The curves traditionally called "Fletcher-Munson" are sort of a lab curiosity. Just think for a moment how you would establish your F-M level with music of varying loudness?
I do like bass boost when playing music softly late at night. But something of an intellectual misunderstanding to think our hearing needs "F-M correcting" like some physical device. This is similar to what is known as the "El Greco fallacy."
Better to simply start with the observation that boosting the bass makes softly played music sound better. Then pick a slope and turn-over point (F-M middle curves are as good as any). Then, if you think it is worth the fuss, make the slope variable with loudness. Then, key the curves to how it sounds to you at different levels (there's no way to do it with a sound level meter).
We don't listen critically when music is played softly. We're just happy to have music in the middle of the night and not wake the kids or neighbours.
Ben
I do like bass boost when playing music softly late at night. But something of an intellectual misunderstanding to think our hearing needs "F-M correcting" like some physical device. This is similar to what is known as the "El Greco fallacy."
Better to simply start with the observation that boosting the bass makes softly played music sound better. Then pick a slope and turn-over point (F-M middle curves are as good as any). Then, if you think it is worth the fuss, make the slope variable with loudness. Then, key the curves to how it sounds to you at different levels (there's no way to do it with a sound level meter).
We don't listen critically when music is played softly. We're just happy to have music in the middle of the night and not wake the kids or neighbours.
Ben
Last edited:
Good day, eh?* I too am a Ben but here I am the Soldermizer! You make some good points...
I'm not familiar with the "El Greco" fallacy but a quick web search gives me the idea it has something to do with incorrectly explaining a misperception. Is this like explaining a mirage is actually not water, but a volatile liquid that evaporates just before the thirsty traveler will reach it?**
While it would be nice to automate the whole process (or would it? You might spend much time tweaking only to find the best solution may be...) or should we just do it manually, adjusting levels depending on our listening volume. In a way, I already do that with my "1/3 octave subjectively flat pink noise" EQ. If you listen at an "average" volume, you will soon dial in to your own optimal spot. After much listening and testing, I have settled on a curve that works fine for me. The only downside I see is if you like to blast once in a while, as I do: then your boosts of low (and high?) will be a bit much for high dB levels. To each his own.
*We yanks do so like to make fun of you Canadians. For your entertainment in return, I offer our political system
** I think I will go spill random puddles of [insert quick-evaporating solvent name here] on the highway to f**k with people's minds.
I'm not familiar with the "El Greco" fallacy but a quick web search gives me the idea it has something to do with incorrectly explaining a misperception. Is this like explaining a mirage is actually not water, but a volatile liquid that evaporates just before the thirsty traveler will reach it?**
While it would be nice to automate the whole process (or would it? You might spend much time tweaking only to find the best solution may be...) or should we just do it manually, adjusting levels depending on our listening volume. In a way, I already do that with my "1/3 octave subjectively flat pink noise" EQ. If you listen at an "average" volume, you will soon dial in to your own optimal spot. After much listening and testing, I have settled on a curve that works fine for me. The only downside I see is if you like to blast once in a while, as I do: then your boosts of low (and high?) will be a bit much for high dB levels. To each his own.
*We yanks do so like to make fun of you Canadians. For your entertainment in return, I offer our political system
** I think I will go spill random puddles of [insert quick-evaporating solvent name here] on the highway to f**k with people's minds.
Last edited:
If El Greco had astigmatism, his stretchy paintings would look distorted to him as he painted them and he would have made them "accurate" (if he wanted to) as he went along. So they look oddly long because he wanted them to look that way, not because eyeglasses (or F-M correction) would have made any difference, astigmatism or not.
Whenever ear versus mic results are discussed at DIYaudio, it ends up that even the most die-hard measurement types (and I fall mostly in that direction) agree that, like Soldermizer, folks prefer a "house curve" with extra oomph in the bass... no matter what the mic says.
I'd say somebody could (and likely has) produced nested F-M-like curves (AKA "house curves") for home music rooms.
Those F-M curves do reflect a reality of hearing. They might be good to use if you listened to pure tones a lot. Otherwise, just EQ as you please.
Ben
Whenever ear versus mic results are discussed at DIYaudio, it ends up that even the most die-hard measurement types (and I fall mostly in that direction) agree that, like Soldermizer, folks prefer a "house curve" with extra oomph in the bass... no matter what the mic says.
I'd say somebody could (and likely has) produced nested F-M-like curves (AKA "house curves") for home music rooms.
Those F-M curves do reflect a reality of hearing. They might be good to use if you listened to pure tones a lot. Otherwise, just EQ as you please.
Ben
Last edited:
Ben
Thank you for posting.
I understand your reservations on the validity of any loudness correction scheme
e.g.
George
Thank you for posting.
I understand your reservations on the validity of any loudness correction scheme
e.g.
Nevertheless, do you happen to know any paper discussing loudness correction contours based on complex –music- signal as opposed to F-M which are based on pure tones?
George
Possibly a good question. Sorry, I am not current on that kind of research.
I'm not sure how the successors to F-M do their research - darn clever, these psychologists, if I may say so myself.
But the original studies sat people down and played a 1000 Hz note. Then they said, "I am going to play a 80 Hz note and you tell me if it is as loud as the 1000 Hz note." That very roughly is how it was done.
Yes, even with screwball instructions like that, you can create stable, repeatable, representative (AKA scientific) curves at different loudness levels of the 1000 Hz note. (And probably somewhat different curves if you used square waves, 1/3 octave white noise, or music.) But does that mean anything to you in your music room?
I know this will disappoint people with an engineering soul, but the "house curve" is the curve you think has good balance or is otherwise satisfying to you.
I'd be curious if representative house curves (maybe even the famous movie-house curves.... which just got adjusted) resemble F-M.
Why do you accuse me of objecting to ALL loudness compensation when I said I use bass boost several times in earlier posts? In fact, I use one of the glorious pre-amp of yesteryear, Kenwood Basic2. Unlike Basic1 which was a wire-with-gain (that I also use), Basic2 is complicated and has the four-pot F-M-like loudness compensating volume control. Moreover, I've done my "gain management" chain so that Kenwood's idea of compensation tracks-and-suits my ears OK. Yup, I often use loudness compensation late at night.
Ben
Last edited:
Assuming somebody could actually do what you describe and get stable results, what could be wrong with that as a means of producing equal "volume" of sound? Seems a suitable method to get equal volume in real-room tests. Unlike the false belief that a mic can produce that curve.
However, just because you achieved a kind of textbook/rational harmony of volume among your room, system, and ears (at least at one loudness level), it doesn't mean that you've achieved the sound balance that makes great listening to the most recordings in your setting even at that loudness level.
That's a different question.
In practice, first with a mic, then with ears, perhaps as Soldermizer has conscientiously done appropriately, then inevitably, tweaking with music.
Ben
Last edited:
There's something in your method (or something similar) that represents the "missing link" between the machinery (and physical measurements) and actual human hearing. Most people just listen to music and say, "Ummm, needs to be a little brighter" rather than use an articulated method like yours.
So I wonder what elaborations of your method would be improvements?
For example.... I've often wondered if there is a kind of orchestral-music random noise (perhaps in the traditional 1/3 octave bands) that reflects the statistics of orchestral music better than pink noise does? Why doesn't somebody make that test signal?
Ben
Hi,
If you want a spanner in the works I can throw you several :
a) The FM curves are equal loudness related to fixed levels
at 1KHz. Whilst it is clear in those curves response to levels
varies at different frequencies, by necessity they include the
ludicrous assumption @ 1KHz ear response to relative levels
is perfectly linearly logarithmic.
b) So FM curves don't tell you how loud it actually is, and
annoyingly I've never seen an attempt to quantify the
perception of different levels @ 1KHz.
c) If your confused say move the the reference point to
3KHz and assumed that is perfectly linearly logarithmic.
The curves will change but won't tell you any more *.
d) IMO you cannot assume any frequency is in any
way is perfectly linearly logarithmic, much to my
frustration in the way F&M information is presented.
e) So if you have twigged F&M curves show equal
loudness curves relative to completely arbitrary
levels at 1KHz, they don't tell you how loud it is.
rgds, sreten.
* But wiil change what you can read into the spacing.
If you want a spanner in the works I can throw you several :
a) The FM curves are equal loudness related to fixed levels
at 1KHz. Whilst it is clear in those curves response to levels
varies at different frequencies, by necessity they include the
ludicrous assumption @ 1KHz ear response to relative levels
is perfectly linearly logarithmic.
b) So FM curves don't tell you how loud it actually is, and
annoyingly I've never seen an attempt to quantify the
perception of different levels @ 1KHz.
c) If your confused say move the the reference point to
3KHz and assumed that is perfectly linearly logarithmic.
The curves will change but won't tell you any more *.
d) IMO you cannot assume any frequency is in any
way is perfectly linearly logarithmic, much to my
frustration in the way F&M information is presented.
e) So if you have twigged F&M curves show equal
loudness curves relative to completely arbitrary
levels at 1KHz, they don't tell you how loud it is.
rgds, sreten.
* But wiil change what you can read into the spacing.
Last edited:
80 year update!
ISO 226:2003 (/2008)
Acoustics -- Normal equal-loudness-level contours
Abstract
This International Standard specifies combinations of sound pressure levels and frequencies of pure continuous tones which are perceived as equally loud by human listeners. The specifications are based on the following conditions: the sound field in the absence of the listener consists of a free progressive plane wave; the source of sound is directly in front of the listener; the sound signals are pure tones; the sound pressure level is measured at the position where the centre of the listener's head would be, but in the absence of the listener; listening is binaural; the listeners are otologically normal persons in the age range from 18 years to 25 years inclusive.
See this great thread at ... the grumpy place:
http://www.diyhifi.org/forums/viewtopic.php?f=11&t=2289&p=47818&hilit=fletcher+munson+curves#p47818
ISO 226:2003 (/2008)
Acoustics -- Normal equal-loudness-level contours
Abstract
This International Standard specifies combinations of sound pressure levels and frequencies of pure continuous tones which are perceived as equally loud by human listeners. The specifications are based on the following conditions: the sound field in the absence of the listener consists of a free progressive plane wave; the source of sound is directly in front of the listener; the sound signals are pure tones; the sound pressure level is measured at the position where the centre of the listener's head would be, but in the absence of the listener; listening is binaural; the listeners are otologically normal persons in the age range from 18 years to 25 years inclusive.
See this great thread at ... the grumpy place:
http://www.diyhifi.org/forums/viewtopic.php?f=11&t=2289&p=47818&hilit=fletcher+munson+curves#p47818
Dear Dev Team
We're patient, but it's now been three and a half years . .
Where does this very popular, not much work plug-in fit in your plans?
Thanks
Otto
If I were miniDSP I would put this at the bottom of my to-do list.
And I don't think it's as simple to implement as everyone might think. You're talking about a continuous level-sensing EQ that modifies the shelving actively at various levels. Plus, you'd need to have the capability to threshold "calibrate" it some way for various speaker efficiency, amplifier voltage gain, source level, etc, combinations.
This is not trivial....IMHO.
A user could easily set up a simple shelving filter with the existing hardware and switch it on/off manually. This would be much the same as a traditional "loudness" switch/control that might have been found on equipment from decades past. I think that would achieve 90% of the goal here.....IMO.
Some of the other DSP platforms have forms of "dynamic" EQ, although I don't like the way those work and it's not, strictly speaking, the same as the request here.
Cheers,
Dave.
And I don't think it's as simple to implement as everyone might think. You're talking about a continuous level-sensing EQ that modifies the shelving actively at various levels. Plus, you'd need to have the capability to threshold "calibrate" it some way for various speaker efficiency, amplifier voltage gain, source level, etc, combinations.
This is not trivial....IMHO.
A user could easily set up a simple shelving filter with the existing hardware and switch it on/off manually. This would be much the same as a traditional "loudness" switch/control that might have been found on equipment from decades past. I think that would achieve 90% of the goal here.....IMO.
Some of the other DSP platforms have forms of "dynamic" EQ, although I don't like the way those work and it's not, strictly speaking, the same as the request here.
Cheers,
Dave.
Last edited:
I'm with Davey... I mean, it would be very cool to have some sort of Labview-like DSP programming visual language to enable these kinds of things, but I can only imagine the development and support costs (having been there, somewhat).
If you want different EQ configurations, just set them up with one of the platforms that allows it and switch with a remote. You don't need infinitely-variable loudness contours, two or three would do it, surely?
If you want different EQ configurations, just set them up with one of the platforms that allows it and switch with a remote. You don't need infinitely-variable loudness contours, two or three would do it, surely?
Yes, I now see that infinitely-variable is a lot of complexity.
2-3 would help. Better than one!
I was going to use one setting for a mild midbass hump.
It'd be nice to have a setting with both a midbass hump, and (for listening at lower volumes) a "loudness contour".
of course the "standard" setting
And I have a *lot of difficulty with the intelligibility of dialog. A setting that emphasises consonants, about 2-5 kHz - I would find that very useful for anything spoken (and possibly song lyrics) etc
I'm no EE, but wonder what part of the miniDSP limits the number of settings currently to a total of four. Memory?
What's most important is to get an "excellent"/ "smooth" crossover.
But then, to be able to tune it to say 3 different sorts of material/ volume?
That would be icing on the cake. I'd be happy to pay some extra, for some extra flexibility.
"Loudness" compensation in playback equipment has always puzzled me since I think the ear/brain adjusts for this automatically. When I go to the Seattle Symphony does the orchestra need to "equalize" their sound when playing softly vice full force? Everything we hear in nature is not equalized depending upon absolute SPL.
Dave.
Everything we hear in nature is not equalized depending upon absolute SPL.Dave.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.