I didn't try many for this speaker, class D for the low end (500Hz and below), class A GaNfet Circlotron (for above compression driver500Hz).
All amps sound different. Mostly different between topologies chosen. Why, well imho it's like food or wine: it's as if there's always a clear character that is inherent and stands out for the topology and/or the components used. I really don't know. Maybe someone with excellent knowledge about the human auditory system could make more insightfull, for me I just take certain things as a given since every amp has it's pros and cons, it's easier for me to choose a certain character and try to get it as good as I can. But, I don't think thd measurements per se are important. The easiest way to.make an amp sound good is getting it good without feedback and then use the feedback to take it a step further. Why e.g. an amp can sound faster but more artificial when comparing feedback vs non feedback remains a mystery to me. What comes into play is not feedback for bandwith but when feedback has to be used to enhance the voltage source behaviour. Just an example of why amps sound different even though they might measure similar on a passive 8 ohm resistor instead of a complex load. Etc etc. Btw those power opamps you use can be really nice!
All amps sound different. Mostly different between topologies chosen. Why, well imho it's like food or wine: it's as if there's always a clear character that is inherent and stands out for the topology and/or the components used. I really don't know. Maybe someone with excellent knowledge about the human auditory system could make more insightfull, for me I just take certain things as a given since every amp has it's pros and cons, it's easier for me to choose a certain character and try to get it as good as I can. But, I don't think thd measurements per se are important. The easiest way to.make an amp sound good is getting it good without feedback and then use the feedback to take it a step further. Why e.g. an amp can sound faster but more artificial when comparing feedback vs non feedback remains a mystery to me. What comes into play is not feedback for bandwith but when feedback has to be used to enhance the voltage source behaviour. Just an example of why amps sound different even though they might measure similar on a passive 8 ohm resistor instead of a complex load. Etc etc. Btw those power opamps you use can be really nice!
If you mean "faster" is reproduce attack time accurately, it is because the amplifier have higher slew rate. Slew rate can be high in amplifier using global negative feedback. But amplifier non global negative feedback is easier to get high slew rate.
Amplifier with global negative feedback is easier to get low distortion, but amplifier non global negative feedback is more difficult to get low distortion.
Global negative feedback or not is not important. The important is good measurement of slew rate, distortion, signal to noise ratio, etc.
You should try an amplifier that high slew rate and low distortion.
My priority for designing an amplifier is to get higher slew rate as possible but still stable (good phase margin and gain margin). It is easier to use CFA (Current Feedback Amplifier), but other topology can have high slew rate, too. It is depend of the designer skill.
Perception only enters into affecting the outcome of a measurement, when such a measuring, is an opinion. Perception cannot affect objective measurements like length, temperature, mass, force and the rest of physical properties. If one is assessing predominantly perceptive experiences, like any sensory experience, then, one is not doing any objective measurement. Without any means of objectively comparing conscious experiences, it is like groping in the dark to compare.
We know, as human beings, we are all perceptual beings. However, irrespective of this, no two human beings can be objectively certain, they experience the same green colour when they look at the same green colour shade. Human beings can only compare their perceptual experiences through language, both written and spoken, which again, depend on perception. Therefore, if person A compares their perceptual experience with person B, BOTH A and B, have first to interpret their perceptual experience, then communicate it. This comparison, has at least, four sources of error: two being the interpretation of a perceptual experience, and another two, being the transcription of these perceptual experiences into language.
In cases like these, I think, as I am NOT an expert in the field, the only solution to reduce dependence on subjective interpretation of a perceptual experience, is to use statistics. As with all statistics, individuals do NOT count. Only large populations of data elements count. That way, for instance, when assessing an amplifier's performance, one can ask such an assessment from a large number of qualified assessors and statistically process the results.
We know, as human beings, we are all perceptual beings. However, irrespective of this, no two human beings can be objectively certain, they experience the same green colour when they look at the same green colour shade. Human beings can only compare their perceptual experiences through language, both written and spoken, which again, depend on perception. Therefore, if person A compares their perceptual experience with person B, BOTH A and B, have first to interpret their perceptual experience, then communicate it. This comparison, has at least, four sources of error: two being the interpretation of a perceptual experience, and another two, being the transcription of these perceptual experiences into language.
In cases like these, I think, as I am NOT an expert in the field, the only solution to reduce dependence on subjective interpretation of a perceptual experience, is to use statistics. As with all statistics, individuals do NOT count. Only large populations of data elements count. That way, for instance, when assessing an amplifier's performance, one can ask such an assessment from a large number of qualified assessors and statistically process the results.
Last edited:
Putzeys has a nice paper about feedback, rise time/slew rate and bandwith:
https://linearaudio.net/sites/linearaudio.net/files/volume1bp.pdf
The slew rate is higher when using no feedback, but the bandwith is lower and the distortion higher. It's all about having a good balance between these.
To be honest, I never could tell an amplifier sounds faster because any of those numbers, I mostly am concerned about having to use feedback to mitigate drive difficulties of the load (mainly impedance related, second is power needed). The rest is easier because it's a design goal one can do on paper and in sims.
I agree. Yet, how does a measurement of an amplifier by e.g. a sine wave compare to the perception of music amplified by it, which is processed by ears and brains?
You can't expect to have different stimuli and different sensors, so to speak, and expect there's always a relation, regardless of errors within that process.
Also, when one would compare the whole chain it's way easier. Let's say I play violin or drums, record it and play it back. I'm sure many would agree there's no description of fast, slow, harmonics, imd etc etc when listening to the real thing. Yet when played back, these things are too often the standard lingo used.
One should better use naturalnes, transparancy or just "closest to the real thing" as a start in evaluating. If the rest of technical lingo needs to be used, you're already out the door in achieving the best amp. That's not to say one shouldn't measure, I wouldn't even dare myself, but it becomes a problem when all lingo is pointed towards finding technical descriptions and there's no way of relating those to the differences experienced when comparing to real instruments.
I hope I got this right a bit. Maybe need to think more to get this clearer in writing.
A good test of this idea is headphones. I have a pair of ATH 900 'Air' that cost me about $350 in 2010 when I lived in Japan. About a year ago I went into the local city here in the UK with $1000 burning a hole in my pocket to buy a new pair of headphones. Luckily I took my old phones with me. I bought nothing. They all sounded boxy, or too bassy or too hissy.
Some of these phones had very good reviews. What was wrong? I hadn't given my brain time to tune to the new sound - the headphones were all ok. I'll only buy new ones when these ones finally konk out. I decided no need to put my ears through a recalibration at this stage of life
A similar thing happened to me. I have to use closed headphones, and can't easily audition new ones, so when my AKG K540 headband and earpads deteriorated, I thought I might try something new. Juno had a pair of 'open box' Shure SRH1540 on offer at half price so I thought I'd take a punt based on the excellent reviews they get across pro-audio and domestic audio magazines.
It took a week or so to adjust to the drier, flatter more detailed sound of the Shure's, but after that I do believe I'm getting more out of music. I retired the AKGs to my office, but on using them again, even on my main system, I find them overblown in the bass response, and a bit claustrophobic in their presentation - even though I loved them before acclimatising to the Shure's!
I do believe the Shure phones are more accurate, but before hearing them, I was totally happy with the AKGs.
Basically, once beyond 'junk', all competently designed audio components are capable of giving great enjoyment - as long as you don't make the mistake of visiting shows and dealers!
Bruno write about amplifier using ordinary Miller compensation and VFA (Voltage Feedback Amplifier).
Width bandwidth open loop gain until 20kHz can be done with two pole compensation. In VFA slew rate depend on LTP current and how to compensated.
Full power bandwidth = slew rate / (2 * pi * Volatge output maximum)
In VFA is easy to get 300 kHz full power bandwidth.
But if CFA is easy to get 1 MHz full power bandwidth.
Higher slew rate can have low distortion. In Dr. Arto Kolinummi's book, he design two amplifiers, one with global negative feedback and other with no global negative feedback. Both have high slew rate and low distortion.
Dadod's CFA have high slew rate and low distortion. It is not only in simulation, but measured by R.N. Marsh. Another member also design such amplifier.
I think you have limited knowledge of designing an audio amplifier. You should read some book:
1. Audio Power Amplifier by Dr. Arto Kolinummi
2. Audio Power Amplifier Design Handbook by Douglas Self
3. Designing Audio Power Amplifiers by Bob Cordell
You don't like other viewpoints much, do ya?
Maybe you should explain why current feedback has anything to do with a difference in perception of sound quality.
As long as one can't be sure about the exact meaning of measurements wrt transparancy, ease of listening, depth of stage, things like being able to make all music seem to have a different feel, atmosphere etc when recording techniques change, the production changes etc etc measurements are just measurements.
Once again, taken into account a certain sufficient technical quality level, you can't compare different stimuli (music vs sine or square wave forms) and different sensory systems (AP, Clio vs ears and brains) to have any meaning.
If you do, then you're extracting meaning in measurements you can't seriously expect to correlate with sound quality.
So use real instruments for your ultimate comparison.
And the need for you to get personal about this speaks volume, you really ought to let that go.
Maybe you should explain why current feedback has anything to do with a difference in perception of sound quality.
As long as one can't be sure about the exact meaning of measurements wrt transparancy, ease of listening, depth of stage, things like being able to make all music seem to have a different feel, atmosphere etc when recording techniques change, the production changes etc etc measurements are just measurements.
Once again, taken into account a certain sufficient technical quality level, you can't compare different stimuli (music vs sine or square wave forms) and different sensory systems (AP, Clio vs ears and brains) to have any meaning.
If you do, then you're extracting meaning in measurements you can't seriously expect to correlate with sound quality.
So use real instruments for your ultimate comparison.
And the need for you to get personal about this speaks volume, you really ought to let that go.
What about pink noise as a useful stimuli for both measurement and perception?
That still is the bottleneck in audio electronics.Probably because your speakers aren't fast enough.
so what would be the pertinent parameters in a speaker for fastness?
my halo integrated sounded quite slow to me in my last setup (straight up comparison to the yamaha it was replacing) my next speaker build was chosen to be faster sounding.
I went with pro components, b&c de250 and eminence delta 10's....my reasoning being strong motor with low mms (I think it's 33g) on the woofers, and high efficiency of a cd, combined with active xo should speed up the perceived sound .....maybe?
my halo integrated sounded quite slow to me in my last setup (straight up comparison to the yamaha it was replacing) my next speaker build was chosen to be faster sounding.
I went with pro components, b&c de250 and eminence delta 10's....my reasoning being strong motor with low mms (I think it's 33g) on the woofers, and high efficiency of a cd, combined with active xo should speed up the perceived sound .....maybe?
Some may question why a sine wave is used as the signal of preference to analyse the frequency response of an amplifier. The basis for adopting the sine wave, stems from mathematics, particularly Fourier's contribution. Mathematically, any periodic function can be expressed as a summation of sines of multiples of a fundamental frequency. The rule also applies to a sine wave, but trying to derive the Fourier Series of a pure sine wave results in the same exact sine wave. This result shows a sine wave cannot be reduced into simpler waveforms, and is therefore, pure. This basis is applied to signal processing in every frequency range.
Applying a sine wave to an amplifier is appying the simplest of signals, which is an advantage compared to using a complex signal of natural occurence. Moreover, since a sine is pure in itself, if an amplifier 'creates' other ghost frequencies, it would mean, it is distorting the original signal by adding ingredients to it that were not present originally.
If amplifier design engineers were to adopt using real audio signals instead of a sine wave, the complexity of such signals would make processing unnecessarily complicated. It can be done for a constant audio signal, but, the analysis will need every harmonic of the orginal signal to be considered.
Applying a sine wave to an amplifier is appying the simplest of signals, which is an advantage compared to using a complex signal of natural occurence. Moreover, since a sine is pure in itself, if an amplifier 'creates' other ghost frequencies, it would mean, it is distorting the original signal by adding ingredients to it that were not present originally.
If amplifier design engineers were to adopt using real audio signals instead of a sine wave, the complexity of such signals would make processing unnecessarily complicated. It can be done for a constant audio signal, but, the analysis will need every harmonic of the orginal signal to be considered.
Last edited:
If you want quality sound from speakers, it will cost you. Yes, this is where the money matters, not DAC, amp and cables. Try >$400 compression drivers and woofers.
True active x-over will help along with good room acoustics.
Pink noise is really helpfull for finding easy flaws indeed. It's propably not used enough as an indicator. Do you use it?
The problem however is that in order to know if you are missing bandwith (let's say top end or really low frequencies), there's no way to tell it is because it must be drawn from memory. Of course at that instance you can use a parallel electric or acoustic measurement to know if there is actually something missing. For short periods, I also like white noise, any sharpness like hiss stands out immediately.
The thing that isn't tested with noise stimuli is attack, transients etc. At least as far as I can tell, I never perceive any of that with listening to noise.
The issue I have with the whole idea of just feeding amps static stimuli and merely focusing on those results of trying them to get cleaner and cleaner, is just that in the end, there's an instrument that needs to be replicated as true as can be. The only thing that matters in the end is that instrument/choir/etc. So why stop at sine/square measurement stimuli.
Of course, the drawbacks are clear as day, there's never a musician around when you need them and comparing an amp can't be done without using the whole chain of recording electronics as well as evaluating the whole playback chain, so it's impossible to isolate, let's say an amplifier, for evaluation. It might also happen that one device equalizes the shortcomings of another in the chain. However, if your goal is to stay true to the original sound, then it's a really good way of knowing if your system alltogether, does just that.
Just for the record: I do believe measurements are the basic tools for developing amps etc. But you should hear a kick drum, snare or violin in real life in your room and I'd like to hear from you when that's done and you compare that with your stereo, if you really think that ultra low thd amps fixed a problem for ya. I just don't think, that above a certain basic quality level, that it's a very important concern anymore.
The problem however is that in order to know if you are missing bandwith (let's say top end or really low frequencies), there's no way to tell it is because it must be drawn from memory. Of course at that instance you can use a parallel electric or acoustic measurement to know if there is actually something missing. For short periods, I also like white noise, any sharpness like hiss stands out immediately.
The thing that isn't tested with noise stimuli is attack, transients etc. At least as far as I can tell, I never perceive any of that with listening to noise.
The issue I have with the whole idea of just feeding amps static stimuli and merely focusing on those results of trying them to get cleaner and cleaner, is just that in the end, there's an instrument that needs to be replicated as true as can be. The only thing that matters in the end is that instrument/choir/etc. So why stop at sine/square measurement stimuli.
Of course, the drawbacks are clear as day, there's never a musician around when you need them and comparing an amp can't be done without using the whole chain of recording electronics as well as evaluating the whole playback chain, so it's impossible to isolate, let's say an amplifier, for evaluation. It might also happen that one device equalizes the shortcomings of another in the chain. However, if your goal is to stay true to the original sound, then it's a really good way of knowing if your system alltogether, does just that.
Just for the record: I do believe measurements are the basic tools for developing amps etc. But you should hear a kick drum, snare or violin in real life in your room and I'd like to hear from you when that's done and you compare that with your stereo, if you really think that ultra low thd amps fixed a problem for ya. I just don't think, that above a certain basic quality level, that it's a very important concern anymore.
Probably because your speakers aren't fast enough.
So my speakers need feedback? [Insert laughing face]
Last edited:
I've only used pink noise recently in a simple listening test I did to see the effect of moving my head on the sound I heard. I was trying initially to see if I could hear any phasing effect or comb filtering between speakers but I found I was getting the same variation in sound even when listening to one full range speaker.
Am I the only one who changed opinion about the perceived speed of a speaker after changing the loudspeaker cables from thin (around 0.5mm2/17 AWG) to really thick (10mm2/7AWG)?
I'm curious how this relates to a transient response measurement. I never did try to verify with a measurement but am familiar with the sonic shift by it.
I'm curious how this relates to a transient response measurement. I never did try to verify with a measurement but am familiar with the sonic shift by it.
If I'm not mistaken, John Atkinson usually starts with noise or similar tones to get a first idea of the speaker he's reviewing.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.