I don't understand what you mean. Everything in this world should comply with science and should be in order. At some point, I don't even believe in "taste".
Jay,
"Sonic" is term for anything that has to do with sound.
Frank, in a word yes, current tests reveal everything required to get high fidelity from audio equipment and speakers. Which tests are quoted, and how they are quoted is another matter all together.
Jay,
High fidelity is about getting signal back out that was put in with only difference allowed being scale of magnitude.
Distortion and perceptions of distortion is relatively easy to cope with in high fidelity system; simply ignore studies pointing at irrelevance and uses equipment and drivers with measurable distortion level below levels perceivable under the most revealing test conditions. From here it is easy to reconstruct tests revealing distortion levels that are masked for most listeners when listening to normal program material.
When I follow these guidelines, and EQ my speakers to measure like this:
And likewise have flat phase response, I find that the greatest variety of recordings play back without any need to change anything other than the volume setting.
In terms of directly measurable fidelity:
In above picture, top track is waveform of recording on CD of a single drum hit, and bottom track is capture of speaker playing back the recording. Time interval is about 94ms.
Here is zoom in to 10ms view from cursor position in above picture:
Once again; review the definition of fidelity. Pursuit of fidelity in audio is the work of science.
"Sonic" is term for anything that has to do with sound.
Frank, in a word yes, current tests reveal everything required to get high fidelity from audio equipment and speakers. Which tests are quoted, and how they are quoted is another matter all together.
Jay,
High fidelity is about getting signal back out that was put in with only difference allowed being scale of magnitude.
Distortion and perceptions of distortion is relatively easy to cope with in high fidelity system; simply ignore studies pointing at irrelevance and uses equipment and drivers with measurable distortion level below levels perceivable under the most revealing test conditions. From here it is easy to reconstruct tests revealing distortion levels that are masked for most listeners when listening to normal program material.
When I follow these guidelines, and EQ my speakers to measure like this:
An externally hosted image should be here but it was not working when we last tested it.
And likewise have flat phase response, I find that the greatest variety of recordings play back without any need to change anything other than the volume setting.
In terms of directly measurable fidelity:
An externally hosted image should be here but it was not working when we last tested it.
In above picture, top track is waveform of recording on CD of a single drum hit, and bottom track is capture of speaker playing back the recording. Time interval is about 94ms.
Here is zoom in to 10ms view from cursor position in above picture:
An externally hosted image should be here but it was not working when we last tested it.
Once again; review the definition of fidelity. Pursuit of fidelity in audio is the work of science.
Wow, congratulations! You have a great system there. But don't forget that not everyone has the money to build such system. Most of us live in far from ideal sound reproduction, too many compromises such that the way we prioritize things will be different from one to another.
Also frank was right, that it is common an expensive system sounds bad simply because of environment settings. Room effect is always a problem to final response captured by ears (or mic in a measurement).
But if we show 5 similar measurement result of equally good systems. A blind test will give consistent ranking between the 5, while visual test of the graphics (without listening of the real thing) will be useless imo.
The music was fine, but the rendition of it wasn't -- as I mentioned earlier, there was another upmarket system on the other side of the room, played the same track, the ol' chalk and cheese thing. Where's fidelity when 2 expensive systems sound completely different?
Yes, this is the sort of testing that should be done, and more sophisticated versions of it, by a big margin. However, I'm now going to be a spoilsport, and point out that that the speaker output is showing easily 10% distortion - look at the 2.3340 time area as an example. In electronics we're worrying about whether -100dB distortion is audible, when the speaker capture shows levels at -20dB happening.
If we had fidelity, then those 2 waveforms should be a much closer match; how is the ear/brain supposed to know that the 2nd version really means the first version ...?
And now an interesting observation: tried a compilation of recent pop "hits" from the library on my ultra, ultra low end player plus TV combo, and it sounds like total sh!te! Of course, many of you would say that what all the recent stuff is about, it's very carefully mangled in the studios to be almost unlistenable to. At normal listening volume the TV speakers are rattling and falling apart, so the volume has to be turned down, and turned down, until it starts to sound vaguely reasonable. And then we have matchbox sound, a little transister radio on the bedside table sort of thing.
So what's happening here? Well, the mastering, and remastering, has made sure the virtual VU meters are sitting on the red line most of the time, plus a good chunk of bass bump boosting. Which means the poor power supplies in the TV set are being pushed hard, really hard even at normal volumes; no internal tweaking done, so they collapse badly under the stress. Thus, the producers of the music have effectively ensured that the tracks sound about the same on almost everything - barely acceptable ...
Having heard this sort of recording many times on a system with decent meat, it doesn't have to be this way. The HT cheapy does a good job here, because the power supplies have the grunt; the intrusive tweaking means no collapse of the sound when a reasonable volume is called for ...
So what's happening here? Well, the mastering, and remastering, has made sure the virtual VU meters are sitting on the red line most of the time, plus a good chunk of bass bump boosting. Which means the poor power supplies in the TV set are being pushed hard, really hard even at normal volumes; no internal tweaking done, so they collapse badly under the stress. Thus, the producers of the music have effectively ensured that the tracks sound about the same on almost everything - barely acceptable ...
Having heard this sort of recording many times on a system with decent meat, it doesn't have to be this way. The HT cheapy does a good job here, because the power supplies have the grunt; the intrusive tweaking means no collapse of the sound when a reasonable volume is called for ...
Last edited:
I'm curious about one thing. If we have to rely on the flatness of frequency response measurement. What if off-axis is up side down compared to on-axis? I have never assumed that I want to listen on-axis! I put my face 10 cm in front of the drivers. And if it sounds good there, I think it should sound good anywhere 
Subjective again I know.
But I guess that there are many builders fall into the trap of not recognizing the limit of measurement. But then again some body will ask me to show or mention some of them. Well, the only thing I can say is that there is no pure subjectivist as there is no pure objectivist. Those considered the purists, must have missed a lot of things.
Subjective again I know.But I guess that there are many builders fall into the trap of not recognizing the limit of measurement. But then again some body will ask me to show or mention some of them. Well, the only thing I can say is that there is no pure subjectivist as there is no pure objectivist. Those considered the purists, must have missed a lot of things.
You have mentioned several times that when you increase the volume, the power supply must work harder and “failed”. In my experience, rarely it is about power supply. Most chip amps (as in a TV set) are low power amps. Their distortion will jump considerably high when they have to be driven pass certain threshold. Many of them are specified with 10% distortion at rated power. So this rated power is useless because 10% is actually unbearable.
No one else has, either, but that's not even close to the whole story. It might be worthwhile for you to read Floyd Toole's book to get a better idea of what speaker measurements are all about and how their correlation with listener preference is done. You'll also find very complete off axis measurements in reviews from the late, lamented Audio and even Stereophile.
I will have to differ with you on this -- I've gone through the exercise over and over again: I retain the amplifying topology, but work hard on improving the performance of the power supply; this has always very substantially lifted performance levels, specifically in the area of allowing a very large range of playback volume with minimal change in tonality. And chip amps have benefited as much as discrete ...
I "learnt" this lesson decades ago: my first "serious" amp was a Perreaux biggie, which nominally had power to burn. But, no ... put your foot down a bit, and at a very precise volume level the thing started to choke. So, over several episodes of effort the standard power supply was disemboweled, and upgraded with more and more energy capacity. By the end it could do a fairly decent drum solo, at a good level ...
I still have this "monster", in working state, but have no special desire to slot it in. The chip amps are doing the job at least as well, are not holding back the sound.
Very good approach ... if people did that more often when the system was playing at realistic levels they would appreciate how dirty the sound often is, especially on "poor" recordings. That's why you don't need measurements to know when a setup is right: it's right when the sound still registers as "clean" with your face this close to a driver ...I put my face 10 cm in front of the drivers. And if it sounds good there, I think it should sound good anywhere
Amplifiers have a PSRR figure. Are we saying that as the stressed power supply is introducing more ripple, simply more ripple is making it to the speakers? Presumably it varies with frequency.
Or is the voltage dropping so low that the amp can no longer function properly at all, and distortion is rising disproportionately? Wouldn't this be apparent from driving the amp at the specified full power continuously into the expected load? (Not that all manufacturers don't indulge in some specmanship!)
Or is the voltage dropping so low that the amp can no longer function properly at all, and distortion is rising disproportionately? Wouldn't this be apparent from driving the amp at the specified full power continuously into the expected load? (Not that all manufacturers don't indulge in some specmanship!)
May be no difference at all. Like in previous case, I mentioned that I'm more into improving things than fixing things.
I'm pretty sure that I tend to use more advanced PS circuit and more audiophile approved parts
When we push any circuit to its limit, then we will be familiar with the bottleneck or the limitation of it. And in this case, I know very well "the sound of distortion". Same thing with discrete top amplifiers. Those who can reach 0.01% really is in a different league, soundwise.Look again at the distortion chart of the chip amp you used. Of course, you can always improve things even by simply changing a stock capacitor into a Black Gate. But nothing you can do to make them sing at volume where distortion is 1%
The design of amplifiers is largely based on the belief, and assumption, that the voltage rails are constant, irrespective of current load. Unfortunately, key aspects of the correct functioning of those circuits, including elements of the FB design are very sensitive to any modulation of those voltages, and the theoretical performance of the circuit won't be achieved.
If one runs Spice simulations of typical amplifier topologies, you will get remarkably good distortion figures IF the rails are assumed to be perfect. However, introduce anything like real power supplies to the model and those distortion measurements will degrade by orders of magnitude immediately. So, even in the virtual world of electrical modelling it is easy to find real effects of less than ideal PSR, which is that the distortion levels of the circuit are largely dependent on the stability of the power supply. And to make things worse, the effects escalate at higher frequencies; this is what one hears in real systems, the treble loses clarity at higher volumes, because the power supplies are modulating more severely, leading to higher distortion of the circuit.
If one runs Spice simulations of typical amplifier topologies, you will get remarkably good distortion figures IF the rails are assumed to be perfect. However, introduce anything like real power supplies to the model and those distortion measurements will degrade by orders of magnitude immediately. So, even in the virtual world of electrical modelling it is easy to find real effects of less than ideal PSR, which is that the distortion levels of the circuit are largely dependent on the stability of the power supply. And to make things worse, the effects escalate at higher frequencies; this is what one hears in real systems, the treble loses clarity at higher volumes, because the power supplies are modulating more severely, leading to higher distortion of the circuit.
Last edited:
One does have to look at the reality of what the content of a typical musical track is: there is relatively little of high amplitude, high frequency material, in other words in the region where that 1% distortion is. One could say that, perhaps, 99.5% of the music is occurring in the sweet spots of the amplifier, which is enough to allow the listening experience to be a good one.
That said, modern, highly compressed pop music is one of the worst offenders for pushing amplifiers of all persuasion into areas where they would prefer not to be. Something like the Adele 21 album is a system killer, because the amp is having to work very, very hard, and precisely at the same time, to hold the sound together - it took quite a bit of effort to recover decent replay from this album - but it is possible ...
So, I take it that you don't read any engineering books or journals and don't actually spend time discussing design with engineers.
Now, what???
OK, let's get into. Is it a matter of minutes, years, or how long to notice it?
.
Essentially correct. I'm looking at the material that's readily available, which always separates the functioning of the core circuitry, from its susceptibility to power supply fluctuations - they're always treated as separate issues. If you can point me to a discussion, and an analysis, that actually fully integrates the real world behaviours of the power supply with that of the audio circuit it's supplying, I would be interested in seeing that ...
I note in Cordell's amplifier book it takes about 5 minutes to absorb what he has to say about the matter, which is essentially that you have to worry about it, and test for it, as a separate issue from all the considerations of the intrinsic circuit behaviour - there is essentially no attempt to compare and contrast how the variations in amplifier design interact with non-perfect supply voltages.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.