amitie madelin

Hello,
This semester I am in an academic research writing class in which I am basing my paper on the importance of flat frequency response in a typical home environment. Now, defining your typical home environment or at least the typical homes I know to have rooms with stereos and home theaters ranging from 10'x12' to 20'x20' with ceilings between 8' and 10', I would like to have good evidence in both Directions for, and against why flat frequency response is either a must, or bust.

John K. Site discusses radiation patterns and power response quite nicely, and possibly shows how flat response should not be a single goal. http://www.musicanddesign.com/Power.html

I am looking for information and sources such as this. I will use the LDC, Speakerbuilding 101, D'Appolito's book on measuring and Tool's book on loudspeakers in rooms.

Opinions do not really count (Well I will take some from the big guns such as Jeff B., DLR, John K, Curt C....) As designer we have many measurements and a host of data that has to be considered and we all choose our methods and weigh what is most important a little differently. I am looking for some data, and analysis of that data. Quality and validity count, i know there are many sound minded individuals here who have found many good references out there around this topic. Any help is appreciated in advance.

You're looking for proof on a subjective topic, IMO -- which apparently isn't worth much.

In a studio, flat is the goal. In a home, there's no right answer. It's just preference. The scores of not-flat playback equipment should be adequate proof that "importance of flat response" is in the ear of the beholder.

You're looking for proof on a subjective topic, IMO -- which apparently isn't worth much.

In a studio, flat is the goal. In a home, there's no right answer. It's just preference. The scores of not-flat playback equipment should be adequate proof that "importance of flat response" is in the ear of the beholder.

Not proof over a subjective subject, but some reasonably proposed data and well grounded explanations such as the example I gave.

Human hearing is not flat.. http://www.webervst.com/fm.htm so a source with ultra linear frequency response will not sound pleasing to most people and this could be considered a detractor to your theory.

Human hearing is not flat.. http://www.webervst.com/fm.htm so a source with ultra linear frequency response will not sound pleasing to most people and this could be considered a detractor to your theory. That doesn't jive with all the testing data performed over the years by Harmon.

Our hearing may not be "flat" but that doesn't mean the reproduction equipment shouldn't be. A violin in the room produces a spectrum that if reproduced by a non-flat system, will not sound like the violin. Harmon's research shows that when given a choice between a number of systems, trained and non-trained listeners alike tend to prefer the systems with the flattest overall response.

Human hearing is not flat.. http://www.webervst.com/fm.htm so a source with ultra linear frequency response will not sound pleasing to most people and this could be considered a detractor to your theory. I so much wish that would quit being stated. Equal loudness curves have nothing to do with whether a speaker should have flat response or not. Those curves represent the way we hear EVERYTHING not just the sound from our speakers.

I think I get what Jeff and Pete are saying. Because our ears don't hear frequencies in a flat line, then NOTHING we hear is perfectly flat. For instance, if we were to play an orchestra from 20-20KHz, note by note, or even the sine wave, how our ears perceive that is not perfectly flat. BUT, in order for our speakers to DUPLICATE the sound as we hear it, it MUST be flat, because if it wasn't flat, then a G wouldn't sound like a G, it would be altered since our ears AREN'T flat. Is that right?

I think I get what Jeff and Pete are saying. Because our ears don't hear frequencies in a flat line, then NOTHING we hear is perfectly flat. For instance, if we were to play an orchestra from 20-20KHz, note by note, or even the sine wave, how our ears perceive that is not perfectly flat. BUT, in order for our speakers to DUPLICATE the sound as we hear it, it MUST be flat, because if it wasn't flat, then a G wouldn't sound like a G, it would be altered since our ears AREN'T flat. Is that right?

The G would still sound like a G, just not necessarily at the same amplitude.

Using the violin example, if it's playing a 440Hz fundamental, there's also 880, 1320, 1760, 2200, etc. Same thing for a trumpet playing the same note. It's the relative levels of each of those harmonics that gives the characteristic and recognizable sound that allows our brains to differentiate the single trumpet from among the dozens of violins playing the same note. If the playback system didn't allow those harmonics to maintain the original relative levels, the trumpet might sound like a french horn or trombone instead of a trumpet.

Additive synthesis takes advantage of this to emulate the sound of real-world sources without having to sample audio data. See: MIDI from the 90s.

I hear a lot about studies that say "people prefer flat speakers", and then I hear "people prefer about +3 to +6dB of bass", and then I also hear "people prefer a slightly downward-angled slope". I wonder if those are the same people.

SPL was what I meant. When I said the G is altered, I meant because the amplitude is lower or higher (I just couldn't think of the right words), the sound is altered overall. This is because we loose the proper level of harmonics that are correct to our ears as we perceive them normally, when speakers are not designed to be flat, because other nearby harmonics are either louder, or lower and therefore distort what we hear. I think I am saying that right for what I mean. Ugh I need to go to bed... LOL

Additive synthesis takes advantage of this to emulate the sound of real-world sources without having to sample audio data. See: MIDI from the 90s.

I hear a lot about studies that say "people prefer flat speakers", and then I hear "people prefer about +3 to +6dB of bass", and then I also hear "people prefer a slightly downward-angled slope". I wonder if those are the same people.

When the sighted/confirmation bias is removed, all kinds of dearly held preferences and dogma go bye-bye. One of the more interesting publicly published round of testing Harmon did was when they built a garbage speaker into a high end looking cabinet and had the listeners compare it to a good speaker in a garbage looking cabinet. They they blinded the listeners. I believe it can be dug out of Sean Olive's blog somewhere, but the results strongly resembled the "brownie tastes better served on fine china" effect...

Similarly, you can fool wine experts by swapping price-tags around. Any bar owner will tell you the advantages of having a high priced beer and a low priced beer on the menu when you want to sell a ton of the middle priced beer. What does it all mean? I don't know - but the evidence about our ability to verify our stated preferences is slim to none and slim just saddled up and is headed out of town.

When the sighted/confirmation bias is removed, all kinds of dearly held preferences and dogma go bye-bye. One of the more interesting publicly published round of testing Harmon did was when they built a garbage speaker into a high end looking cabinet and had the listeners compare it to a good speaker in a garbage looking cabinet. They they blinded the listeners. I believe it can be dug out of Sean Olive's blog somewhere, but the results strongly resembled the "brownie tastes better served on fine china" effect...

Similarly, you can fool wine experts by swapping price-tags around. Any bar owner will tell you the advantages of having a high priced beer and a low priced beer on the menu when you want to sell a ton of the middle priced beer. What does it all mean? I don't know - but the evidence about our ability to verify our stated preferences is slim to none and slim just saddled up and is headed out of town.

The difference in the Harmon studies was that the audience didn't know what they were listening to. The speakers were not visible, only audible. And the statistical result showed a significant preference for flat response.

Now, do we like to turn up the bass knob? Sure, when the recording is thin. And because we don't control the mastering process, we are given tone controls to adjust to preference.

Personally, I need to crank the bass quite a bit listening to old Rush albums, heck, MOST of the vintage classic rock. The engineers in the booth at the time must have mixed it thinking that the end user would have the bass knob cranked all the way up, so they purposely cut the bass to be able to raise the average recording level, a type of compression.

Anymore though, I find myself listening to music for the pure enjoyment of a good recording. A good artist that has come across the server lately is Fink, someone I'd never heard of until DoubleTap saw his name as a credit on Walking Dead. No need for any tone controls as the overall balance from low to high is superb. That and the music is fantastic!!!

The difference in the Harmon studies was that the audience didn't know what they were listening to. The speakers were not visible, only audible. And the statistical result showed a significant preference for flat response.

Now, do we like to turn up the bass knob? Sure, when the recording is thin. And because we don't control the mastering process, we are given tone controls to adjust to preference.

Personally, I need to crank the bass quite a bit listening to old Rush albums, heck, MOST of the vintage classic rock. The engineers in the booth at the time must have mixed it thinking that the end user would have the bass knob cranked all the way up, so they purposely cut the bass to be able to raise the average recording level, a type of compression.

Anymore though, I find myself listening to music for the pure enjoyment of a good recording. A good artist that has come across the server lately is Fink, someone I'd never heard of until DoubleTap saw his name as a credit on Walking Dead. No need for any tone controls as the overall balance from low to high is superb. That and the music is fantastic!!!

+1

Adding a Behringer DEQ to my system was one of the best improvements I've ever made to my system. You can readily hear changes in the test noise signal as it auto-EQs with the mic in the listening position and the improvement in SQ was not subtle. The other thing the DEQ does is lets you visualize the recording with it's RTA. I've seen rock recordings with nothing below 200hz and on the other end it's no surprise that some of the music we like to use to demo bass often has a boosted low end. Are the vocals buried in the mix or up front - you can see it with the RTA. I also find that the recordings that sound good to me look flat on the RTA. You can also see where exactly you need to boost the response of those old recordings and toss a little parametric EQ at it.

Ron

I love the brownie on a china plate story. That and many many more studies about perception effecting sense of taste and other things can be found in the book Mindless Eating which I thoroughly enjoyed. I think it tells us as much about speakers and cars and sporting goods equipment as it does about food.

To the topic in general:
Having said that when you strum a guitar and it makes, whatever sound it makes, you're hearing that based on the equal loudness curve of human hearing. This is assuming your standing next to said guitar. Now when something records that guitar do you want it to record it exactly as it sounds or do you want it to record it and then apply the equal loudness curves to what it records. Then when you play back that recording do you want the speaker to have it's frequency response mapped to equal loudness curves? Or do you want it to play back the recorded guitar strum as it came out of the guitar, as you'd hear it when you were standing next to it exactly without changing anything?

I don't understand at all any of the "speakers response should conform to equal loudness curve" talk. It makes no sense at all? Is ruler flat best? Who knows, probably subjective, but flattish is definitely going to be way desirable. It's like saying "The human eye is most sensitive to green, so cameras should be most sensitive to green, then the green pixles on your monitor should be brighter than the red and blue pixels, because it matches the our visual preceptory profile." No, then everything would just be super green. As far as I see it speakers, monitors, cameras and whatever else exist purely to take something in, and then spit it back out as close as possible as to how it went in. That's the way it will most resemble the way it would have looked/sounded if you were there. Obviously there are room interactions and psychoacoustics to deal with, but that's a whole other critter.

Edit: Afterthought, if you wanted something to sound flat it'd have to be totally inverse to the equal loudness curves, and that might sound really bad.
or maybe not... http://en.wikipedia.org/wiki/Smiley_face_curve

When the sighted/confirmation bias is removed, all kinds of dearly held preferences and dogma go bye-bye. One of the more interesting publicly published round of testing Harmon did was when they built a garbage speaker into a high end looking cabinet and had the listeners compare it to a good speaker in a garbage looking cabinet. They they blinded the listeners. I believe it can be dug out of Sean Olive's blog somewhere, but the results strongly resembled the "brownie tastes better served on fine china" effect...

Similarly, you can fool wine experts by swapping price-tags around. Any bar owner will tell you the advantages of having a high priced beer and a low priced beer on the menu when you want to sell a ton of the middle priced beer. What does it all mean? I don't know - but the evidence about our ability to verify our stated preferences is slim to none and slim just saddled up and is headed out of town.

amitie madelin

Wednesday, April 10, 2013

Flat Frequency Response - Techtalk Speaker Building, Audio, Video ...

Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Re: Flat Frequency Response

Thread Information

Users Browsing this Thread

Posting Permissions

No comments:

Post a Comment