listening software produced as close as possible to a flat response curve. This leads to a lot of designs whereby the overall frequency balance to my ears is tilted upwards. These speakers may be what the computer finds as flat response, but to my ears they often sounded overly bright and irritatingly harsh. These are the kinds of speaker that sound realistic for ten minutes of listening. After that they are simply grating and irritating. My ears were physically hurting for two days after doing these auditions. The question I kept asking myself is how could I possibly enjoy listening to such speakers over the long term in my home. Along with this seemed to be a desire on the part of builders to not provide any high frequency level control to compensate for what are obvious differences in various rooms’ ability to absorb sound. I find this curious at best"
Scroll down to BACKGROUND;
http://members.aol.com/_ht_a/Debertin/spbuild.htm
"Overwhelming Highs. Of course, there was a lot of emphasis on projects that when tested using computer listening software produced as close as possible to a flat response curve. This leads to a lot of designs whereby the overall frequency balance to my ears is tilted upwards. These speakers may be what the computer finds as flat response, but to my ears they often sounded overly bright and irritatingly harsh. These are the kinds of speaker that sound realistic for ten minutes of listening. After that they are simply grating and irritating. My ears were physically hurting for two days after doing these auditions. The question I kept asking myself is how could I possibly enjoy listening to such speakers over the long term in my home. Along with this seemed to be a desire on the part of builders to not provide any high frequency level control to compensate for what are obvious differences in various rooms’ ability to absorb sound. I find this curious at best"
Scroll down to BACKGROUND;
http://members.aol.com/_ht_a/Debertin/spbuild.htm
The Vifa D25 tweeter measured reasonably flat and in the systems I have designed, no matter how much I attenuated the tweeter in relation to the woofer, the treble sounded harsh and irritating to my ears.
As alluded to before, a flat measurement can hide a qualitative deficiency in a driver which the ear recognises only too quickly.
As alluded to before, a flat measurement can hide a qualitative deficiency in a driver which the ear recognises only too quickly.
Problem with projects like that is that the flat response is raised as a goal in itself no matter what. Typically brute force is used to make the response flat - resonances notched, dips raised, etc.
The measurements are done with slow sweep of pure sinewave.
What this doesn't show is dynamic distortions, intermodulations, garbage and ringing that gets additionally produced.
Thats the only way I can see why flat speakers sound awful.
Check this: http://www.linkwitzlab.com/mid_dist.htm
The measurements are done with slow sweep of pure sinewave.
What this doesn't show is dynamic distortions, intermodulations, garbage and ringing that gets additionally produced.
Thats the only way I can see why flat speakers sound awful.
Check this: http://www.linkwitzlab.com/mid_dist.htm
wimms said:
This last statement just isn't true. Flat response ONLY means equal SPL output form equal power input accross the frequency spectrum. It does not necessarily indicate sound quality.
A perfect example:
I use some cheap but highly efficient Pyle bullet tweeters as super tweets in my line arrays. These things are god awful out of the box, just ear bleedingly harsh at any volume. I just stuff a few pieces of foam rubber down in the throat and all of the harshness disappears and the detail comes out. They're quite nice after tweaking with very little loss of sensitivity. It's a night and day difference in sound quality with negligble effect on the Frequency Response.
Get your sound quality first, then try to flatten response to the extent you want, but without sacrificing SQ. If a flat FR is your only consideration, you're doomed from the start or simply lucky.
I suppose the other two important quantitative measurements you need to look at other than frequency response are linear and non-linear distortion. FR tells you how accurate your speakers are at reproducing frequecies for a particular voltage input with respect to SPL, but it doesn't tell you anything about how much the speaker "smears" certain frequencies (that is, continues to resonant even after the input is taken away) or what other frequencies are produced when you attempt to play a particular frequency. All of these are measures of accuracy and none of them by themselves can give you a good idea for how a speaker sounds.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.