LineArray said:
I know Dr. Blauert personally and I have read his book. It had a big influence on my work.
You are right that a CD waveguide would be quite large at LF, but
I do not agree with the need for directivity below 500 Hz. as the room reflections, etc. are not a perceptual factor in that range. You admit as much yourself. But above 500 Hz. and most critically above 1 kHz. smooth controlled directivity is a must.
I have built and used shaded line arrays, but they really don't come close to the control possible with a well designed waveguide.
I have toyed with the idea of a dipole mid bass, but this leads to more design problems, which I am not sure are good tradeoffs. Clearly the dipole is a disaster from a practical point of view at LF. The Summa is a good tradeoff of conflicting constraints IMO.
I declare a following LAW for an obsessed audiophile.
1 shot of liquor=$5000.00 in cables and upgrades.
No, seriously guys. Coming back to the subject.
There’s no substitute for a live performance. I remember traveling to Prague and listening to a live pipe organ performance. There’s just nothing like that. The ambience, the emotions, the sound.
By the same token I remember seating in St.Petersburg Philharmonic and having old babushka sneezing next to me every 5 seconds.
The main thing to remind yourself of: “It’s about the music”
I do like building speakers and it is my hobby. I also enjoy listening to Yasha Hifiz or Paco Delucia. You can’t always go live performance
1 shot of liquor=$5000.00 in cables and upgrades.
No, seriously guys. Coming back to the subject.
There’s no substitute for a live performance. I remember traveling to Prague and listening to a live pipe organ performance. There’s just nothing like that. The ambience, the emotions, the sound.
By the same token I remember seating in St.Petersburg Philharmonic and having old babushka sneezing next to me every 5 seconds.
The main thing to remind yourself of: “It’s about the music”
I do like building speakers and it is my hobby. I also enjoy listening to Yasha Hifiz or Paco Delucia. You can’t always go live performance
The first true eye opening experience I had with hifi was with my first boombox experiment when after having let it run in for a few days took it to a beach party and played music all day and all night while grilling right there on the beach listening to good music in reasonable quality outside in the sun with friends and beers.
My current version of the boombox is vastly better than that first try and is very close to audiophile quality. Sure it could go deeper, sure the treble could be a little more nuanced but it still beats the hell out of most home hifi systems, and you can take it with you ... outside. And listen to it forever and ever.
Oh, and just for the record, yes, I have a hi-end stereo system in the $10,000 range but I these days I actually prefer listening to the boombox, it's just more natural and lifelike in sound.
My current version of the boombox is vastly better than that first try and is very close to audiophile quality. Sure it could go deeper, sure the treble could be a little more nuanced but it still beats the hell out of most home hifi systems, and you can take it with you ... outside. And listen to it forever and ever.
Oh, and just for the record, yes, I have a hi-end stereo system in the $10,000 range but I these days I actually prefer listening to the boombox, it's just more natural and lifelike in sound.
gedlee said:
Magnetar and Linearray to comment on this,
Many speakers designers including Zalph, Jay, Mark and others concentrate on trying to get the drivers distortions linear and non-linear to the lowest level for the Speakers to sound any good.
I don't have that many years of design experience, but my gut feel tells me distortion in drivers are important, however below certain level(what level I don't know) it is really not that important other factors you mention dorminant. In fact both of don't even mention distortions in drivers as important!
Anyone care to elaborate on the important role of distortions in drivers that they play which will enable the speakers to sound really good.
gedlee said:
An excellent article by Keith Howard. You and your colleague suggested a nice realistic measure of nonlinear distortions. Yeah, I agree that a proper measure should consider the sensitivity of human perception of higher order harmonics at low SPLs and also the phase of distortion amplitude over frequency.
As you may know, part of this is in fact exactly what quite a few hobbysts with good measurement capability (like Zaph and Mark K) have been aware of and told people. That's why they don't simply give a single THD or IMD number of a driver but instead provide its single tone HD sweeps or IMD type distortion tests over full frequency range (not performed at an extreme listening level but at a normal listening level) to interpret the contribution of harmonics of all orders. Together with linear distortions, PROPERLY measured and evaluated nonlinear distortions are an important factor in good sound.
ttan98 said:
I use electronic crossovers and high sensitivity low distortion drivers in systems that have extreme sensitivity and headroom. I used crude methods in the past to measure this and now have a measurement system that is more sensitive and revealing of what is going on. I recently tested a bass reflex JBL 15" 2226H woofer in a bass reflex cabinet and was measuring 15-20 percent distortion at 110 db levels a foot away- I compared these 'low distortion' 15's to the open baffles, which by the way are 6 db more sensitive, they are below 3 percent in the same 60 to 300 cycle range. This does not tell the whole story though - the dipoles drive the room differently - I believe that is more important than the lower distortion because the dipoles are much more linear at the listening area.
Jay_WJ said:
I just read the power Point Presentation about
weighting of distortion on the homepge of Dr. Geddes.
Accepting the statement, that higher order distortions are less
masked at low signal levels, so they become audible, there may
be a practical impact i'd like to discuss.
Most dome Tweeters tend to have significantly higher (unweighted) distortion at low signal level than at normal
listening level. This implies, that the weighted distortion is very large and audible at low levels, when using such a system.
In my Dipol Line Arrays i use two dome tweeters on the rear of
the baffle to fill the missing HF from the rear of the
Fullrange speakers.
Since the tweeters have good efficiency, and are only used to fill
up the backward lobe of the dipole, which is already supressed
by a lining Material, they are driven over a voltage divider.
The attenuation is about 26 dB.
One time while experimenting i forgot to install the highpass capacitor...
The sound in the far field, where the effect of the backward
tweeters is audible, was more detailed and cleaner. Especially
with gentle sounds from a triangle or cymbals e.g.
I tested several high quality capacitors and dimensioned them
to produce a cutoff frequency 2 Octaves below the cutoff
of the tweeter.
No configuration sounded as good as the "no capacitor" version.
Differences between capacitors were less than the
effect observed without capacitor.
I believe the effect is due to distortion at very low excursion of
the tweeters, where force is not proportional to excursion.
There may be also mealleable effects in the suspension at
very low excursions.
Finally i designed the tweeter panel in a way that i got around a highpass at all using only the rolloff of the tweeters itself.
(Yes, i already know it is unsane.)
The excursion from low frequency portion seems to help reducing
those distortions at very low levels, even if the tweeter is
not able to radiate sound at low frequencies.
It just keeps moving. Thats how i explain it, i dont believe the capacitor itself is responsible for the effect.
I have not proven the effect scientifically, its just a subjective
observation.
The Keith Howard paper and the GedLee Powerpoint presentation both make for good reading. I've been an advocate for D.E.L. Shorter's work since the mid-Seventies, and it's good to see material that extends and expands it from a completely different direction. Excellent work, much appreciated, looking forward to seeing more.
It was a D.E.L. Shorter comment, made some time in the mid-Fifties in Wireless World, that first got my attention back in the Audionics days in the Seventies (I had an extensive collection of WW back issues at the time). Mr. Shorter did a bit of math and discovered that for signals more complex than three tones, the IM sum-and-difference products started to outnumber the simple harmonics. This has obvious implications for the source material:
First, that almost all music has a more complex structure than three tones, thus IM distortion actually dominates in the real world. Harmonic distortion only predominates with extremely simple spectra - spectra so sparse that it is musically uninteresting. The novelty of listening to a tune performed by sine-wave wears off pretty fast.
Second, that different kinds of music have different spectral densities, depending on the number of instruments playing at any given moment, and the spectral density of the instruments themselves. A single guitar is going to have a very different spectral density than a full symphony orchestra and choir. A spectral analysis of the latter is going to look like broadband noise (weighted towards the midband), but is obviously nothing of the sort.
The greater the number of tones, the greater the number of intermodulation sum-and-difference products - and these increase geometrically, not in simple proportion. With real music, the number of sum-and-difference IM terms is so great it can resemble a dynamic noise floor more than discrete easily identified artifacts. I would gently suggest this dynamic noise floor does in fact degrade the reproduction of spectrally dense music due to over-riding the subtler music content with what amounts to a type of modulation noise, while spectrally simpler music may escape this fate (and sound better).
With loudspeakers, we have an additional problem that excursion-induced distortion has a strong frequency-related component, since direct-radiators are constant-acceleration devices and horns are constant-velocity devices (within the working passband). This can lead to the result that a tweeter excited with LF material (out of its passband) can create gross IM distortion that comes and goes apparently randomly, since the out-of-band stimulus can be an artifact of low-slope and/or ineffective crossover filtering. This can be very annoying to track down and remove, since the offending stimulus may not be actually audible coming from the tweeter (since it is masked by LF content from the woofer).
The density of the spectra of acoustic instruments, as well as the complex changes in harmonic structure as the sound develops over time, makes them extremely hard to mimic electronically - not to mention the complex spatial emission characteristic, which leads to different sound depending on the local acoustic. An object as seemingly commonplace as a piano has an extremely complex emission structure, and is not trivial to tune due to pitch-shifting of dense groups of partials as the note decays over time.
I do feel that the type of reproduction system (from microphone to loudspeaker) is not necessarily going to be even-handed in the types of music it can best reproduce. Some types of music will tolerate electronic reproduction better than others, and this in turn has affected popular tastes as different types of music systems have come and gone over the decades, starting with all-tube table radios in the Twenties and Thirties, and extending to the latest 7.1 home theater confections of today.
It was a D.E.L. Shorter comment, made some time in the mid-Fifties in Wireless World, that first got my attention back in the Audionics days in the Seventies (I had an extensive collection of WW back issues at the time). Mr. Shorter did a bit of math and discovered that for signals more complex than three tones, the IM sum-and-difference products started to outnumber the simple harmonics. This has obvious implications for the source material:
First, that almost all music has a more complex structure than three tones, thus IM distortion actually dominates in the real world. Harmonic distortion only predominates with extremely simple spectra - spectra so sparse that it is musically uninteresting. The novelty of listening to a tune performed by sine-wave wears off pretty fast.
Second, that different kinds of music have different spectral densities, depending on the number of instruments playing at any given moment, and the spectral density of the instruments themselves. A single guitar is going to have a very different spectral density than a full symphony orchestra and choir. A spectral analysis of the latter is going to look like broadband noise (weighted towards the midband), but is obviously nothing of the sort.
The greater the number of tones, the greater the number of intermodulation sum-and-difference products - and these increase geometrically, not in simple proportion. With real music, the number of sum-and-difference IM terms is so great it can resemble a dynamic noise floor more than discrete easily identified artifacts. I would gently suggest this dynamic noise floor does in fact degrade the reproduction of spectrally dense music due to over-riding the subtler music content with what amounts to a type of modulation noise, while spectrally simpler music may escape this fate (and sound better).
With loudspeakers, we have an additional problem that excursion-induced distortion has a strong frequency-related component, since direct-radiators are constant-acceleration devices and horns are constant-velocity devices (within the working passband). This can lead to the result that a tweeter excited with LF material (out of its passband) can create gross IM distortion that comes and goes apparently randomly, since the out-of-band stimulus can be an artifact of low-slope and/or ineffective crossover filtering. This can be very annoying to track down and remove, since the offending stimulus may not be actually audible coming from the tweeter (since it is masked by LF content from the woofer).
The density of the spectra of acoustic instruments, as well as the complex changes in harmonic structure as the sound develops over time, makes them extremely hard to mimic electronically - not to mention the complex spatial emission characteristic, which leads to different sound depending on the local acoustic. An object as seemingly commonplace as a piano has an extremely complex emission structure, and is not trivial to tune due to pitch-shifting of dense groups of partials as the note decays over time.
I do feel that the type of reproduction system (from microphone to loudspeaker) is not necessarily going to be even-handed in the types of music it can best reproduce. Some types of music will tolerate electronic reproduction better than others, and this in turn has affected popular tastes as different types of music systems have come and gone over the decades, starting with all-tube table radios in the Twenties and Thirties, and extending to the latest 7.1 home theater confections of today.
I would like to chime in with my comments on the original question regarding the problem with hi fi.
I agree with much of the sentiment here regarding open baffles, to me they just sound so much more like a real performance, most everything else sounds like ...well a box.
But the real issue to my mind is a lack of synergy. Most people buy systems that are an amalgam of various bits a pieces, that individually are supposed to be the best sounding or best value or best bass or whatever. In the end what you get is a system that reproduces the recorded music but does so by adding a good deal of its own colouration.
I have grappled with this question repeatedly over the past couple of years and my answer was to DIY the complete system with all bits tuned so to speak to work together towards one end.
Of course all of this needs to take into account the listening room as well and other variables such as your own hearing.
Overall I feel most hi-fi has a fairly typical non-live sound, but it has become what most folk expect of a system......but it is a long way from live performance.
The tragedy I fear for those who go down the equipment tripping path is that they will never get what they are after, they will just spend ever greater amounts of money for diminishing returns.
My advice start with a good set of OBs some room tweaking and work your way back through your system from there and adhere to the concept that throwing more money at a problem does not automatically solve the problem.
I agree with much of the sentiment here regarding open baffles, to me they just sound so much more like a real performance, most everything else sounds like ...well a box.
But the real issue to my mind is a lack of synergy. Most people buy systems that are an amalgam of various bits a pieces, that individually are supposed to be the best sounding or best value or best bass or whatever. In the end what you get is a system that reproduces the recorded music but does so by adding a good deal of its own colouration.
I have grappled with this question repeatedly over the past couple of years and my answer was to DIY the complete system with all bits tuned so to speak to work together towards one end.
Of course all of this needs to take into account the listening room as well and other variables such as your own hearing.
Overall I feel most hi-fi has a fairly typical non-live sound, but it has become what most folk expect of a system......but it is a long way from live performance.
The tragedy I fear for those who go down the equipment tripping path is that they will never get what they are after, they will just spend ever greater amounts of money for diminishing returns.
My advice start with a good set of OBs some room tweaking and work your way back through your system from there and adhere to the concept that throwing more money at a problem does not automatically solve the problem.
I have never owned any of the more esoteric equipment which is often mentioned on these forums and have only been to a few live concerts in my life. Some of these had the sound electronically augmented. I was affected by the emotion of the music I heard live but have also been affected by the emotion of recorded music in my own house and I guess that is what I seek. Maybe my goals are flawed from a purist standpoint.
Recorded music is not live, that is purely acoustic with no electrical augmentation, and therefore will not sound the same. It is a matter of acceptance. Of course there is nothing wrong with trying to find ways to improve what we hear but if the law of diminishing return is neglected there is a tendency to become obsessed. This of course is great for the consumer society in which we live where some think nothing of spending hundreds of thousands of dollars to listen to a record, or need to waste gallons and gallons of diesel to run one mile in super luxury yachts.
In a way I am saying "listen to the music - not the system.
Another thing that is worth remembering is that one pays for the recording once but can listen over and over again, but one pays for a concert once and hears it once.
jamikl
Recorded music is not live, that is purely acoustic with no electrical augmentation, and therefore will not sound the same. It is a matter of acceptance. Of course there is nothing wrong with trying to find ways to improve what we hear but if the law of diminishing return is neglected there is a tendency to become obsessed. This of course is great for the consumer society in which we live where some think nothing of spending hundreds of thousands of dollars to listen to a record, or need to waste gallons and gallons of diesel to run one mile in super luxury yachts.
In a way I am saying "listen to the music - not the system.
Another thing that is worth remembering is that one pays for the recording once but can listen over and over again, but one pays for a concert once and hears it once.
jamikl
Lynne
One thing that might make me more comfortable with your position is if we could agree on some understandings regarding nonlinearity. THD and IMD are simply symptoms of the problem, which is nonlinearity, hence they are not different things. One is the result for a single tone and the other the result for multiple tones. Its far better to discuss the root problem which is the nonlinearity, not the symptoms which can be IMD or THD or multi-tone, etc.
We, and others, have looked at the signal dependency of nonlinearities and everyone finds the same thing. Large orchestral music is the least susceptable to audibilty of a nonlinearity. The density of the instruments apparently masks the errors.
Now I don't doubt your perceptions about the reproduction of orchestral music one bit, I agree with them completely. But all of the exsiting data says that your explaination of this problem being due to IM distortion is incorrect.
I tend to put the blame on poor micing techniques which destroy the coherence of the signal. Nothing is so sensitive to recording technique as a full orchestra, which is what makes it so challenging. This belief came from the late John Eargle who did a lot of recording of these types of performances.
One thing that might make me more comfortable with your position is if we could agree on some understandings regarding nonlinearity. THD and IMD are simply symptoms of the problem, which is nonlinearity, hence they are not different things. One is the result for a single tone and the other the result for multiple tones. Its far better to discuss the root problem which is the nonlinearity, not the symptoms which can be IMD or THD or multi-tone, etc.
We, and others, have looked at the signal dependency of nonlinearities and everyone finds the same thing. Large orchestral music is the least susceptable to audibilty of a nonlinearity. The density of the instruments apparently masks the errors.
Now I don't doubt your perceptions about the reproduction of orchestral music one bit, I agree with them completely. But all of the exsiting data says that your explaination of this problem being due to IM distortion is incorrect.
I tend to put the blame on poor micing techniques which destroy the coherence of the signal. Nothing is so sensitive to recording technique as a full orchestra, which is what makes it so challenging. This belief came from the late John Eargle who did a lot of recording of these types of performances.
I think a lot of people are discounting the effect of dopamine.
Here's a couple of examples which illustrate this.
Go to a stereo show one of these days, and commit yourself to listening for a few hours. You'll find that the first two or three systems you listen to sound fantastic. But after an hour or two, they all sound the same. By the third or fourth hour, you'll wind up doing anything BUT enjoying the music. Perhaps you'll go look for something to eat, or browse through the records for sale, or read the literature you picked up.
What you're experiencing is listening fatigue. What you may not realize is that there are specific brain chemicals which are depleted, and your perception of fatigue is associated with that.
Next time you're at an audio show, buy a couple of stiff cocktails once fatigue begins to set in. You'll find that the music sounds good again.
This is because alcohol increases dopamine levels in the brain.
Now you know why most audio shows have a fully stocked bar! It also lowers inhibitions, which might seal that big sale.
To make a long story short, don't trust your perception, it can be fooled very easily.
Here's one of the mechanisms at work; there are many more:
http://thebrain.mcgill.ca/flash/i/i_03/i_03_m/i_03_m_par/i_03_m_par_alcool.html
Here's a couple of examples which illustrate this.
Go to a stereo show one of these days, and commit yourself to listening for a few hours. You'll find that the first two or three systems you listen to sound fantastic. But after an hour or two, they all sound the same. By the third or fourth hour, you'll wind up doing anything BUT enjoying the music. Perhaps you'll go look for something to eat, or browse through the records for sale, or read the literature you picked up.
What you're experiencing is listening fatigue. What you may not realize is that there are specific brain chemicals which are depleted, and your perception of fatigue is associated with that.
Next time you're at an audio show, buy a couple of stiff cocktails once fatigue begins to set in. You'll find that the music sounds good again.
This is because alcohol increases dopamine levels in the brain.
Now you know why most audio shows have a fully stocked bar! It also lowers inhibitions, which might seal that big sale.
To make a long story short, don't trust your perception, it can be fooled very easily.
Here's one of the mechanisms at work; there are many more:
http://thebrain.mcgill.ca/flash/i/i_03/i_03_m/i_03_m_par/i_03_m_par_alcool.html
gedlee,
I think there are one or two audiodesigners and recording/master/mixing engineers that are more along the lines of Lynn than those of yours.
Can you point to some work on this?
Simple material is known to sound good or decent on non linear products but once you throw complex stuff on the medioker gear it all colapses. Going over to more linear components the complex stuff sounds so much more real and clean. It's the typical hifi-show scenario where obscure manufacturers demo their jewleryboxes with soft jazz or a little girl with guitar kind of music.
I am impressed by your work but I do not agree that for example loudspeaker distortion has so low importance as you seems think.
The best loudspeaker i have listened to have had HD at or below -60dB at aprox. 90dB.
Frequency response and directivity of course is extremly important none the less.
/Peter
I think there are one or two audiodesigners and recording/master/mixing engineers that are more along the lines of Lynn than those of yours.
Can you point to some work on this?
Simple material is known to sound good or decent on non linear products but once you throw complex stuff on the medioker gear it all colapses. Going over to more linear components the complex stuff sounds so much more real and clean. It's the typical hifi-show scenario where obscure manufacturers demo their jewleryboxes with soft jazz or a little girl with guitar kind of music.
I am impressed by your work but I do not agree that for example loudspeaker distortion has so low importance as you seems think.
The best loudspeaker i have listened to have had HD at or below -60dB at aprox. 90dB.
Frequency response and directivity of course is extremly important none the less.
/Peter
The most revealing music for testing a system is large scale choral works. In a sense, all of the singers are like a mass of out of tune violins, each slightly different in pitch. Heard live, these variations combine to make something that can be breathtakingly beautiful. Through >95% of highend equipment, they can range from sounding euphonic on one extreme, to the ol' fingernails on the chalkboard on the other...this with the same recording.
I've always taken these differences(especially of the fingernails on the chalkboard variety) to be related to IMD levels generated by phono cartridge mistracking and poorly performing loudspeakers.
I've always taken these differences(especially of the fingernails on the chalkboard variety) to be related to IMD levels generated by phono cartridge mistracking and poorly performing loudspeakers.
Patrick Bateman said:
I once experienced the same thing at a hifi show near Heathrow, with my friend Jack who was introducing me to the hobby. We ended up exhausted in the Lowther room. The bloke in there asked us if we'd like to choose a record which we did, and would we like a cup of tea which we accepted.
The Lowthers weren't being played too loud, we relaxed with our tea and it suddenly dawned upon both of us that this was what we had been searching for.
Later Jack bought some plans and made a pair of Acoustas with PM6's they were great speakers. I was only about 19 at the time and couldn't afford to change my mind.
I've been looking for that free, open and up front sound ever since and still can't afford to buy any Lowthers
Jem
Pan said:
I can't help what others think or believe, I thought that way myself until I did the studies, but the data simply says otherwise. Now I'm not saying (and I've made this point many times before) that nonlinearity in a loudspeaker is always insignificant, only that it can easily be made to be insignificant with a decent design. It can also be made to be significant by screwing up the design. My point is that THD and IMD numbers etc. are meaningless and any attempt to explain sound quality with them is futile. Read my work - and that of so many others who have done the same thing, they all come to the same conclusion. You can ignore the results or you can accept them and try to figure out what it is that really is the problem. The choice is yours. It should be clear which way I lean and I don't regret that decision one bit since I now "see the light".
Admittedly, I have access to data and test results done for clients, etc. that go well beyond what has been published and these results can make me more confident in my position than others might be. I just telling you what I have concluded from what has been done.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.