I don't have any reference handy atm, it is pretty complex. But the gist of it is that there are muscles in ears that react to loud environments and reduce the sensitivity. That takes time (temporal masking), and the force spent on them is also indirect cue of loudness. Think of analogy with eye pupil reacting to intense light. Effects of masking are related because dynamic range of our auditory nerves is limited, like a window on the total dynamic range we can perceive. Shifting of that window is result of SPL, and we are aware of that as loudness, and that shifting is masking input that drops below dynamic range of that 'window'. Why I thought of 'acoustic energy delivered', is because our ears are not linear, and react not on SPL produced by speakers, but on SPL reaching the nerves. And even then, main purpose of sensitivity shifting is to protect auditory nerves from damage, thus logically ears need to react to acoustic energy that can cause the damage, not so much on the SPL (depends on frequency, in-ear resonances, etc), not even on average SPL.
The auditory processing is relatively independent from the sensitivity shifting, which basically is compressor. I understand that by feeding ears with precompressed signals, ears are fooled into believing the sensitivity shifting triggered more than actually, and thus sounds incoming must be louder than they really are. Basically brain is equipped to handle compression. I wonder if that is reason why mild compression is perceived as sounding better.
But without artificial compression, perception of loudness is result of both sensitivity shifting due to acoustic energy, and the acoustic energy itself reaching auditory nerves.
Compression loudness is due to limited ratio of peak SPL to average SPL. Loudspeakers are going into compression way sooner with mid-fi than with studio monitors. Yes that is accompanied with HD, but HD isn't neccesarily the cause of perceived loudness. Good uncompressed source sounds less loud because its dynamic range, peak to average ratio, is wider. Increasing amp volume on already compressing speakers makes it sound louder, not because more SPL is produced, but because average SPL is rising. Doesn't imho matter much on what the harmonic distribution is, else perceived volume of music would depend on spectral content.
Also, when SPL limited speakers are forced with amp, the energy imparted to them and not escaping as acoustic energy, is stored, and converted later. Energy storage of drivers. That stored energy fills the gaps between transients, increasing average SPL.
Actually, I'm tempted to attribute all sorts of listening fatigue to exactly issues of compression and energy storage, only after that to HD.
I thought I'm reasonably safe using term 'almost' all recordings. But aren't you confusing 'loudness race' with decent amounts of compression, that is inherent part of mastering? Chesky *is* using compression. They just avoid overcompression. Of course I'm not saying that there aren't those few CD's with zero compression. I'd just be inclined to think these are academic experiments. There is little use for such CD's without being reproduced through PA systems on dedicated venues. *Real* 120db dynamic range is not exactly what neighbors tolerate.
You seem to like resolute objection. I'd rather refer you here: http://www.lenardaudio.com/education/08_live.html I'm going on several assumptions. It obviously requires a major leap in faith to think that heavy metal bands don't use any compression in live events. But thats not where CD's sound comparatively bad either 😉
There is a section on live sound problems on Lenard pages. He specifically lists issues, and compression as one. "Continuous compression-limiting in live performances causes the music to sound flat and lifeless." Thus my assumption is that not one *Pro* is ever running their live performance into compression, but solve their loudness requirements by brute force - kilowatts and speaker arrays.
Why bring up this matter here? I'm often wondering that soft-limiting and ear sensitivity shifting are quite similar in many ways, both causing specific nonlinear distortions. If brain is accustomed to such distortion and compensates for it, that makes you think that many forms of nonlinear distortions arising in amps and speakers are basically transformed into loudness cues, not so much of HD. Valves, speakers both exhibit some form of soft-limiting. These create low order harmonics. So perhaps low order harmonics are not so much masked, but supressed by brain as sensitivity shifting induced artifacts? Would explain why some distorting amps sound more natural than we'd expect.
Hi, Francis,
I myself do not believe one can hear 22khz. 😀 What I said in post #77 is "feel". I just realize there are 2 parts, subsonic and supersonic.
I first tought about it in subsonic. Like you said we can sense it. I think our body can feel even 1hz.
In churches (old churces), the seats usually made by hard wood, without any pillow. When the music played is so "glorious", I noticed that these seats are "slightly vibrating", following the music. Because we sit on these wooden chairs, the vibration of music is transfered to our body who touches the chair. I think these vibration transfer also impact one's sense in enjoying the music. Realizing it or not, our skin who touches the vibrating medium adds pleasure in enjoying the music.
Another example is in car audio. If you play a music, with your hands off (not resting on the door) and compare it with your hand resting on the door (where the speaker are mounted on the door), you can enjoy the music differently, just because our hand feels the mid-bass.
Car audio manufacturer also makes strange devices, called "Aura Shaker", even Pioneer car audio division make it. It 's actually a vibrator, not a speaker, but adds the sense of music in the car.
What I'm saying here, maybe we loss some of the emotion of the live performance because the reproduction bandwith only considers what can be heard, lacking what can be feeled.
I dont know how supersonic affects this. Maybe what MBK said is right, phase shift. Or maybe the 10's khz can modulate the frequency between 20-20khz differently, producing different sense in 20-20khz.
I just know that with system that has no frequency bandpass bandwith, the sense of hearing the same system is different.
Audio CD's is limited to 22.05khz. Turntable do not have this.
In the same system, I can hear the turntable has "more spirit" with the same album heard from CD player.
Maybe it's the bandwith thing, because the CD player is limited by 22.05khz, and turntable do not have this. What else can make the sound of turntable different from CD player with the same album (considering both are good quality gears)
A guru here once write that turntable can "spit", producing signals with slew rate 500V/us. So turntable's bandwith must be more than 22.05khz. Could it be that what is contained upper than 22.05khz is what giving the turntable the "spirit" that CD player do not have?
I myself do not believe one can hear 22khz. 😀 What I said in post #77 is "feel". I just realize there are 2 parts, subsonic and supersonic.
I first tought about it in subsonic. Like you said we can sense it. I think our body can feel even 1hz.
In churches (old churces), the seats usually made by hard wood, without any pillow. When the music played is so "glorious", I noticed that these seats are "slightly vibrating", following the music. Because we sit on these wooden chairs, the vibration of music is transfered to our body who touches the chair. I think these vibration transfer also impact one's sense in enjoying the music. Realizing it or not, our skin who touches the vibrating medium adds pleasure in enjoying the music.
Another example is in car audio. If you play a music, with your hands off (not resting on the door) and compare it with your hand resting on the door (where the speaker are mounted on the door), you can enjoy the music differently, just because our hand feels the mid-bass.
Car audio manufacturer also makes strange devices, called "Aura Shaker", even Pioneer car audio division make it. It 's actually a vibrator, not a speaker, but adds the sense of music in the car.
What I'm saying here, maybe we loss some of the emotion of the live performance because the reproduction bandwith only considers what can be heard, lacking what can be feeled.
I dont know how supersonic affects this. Maybe what MBK said is right, phase shift. Or maybe the 10's khz can modulate the frequency between 20-20khz differently, producing different sense in 20-20khz.
I just know that with system that has no frequency bandpass bandwith, the sense of hearing the same system is different.
Audio CD's is limited to 22.05khz. Turntable do not have this.
In the same system, I can hear the turntable has "more spirit" with the same album heard from CD player.
Maybe it's the bandwith thing, because the CD player is limited by 22.05khz, and turntable do not have this. What else can make the sound of turntable different from CD player with the same album (considering both are good quality gears)
A guru here once write that turntable can "spit", producing signals with slew rate 500V/us. So turntable's bandwith must be more than 22.05khz. Could it be that what is contained upper than 22.05khz is what giving the turntable the "spirit" that CD player do not have?
I don't want to hijack this thread and take it too far astray from the topics that relate directly to things like amps and their distortion...
There is a threshold where the ear shifts... most hi-fi listening is probably done with the threshold "shifted" up...
Not my understanding at all... but perhaps my info has been eclipsed with more recent research?
Energy storage of drivers *is* distortion. It is/wasn't my intent to limit my statement that: of two speakers running at the same nominal SPL, the one with the higher distortion will sound louder - to simply "Harmonic Distortion".
Ummm... I think ur confusing things a bit. No CD has 120dB dynamic range, they're limited to 96dB because of the 16 bit limitation. My present system will probably get very near an SPL of 120dB, the HF section definitely can... not that I would *want to be in the same room*. 😀
The question isn't how "loud" a CD is, but how many dB down from "0" is the average AND how much dynamic range is there between said "average" above and below. All "mastering" in the digital age is first trying to get the peak level just at "0"... beyond that is where compression or expansion (etc) may or may not be applied.
I can't speak to exactly what Chesky does or does not do in reality in their mastering, I can speak to Dorian and on the basis of having spoken to some others as well. Compression is not required or usually applied.
The "loudness race" is something that pop recordings are interested in, and FM conglomerate's radio stations... not audiophile labels or recordings.
Dunno what this shows. If you want to point to a specific page, please do.
I can tell you that there are NO pro touring PA systems, especially of the sort pictured on that splash page above that do NOT extensively use signal conditoning and at the VERY MINIMUM soft limiters going out to the mains. Most use much more processing.
Any that fail to do this quickly go out of biz due to huge quantities of blown amps and drivers.
Your assumptions are not correct. What is used is a LIMITER... compression may or may not be applied depending on the source usually on a per input channel basis. No one applies heavy compression to the main feed, unless there is some specific reason to do so... which, btw is effectively what you get if you slam the mains high enough into a soft limiter...
Interesting hypothesis, but I think your idea about sensitivity shifting being a *fast* and *realtime* process is incorrect.
_-_-bear
There is a threshold where the ear shifts... most hi-fi listening is probably done with the threshold "shifted" up...
Not my understanding at all... but perhaps my info has been eclipsed with more recent research?
Energy storage of drivers *is* distortion. It is/wasn't my intent to limit my statement that: of two speakers running at the same nominal SPL, the one with the higher distortion will sound louder - to simply "Harmonic Distortion".
Ummm... I think ur confusing things a bit. No CD has 120dB dynamic range, they're limited to 96dB because of the 16 bit limitation. My present system will probably get very near an SPL of 120dB, the HF section definitely can... not that I would *want to be in the same room*. 😀
The question isn't how "loud" a CD is, but how many dB down from "0" is the average AND how much dynamic range is there between said "average" above and below. All "mastering" in the digital age is first trying to get the peak level just at "0"... beyond that is where compression or expansion (etc) may or may not be applied.
I can't speak to exactly what Chesky does or does not do in reality in their mastering, I can speak to Dorian and on the basis of having spoken to some others as well. Compression is not required or usually applied.
The "loudness race" is something that pop recordings are interested in, and FM conglomerate's radio stations... not audiophile labels or recordings.
Dunno what this shows. If you want to point to a specific page, please do.
I can tell you that there are NO pro touring PA systems, especially of the sort pictured on that splash page above that do NOT extensively use signal conditoning and at the VERY MINIMUM soft limiters going out to the mains. Most use much more processing.
Any that fail to do this quickly go out of biz due to huge quantities of blown amps and drivers.
Your assumptions are not correct. What is used is a LIMITER... compression may or may not be applied depending on the source usually on a per input channel basis. No one applies heavy compression to the main feed, unless there is some specific reason to do so... which, btw is effectively what you get if you slam the mains high enough into a soft limiter...
Interesting hypothesis, but I think your idea about sensitivity shifting being a *fast* and *realtime* process is incorrect.
_-_-bear
Regarding hearing above 20kHz.
Before I was about 20 I could reliably hear all sorts of nasty things above 20kHz. Of this there is no doubt. 😀
One of the things that this hearing ability made excruciating was going the the Museum Of Natural History's gem room (NYC) some time after Murf The Surf lifted the Star Of India Ruby. They had installed these early vintage ultrasonic motion detectors (I assume that's what they were) which apparently did not shut off the emitters during the day! While I loved the room, I could not stand to be in there. I asked people, guards "do you hear that"? Puzzled looks. As if to say "what are ya nuts boy?" 😀
Table saws, bus brakes, subway brakes, excruciating.
It's kind of a blessing of sorts not to hear that stuff any more.
So, people can hear some of that stuff... and the folding back (IM again) of the higher harmonic stuff does apparently have some effect too...
_-_-bear
Before I was about 20 I could reliably hear all sorts of nasty things above 20kHz. Of this there is no doubt. 😀
One of the things that this hearing ability made excruciating was going the the Museum Of Natural History's gem room (NYC) some time after Murf The Surf lifted the Star Of India Ruby. They had installed these early vintage ultrasonic motion detectors (I assume that's what they were) which apparently did not shut off the emitters during the day! While I loved the room, I could not stand to be in there. I asked people, guards "do you hear that"? Puzzled looks. As if to say "what are ya nuts boy?" 😀
Table saws, bus brakes, subway brakes, excruciating.
It's kind of a blessing of sorts not to hear that stuff any more.
So, people can hear some of that stuff... and the folding back (IM again) of the higher harmonic stuff does apparently have some effect too...
_-_-bear
Ultrasonics:
Yes, I agree it must only be "feel" but in the end it amounts to the same problem - you must find some physical entity that does the feeling.
I remember those ultrasonic sensors too - and could hear them as well. Long time ago. It is not unlikely that sensing them was an intermodulation or other in the ear distortion issue. They were loud, very loud. So it may have been possible that a sub-harmonic was generated that was just on the edge of a young persons hearing range. I can still hear a horizontal scan - but luckily modern TV sets are much better made than the old ones. I remember being able to walk down the road and being able to tell if a household was watching TV. That is seriously loud.
Infrasonics:
Yup, I actually have a couple of the Aura drivers somewhere. Evil bits of work. 🙂 Ages ago we had some fun, I fitted one up under a cushion on the prime listening chair of a friends system. He has big Magnaplanars, so no really low bass. It was fun pushing the really low bass through the chair, but in the end we all agreed that it really didn't add anything to the music.
This rather points to a crucial point about all sensation of sound outside the conventional frequency envelope. For it to be musically valuable we must be able to resolve pitch. If you are simply shaken with an infra-bass driver, I don't think your stomach is appropriately linked to the hearing centre of the brain to be able to resolve its pitch. Perhaps it is, again an experiment to be done.
Live music.
Bear (I assume Randy) certainly matches my experience with live Rock and Roll. They use compressor/limiters on everything. There is a real art to mixing a live band. It is not just about dynamic range, it is about filling the spectrum. Typically a rig will have a compressor/limiter on each of a set of frequency bands, and will be tweaked so that the music always seems to fill the spectrum. There is also scant energy below about 60Hz, those big bass horns just can't deliver it anyway. Then you get such lovely devices as the Aural-exciter.
Yes, I agree it must only be "feel" but in the end it amounts to the same problem - you must find some physical entity that does the feeling.
I remember those ultrasonic sensors too - and could hear them as well. Long time ago. It is not unlikely that sensing them was an intermodulation or other in the ear distortion issue. They were loud, very loud. So it may have been possible that a sub-harmonic was generated that was just on the edge of a young persons hearing range. I can still hear a horizontal scan - but luckily modern TV sets are much better made than the old ones. I remember being able to walk down the road and being able to tell if a household was watching TV. That is seriously loud.
Lets not go down the digital vs. vinyl argument too far. Suffice it to say that you need to be very careful. Whereas a vinyl system has no specifically mandated high end roll off, just about every LP chain in existence does. The physics of creating an LP make it essentially impossible to avoid this, particularly the physics of a tape recorder. You would need to show that the recording studio was able to preserve those ultrasonics. Also the cutting lathe. Think about the cutting lathe amplifier. Interesting to think about the design of those. Not exactly low power. I bet there is more than a bit of feedback in their design, and not too many SET amplifiers in evidence. 🙂 They would need very careful compensation to be stable. The brief dalliance with quadraphonics did try to add an octave of response, but it was hard won, and the stylus would cheerfully polish the wiggles off the groove after a few playings anyway.
The final post production step. This is typically done by the pressing plant and is done to optimise the sound for the media. LP has always been tweaked at this stage to optimise it for the vagaries of the cutting and pressing process. One of the dreadful issues with very early CDs was that they took the LP post production master and cut a CD. It sounded strident and nasty. Because it had been optimised for LP. There were CD pressing plants that had no digital production facilities, so when they were sent a digital master, that ran it through a DAC, through their analog post production chain, and re-digitalised it. Totally ruining the dithering of the master. And so on. The point? You cannot compare a CD and an LP of the same recording. The assumption that they come from the same source is not true. The provenance of the source may actually be very unfortunate. Quite a few bands have done multiple reissues of CDs, each time re-mastering.
Infrasonics:
Yup, I actually have a couple of the Aura drivers somewhere. Evil bits of work. 🙂 Ages ago we had some fun, I fitted one up under a cushion on the prime listening chair of a friends system. He has big Magnaplanars, so no really low bass. It was fun pushing the really low bass through the chair, but in the end we all agreed that it really didn't add anything to the music.
This rather points to a crucial point about all sensation of sound outside the conventional frequency envelope. For it to be musically valuable we must be able to resolve pitch. If you are simply shaken with an infra-bass driver, I don't think your stomach is appropriately linked to the hearing centre of the brain to be able to resolve its pitch. Perhaps it is, again an experiment to be done.
Live music.
Bear (I assume Randy) certainly matches my experience with live Rock and Roll. They use compressor/limiters on everything. There is a real art to mixing a live band. It is not just about dynamic range, it is about filling the spectrum. Typically a rig will have a compressor/limiter on each of a set of frequency bands, and will be tweaked so that the music always seems to fill the spectrum. There is also scant energy below about 60Hz, those big bass horns just can't deliver it anyway. Then you get such lovely devices as the Aural-exciter.
Hi, Francis,
Thanks for the explenation.
In home audio system, there alot of "mumbo-jumbo" stuff, some said they can hear the difference, some said it is a hoax. I still don't accept those, make no sense. They said it is have to do with harmonics and vibration, but how?
Those things are (not cheap)
1. A small wooden circle /carpet/ metal tube that can bend direction of sound focus
2. Special Spikes placed on every equipment leg
3. A special mattras for placing audio equipment
4. Wooden block just to lift RCA cable a few cm above ground
5. A very expensive/specially design chair just for listening to Hi-Fi
Are they really working on vibration/harmonics or just hoax? I think they do not work on 20-20khz audio. IF they work, what spectrum they operating on?
Thanks for the explenation.
In home audio system, there alot of "mumbo-jumbo" stuff, some said they can hear the difference, some said it is a hoax. I still don't accept those, make no sense. They said it is have to do with harmonics and vibration, but how?
Those things are (not cheap)
1. A small wooden circle /carpet/ metal tube that can bend direction of sound focus
2. Special Spikes placed on every equipment leg
3. A special mattras for placing audio equipment
4. Wooden block just to lift RCA cable a few cm above ground
5. A very expensive/specially design chair just for listening to Hi-Fi
Are they really working on vibration/harmonics or just hoax? I think they do not work on 20-20khz audio. IF they work, what spectrum they operating on?
Snake oil
Oh dear, don't get me started. And perhaps we should avoid too much discussion anyway - these things aren't about distortion.
My thesis is simple. There needs to be a physical underpinning - and it needs numbers. Probably just about any device or tweak you can think of will have some tiny effect, but if you do the numbers you will find that it is so totally swamped by other effects that it is totally meaningless. Things that are 250 dB down etc.
You are right about the issue of the spectrum they work in. That is more often than not the downfall of the idea.
All except for a chair. I don't know about "special" chairs, but two things are important. The bit of chair behind your ears, and how comfortable it is. I have an Ekornes Stressless, which I first heard about recommended on rec.audio.high-end. It is a truly wonderful chair. For the money it had better be. But it has nothing to do with the sound. Just being comfortable and able to really relax and enjoy the music. It is astounding how easy it is to stay in a stressed position, and not notice, until you relax, and realise how much better everything is.
More on distortion soon.
Oh dear, don't get me started. And perhaps we should avoid too much discussion anyway - these things aren't about distortion.
My thesis is simple. There needs to be a physical underpinning - and it needs numbers. Probably just about any device or tweak you can think of will have some tiny effect, but if you do the numbers you will find that it is so totally swamped by other effects that it is totally meaningless. Things that are 250 dB down etc.
You are right about the issue of the spectrum they work in. That is more often than not the downfall of the idea.
All except for a chair. I don't know about "special" chairs, but two things are important. The bit of chair behind your ears, and how comfortable it is. I have an Ekornes Stressless, which I first heard about recommended on rec.audio.high-end. It is a truly wonderful chair. For the money it had better be. But it has nothing to do with the sound. Just being comfortable and able to really relax and enjoy the music. It is astounding how easy it is to stay in a stressed position, and not notice, until you relax, and realise how much better everything is.
More on distortion soon.
My mistake
My mistake. We shouldn't go into those SnakeOil stuff. The distortion in power amps is more important. 😀
My mistake. We shouldn't go into those SnakeOil stuff. The distortion in power amps is more important. 😀
Hi MBK,
My thoughts match yours regarding the importance of initial transients, and thus the need for an exact phase coherence within the full audio spectrum, plus a minimum of loudspeaker time shifted back-emf induced amplifier reactivity.
If 'stage' shifting due to head movements are from detecting lower frequencies then our capabilities for detecting phase shift/timing must be incredible.
This is my reason for simulation testing amplifier circuits with suddenly starting sinewaves, not only input but reverse too, to check how a design controls 'back-emfs', and to heck with all the theoretically based criticism that such methodology invokes. I can assure everyone that reverse 'loudspeaker back-emf' tests show up strikingly different requirements than those for merely minimising thd, afterall, I cannot think of a much better way for inducing input intermodulation.
Hi Francis,
I have just read comments on the Benchmark DAC1 reporting headphone listening of piano music being accurate, when less so with the same feed to an amplifier/loudspeaker system. This matches what you were saying, but maybe there is still more to be done in rendering NFB controlled amplifiers impervious to loudspeaker induced back-emfs in order to reduce audible distortion.
Hi Mike,
Maybe loudspeaker back-emf would be number (4) in your list of intermodulation sources, and this constantly changes with not only all of your listed aspects, but with the loudspeaker itself.
Change the loudspeaker - change the sound - change the amplifier-loudspeaker interface induced 'phantom' errors.
So I conclude that tests *must* be done with a loudspeaker load.
Hi Lumanauw,
My long battered ears are getting old, and thus I have little genuine hearing above 10kHz. But a 23kHz animal repellant 'sounds' bloody awful if I get in line, and last summer I was able to 'home in' on a baby bat that was calling for its mother. I could not 'hear' the bat, but its 'sound' set up an extremely uncomfortable auditory resonance. Beyond a certain distance, or out of line with a *single* ear and I could not 'hear' it.
So yes, frequencies above 20kHz remain vitally significant, even though we cannot 'hear' them. Maybe there is some threshold amplitude below which they are cannot be sensed.
Some vinyl cartridges can respond to 30kHz, so whilst the CD brick-wall is not going to especially degrade voices, it might still detract from a harmonic ambience.
Hi Wimms.
Yes the energy storage of drivers is what causes reverse transduction to be reflected back into an amplifier's output stage, this I describe as time shifted. Also when driver excursions increase, then voice coil impedances change considerably, such that composite loudspeaker characteristics also change. Another reason for many small parallel acting drivers sounding much less distorted.
Cheers ........... Graham Maynard.
My thoughts match yours regarding the importance of initial transients, and thus the need for an exact phase coherence within the full audio spectrum, plus a minimum of loudspeaker time shifted back-emf induced amplifier reactivity.
If 'stage' shifting due to head movements are from detecting lower frequencies then our capabilities for detecting phase shift/timing must be incredible.
This is my reason for simulation testing amplifier circuits with suddenly starting sinewaves, not only input but reverse too, to check how a design controls 'back-emfs', and to heck with all the theoretically based criticism that such methodology invokes. I can assure everyone that reverse 'loudspeaker back-emf' tests show up strikingly different requirements than those for merely minimising thd, afterall, I cannot think of a much better way for inducing input intermodulation.
Hi Francis,
I have just read comments on the Benchmark DAC1 reporting headphone listening of piano music being accurate, when less so with the same feed to an amplifier/loudspeaker system. This matches what you were saying, but maybe there is still more to be done in rendering NFB controlled amplifiers impervious to loudspeaker induced back-emfs in order to reduce audible distortion.
Hi Mike,
Maybe loudspeaker back-emf would be number (4) in your list of intermodulation sources, and this constantly changes with not only all of your listed aspects, but with the loudspeaker itself.
Change the loudspeaker - change the sound - change the amplifier-loudspeaker interface induced 'phantom' errors.
So I conclude that tests *must* be done with a loudspeaker load.
Hi Lumanauw,
My long battered ears are getting old, and thus I have little genuine hearing above 10kHz. But a 23kHz animal repellant 'sounds' bloody awful if I get in line, and last summer I was able to 'home in' on a baby bat that was calling for its mother. I could not 'hear' the bat, but its 'sound' set up an extremely uncomfortable auditory resonance. Beyond a certain distance, or out of line with a *single* ear and I could not 'hear' it.
So yes, frequencies above 20kHz remain vitally significant, even though we cannot 'hear' them. Maybe there is some threshold amplitude below which they are cannot be sensed.
Some vinyl cartridges can respond to 30kHz, so whilst the CD brick-wall is not going to especially degrade voices, it might still detract from a harmonic ambience.
Hi Wimms.
Yes the energy storage of drivers is what causes reverse transduction to be reflected back into an amplifier's output stage, this I describe as time shifted. Also when driver excursions increase, then voice coil impedances change considerably, such that composite loudspeaker characteristics also change. Another reason for many small parallel acting drivers sounding much less distorted.
Cheers ........... Graham Maynard.
Graham,
I am inclined to be very sympathetic with your idea(s) on the effect(s) of reactive speaker loads on the response of the speaker and amplifier... it is certainly an important area that has not be adequately explored or quantified, imho.
😀
_-_-bear
I am inclined to be very sympathetic with your idea(s) on the effect(s) of reactive speaker loads on the response of the speaker and amplifier... it is certainly an important area that has not be adequately explored or quantified, imho.
😀
_-_-bear
Graham Maynard
you've suggested that amps should be tested on reactive load.
Thats fine
but...
would you agree with this statement:
Speaker load is much more complicated. It cannot be simulated only by resistors, caps and coils. It is electroacoustic transducer, which 'products' acoustic energy and reacts with acoustic energy transducing it back to electric. Considering sound speed is far lower than current speed, we should simulate (digitally) acoustic energy response as a transmission line depending on room the speaker works in. And OK. the transmission line can be ilustrated by caps, coils... for one defined freqency. But for more complex signal it is a nightmare- he load is delaying and Zin(f) is a pure nightmare. So this approach is only a first step into a deep forest.
regards
you've suggested that amps should be tested on reactive load.
Thats fine
but...
would you agree with this statement:
Speaker load is much more complicated. It cannot be simulated only by resistors, caps and coils. It is electroacoustic transducer, which 'products' acoustic energy and reacts with acoustic energy transducing it back to electric. Considering sound speed is far lower than current speed, we should simulate (digitally) acoustic energy response as a transmission line depending on room the speaker works in. And OK. the transmission line can be ilustrated by caps, coils... for one defined freqency. But for more complex signal it is a nightmare- he load is delaying and Zin(f) is a pure nightmare. So this approach is only a first step into a deep forest.
regards
I certainly wouldn't worry, the trees are really short, not even shrubs
you might look into the reciprocity principle
dynamic loudspeaker drivers are very inefficient in turning electrical energy into sound, and therefore equally inefficient in turning sound back into electrical energy
most electrical-mechanical-acoustic-mechanical-electrical "round trip" scenarios give way less than 1% impedance variation from acoustic coupling
the voice coil is more strongly coupled to the cone/dome/membrane acoustic radiator so radiator mechanical resonances might be visible in loudspeaker impedance curves but to get good acoustic radiation flatness most loudspeaker radiator designs are mechanically well damped
compared to 1st order effects of the magnitude of Theil-Small electroacoustic parameters, voice coil heating, inductance variation and crossover impedance effects any likely reflected electrical-acoustic-electrical terms must be really distant second order effects
Graham clearly has a point in regards to a dominant pole compensated class B – weak AB stage easily being poorly designed to not be able to meet the I-V / dt demands of a worst case dynamic driver + crossover load but I would also caution against overemphasizing the point or trying to stretch it to cover all “unexplained” perceptual differences
An amplifier’s response/distortion with respect to loudspeaker back emf is well described by its output impedance over the real I-V & I-V / dt output range it is designed to operate within
you might look into the reciprocity principle
dynamic loudspeaker drivers are very inefficient in turning electrical energy into sound, and therefore equally inefficient in turning sound back into electrical energy
most electrical-mechanical-acoustic-mechanical-electrical "round trip" scenarios give way less than 1% impedance variation from acoustic coupling
the voice coil is more strongly coupled to the cone/dome/membrane acoustic radiator so radiator mechanical resonances might be visible in loudspeaker impedance curves but to get good acoustic radiation flatness most loudspeaker radiator designs are mechanically well damped
compared to 1st order effects of the magnitude of Theil-Small electroacoustic parameters, voice coil heating, inductance variation and crossover impedance effects any likely reflected electrical-acoustic-electrical terms must be really distant second order effects
Graham clearly has a point in regards to a dominant pole compensated class B – weak AB stage easily being poorly designed to not be able to meet the I-V / dt demands of a worst case dynamic driver + crossover load but I would also caution against overemphasizing the point or trying to stretch it to cover all “unexplained” perceptual differences
An amplifier’s response/distortion with respect to loudspeaker back emf is well described by its output impedance over the real I-V & I-V / dt output range it is designed to operate within
True, but funny things can and do happen when a signal is present on the "other" input terminal. Feedback of course from the output.
-Chris
-Chris
jcx I have to think it all over again
Yes anatech I agree
about the 'other' input terminal-do you maen what I fear?:
1% is quite high for me
especially if the amp has poor open-loop damping factor (like ccs common emitter/drain output).
Compared to open loop gain (for the 'other' input terminal of course) 0.01% is quite high I would say. And the problem is that acoustic energy transduced to voltage again may have very little corelation with input signal (it may 'come' from the other speaker driven by the other amp or be a product of flatter echo etc.)
best regards
Yes anatech I agree
about the 'other' input terminal-do you maen what I fear?:
1% is quite high for me
especially if the amp has poor open-loop damping factor (like ccs common emitter/drain output).
Compared to open loop gain (for the 'other' input terminal of course) 0.01% is quite high I would say. And the problem is that acoustic energy transduced to voltage again may have very little corelation with input signal (it may 'come' from the other speaker driven by the other amp or be a product of flatter echo etc.)
best regards
Hi darkfenriz,
Yup. 'Tis what you fear. The speaker output terminal is an attenuated input terminal on most amps. Mr Pass's Stassis amp gets around that by not including the outputs in the feedback loop. I'm sure there are other similar designs including some of mine that do the same. A low feedback amp will not suffer this to the same extent as a high feedback amp (by definition).
Surprising how many of us forget about this "input".
-Chris
Yup. 'Tis what you fear. The speaker output terminal is an attenuated input terminal on most amps. Mr Pass's Stassis amp gets around that by not including the outputs in the feedback loop. I'm sure there are other similar designs including some of mine that do the same. A low feedback amp will not suffer this to the same extent as a high feedback amp (by definition).
Surprising how many of us forget about this "input".
-Chris
Hi amplifierguru,
What I was talking about was a signal getting into an amplifier from the speaker terminal. Generally more sensitive to this as the interfering frequency goes up due to parallel R C in the feeback loop. CB transmission can be especially troublesome.
Any signal applied here will be attenuated to some degree depnding on the feedback ratio, damping factor and source impedance. If there is no cancelling signal, this will be amplified as a normal signal would.
Just another noise depending on your speakers and amp combination.
-Chris
What I was talking about was a signal getting into an amplifier from the speaker terminal. Generally more sensitive to this as the interfering frequency goes up due to parallel R C in the feeback loop. CB transmission can be especially troublesome.
Any signal applied here will be attenuated to some degree depnding on the feedback ratio, damping factor and source impedance. If there is no cancelling signal, this will be amplified as a normal signal would.
Just another noise depending on your speakers and amp combination.
-Chris
Hi, Mr.Maynard,
JCX's writing gives interesting input. Loudspeakers cannot be prepresentated by R+L+C, because it is an electomechanical device. You move the speaker's cone, you will get voltage accross the speaker coil. Maybe this is gives back EMF that cannot replaced by RLC, and why testing on real loudspeaker different than testing on dummy load (even reactance one), because loudspeaker can generate voltage when we move the cones.
JCX's writing gives interesting input. Loudspeakers cannot be prepresentated by R+L+C, because it is an electomechanical device. You move the speaker's cone, you will get voltage accross the speaker coil. Maybe this is gives back EMF that cannot replaced by RLC, and why testing on real loudspeaker different than testing on dummy load (even reactance one), because loudspeaker can generate voltage when we move the cones.
Hi Darkfenriz,
Re your post #91.
But of course! An R+L+C model is only a first step, so it is either a case of simulating with some more realistic loads, or endless discerning empiricism with critical reviewing of the outcome.
As you say - the load is delaying, but as Wimms has noted with R+L+C alone, the current often becomes leading.
This is my reason for examining amplifiers when reverse driven too. The worst fractional electrical disturbances tend to arise as a result of crossover reactivity around the upper crossover frequency. In real life I have also noted some serious problems due to dynamically induced bass driver back emf arising when there is inadequate amplifier power rail capability.
An amplifier that has never been tested with suddenly starting sinewaves, input and output, cannot be assumed to be satisfactory at maintaining dynamic loudspeaker control when compared to one that has.
This is also my understanding as to why NFB/divider linked bridged amplifiers sound inferior to those that are driven by an external phase splitter. Errors fractionally circulate, instead of being finite, and those errors manifest as a 'confusion' or 'muddying' of higher frequencies due to lower frequency drive.
Hi jcx,
I understand your point about electromagnetic inefficiency, but occasionally a loudspeaker can be dynamically driven to store considerable amounts of 'air-spring' energy, and this is not just reverse transduced as a phase shifting component, but at the fundamental generating frequency for the 'charged' system, and this comes back as a delayed leading current wave that has nothing to do with the energising signal waveform. Hence a need for superior power rail capability with sealed box bass drivers.
Of course NFB/dominant pole does not cover all perceptible differences. More arise due to the extremely low driving resistance of high NFB amplifiers. Any (series output choke and cable) impedance between a loudspeaker and an amplifier can become fractionally energised both by forward current flow wrt the loudspeaker and loudspeaker system generated back-emf wrt the amplifier, the resultant appearing at the loudspeaker terminals being a distortion of that at the amplifier's NFB controlled terminals. Different amp/cable/spkr - different sound. The differences might be only 0.1dB at some frequency when measured with a steady sinewave, but the alteration of voice/music reproduction can be a difference between acceptable and really awful.
I am soooooo fed up of seeing thd figures and frequency response curves in specification sheets, these tell us nothing about how an amplifier will sound.
We need to present a phase response figure in place of frequency, and as a first approach to getting away from time isolated thd measurement at output, I examine amplifiers for their first cycle distortion wrt input at 10kHz. The best in this regard so far is the JLH class-A and variants (Nelson's DoZ). These class-As are also as near as good as is possible when reverse examined for NFB/dominant pole induced error.
Don't forget that the output impedance is NFB loop generated, this is variable with amplitude, frequency and loading, and holds only as long as the input or load has not rendered the amplifier non-linear. Are theoretical 'design' parameters checked in many working real world situations, or simply assumed to be theoretically covered?
Hi anatech,
Yes. The output impedance of most NFB amplifiers rises sharply beyond 20kHz, often becoming negligible by 1MHz, also, in quadrature or worse (as you say - amplifying), so that some amplifiers haven't a hope of being click free, unless, sited behind/beneath a loudspeaker as monoblocks and away from AC cables.
Hi Lumenauw.
Your post #99, exactly!!!!!
My R+L+C is only a starting point, but it can illustrate how crossover networks pose a significant problem for all NFB amplifiers.
The products developed by back-emf are only fractional, but they can be MUCH greater at a loudspeaker terminal than those observed when resistor testing an amplifier for linearity to better than as little as 0.1%thd, no matter what importance is placed upon harmonic grading.
Indeed, from the above 'musical harmonic' contributions within this column, I see evidence for an increased need of aural review of amplifier performance, no matter what bench or simulator measurement might imply.
Cheers ........... Graham Maynard.
Re your post #91.
But of course! An R+L+C model is only a first step, so it is either a case of simulating with some more realistic loads, or endless discerning empiricism with critical reviewing of the outcome.
As you say - the load is delaying, but as Wimms has noted with R+L+C alone, the current often becomes leading.
This is my reason for examining amplifiers when reverse driven too. The worst fractional electrical disturbances tend to arise as a result of crossover reactivity around the upper crossover frequency. In real life I have also noted some serious problems due to dynamically induced bass driver back emf arising when there is inadequate amplifier power rail capability.
An amplifier that has never been tested with suddenly starting sinewaves, input and output, cannot be assumed to be satisfactory at maintaining dynamic loudspeaker control when compared to one that has.
This is also my understanding as to why NFB/divider linked bridged amplifiers sound inferior to those that are driven by an external phase splitter. Errors fractionally circulate, instead of being finite, and those errors manifest as a 'confusion' or 'muddying' of higher frequencies due to lower frequency drive.
Hi jcx,
I understand your point about electromagnetic inefficiency, but occasionally a loudspeaker can be dynamically driven to store considerable amounts of 'air-spring' energy, and this is not just reverse transduced as a phase shifting component, but at the fundamental generating frequency for the 'charged' system, and this comes back as a delayed leading current wave that has nothing to do with the energising signal waveform. Hence a need for superior power rail capability with sealed box bass drivers.
Of course NFB/dominant pole does not cover all perceptible differences. More arise due to the extremely low driving resistance of high NFB amplifiers. Any (series output choke and cable) impedance between a loudspeaker and an amplifier can become fractionally energised both by forward current flow wrt the loudspeaker and loudspeaker system generated back-emf wrt the amplifier, the resultant appearing at the loudspeaker terminals being a distortion of that at the amplifier's NFB controlled terminals. Different amp/cable/spkr - different sound. The differences might be only 0.1dB at some frequency when measured with a steady sinewave, but the alteration of voice/music reproduction can be a difference between acceptable and really awful.
I am soooooo fed up of seeing thd figures and frequency response curves in specification sheets, these tell us nothing about how an amplifier will sound.
We need to present a phase response figure in place of frequency, and as a first approach to getting away from time isolated thd measurement at output, I examine amplifiers for their first cycle distortion wrt input at 10kHz. The best in this regard so far is the JLH class-A and variants (Nelson's DoZ). These class-As are also as near as good as is possible when reverse examined for NFB/dominant pole induced error.
Don't forget that the output impedance is NFB loop generated, this is variable with amplitude, frequency and loading, and holds only as long as the input or load has not rendered the amplifier non-linear. Are theoretical 'design' parameters checked in many working real world situations, or simply assumed to be theoretically covered?
Hi anatech,
Yes. The output impedance of most NFB amplifiers rises sharply beyond 20kHz, often becoming negligible by 1MHz, also, in quadrature or worse (as you say - amplifying), so that some amplifiers haven't a hope of being click free, unless, sited behind/beneath a loudspeaker as monoblocks and away from AC cables.
Hi Lumenauw.
Your post #99, exactly!!!!!
My R+L+C is only a starting point, but it can illustrate how crossover networks pose a significant problem for all NFB amplifiers.
The products developed by back-emf are only fractional, but they can be MUCH greater at a loudspeaker terminal than those observed when resistor testing an amplifier for linearity to better than as little as 0.1%thd, no matter what importance is placed upon harmonic grading.
Indeed, from the above 'musical harmonic' contributions within this column, I see evidence for an increased need of aural review of amplifier performance, no matter what bench or simulator measurement might imply.
Cheers ........... Graham Maynard.
- Status
- Not open for further replies.