I see data of preamps referring to max. level at input and output, but what about the smallest levels of a musical signal coming in and out.
I know that the smallest level of actual music signal that can be retrieved is relative to the system noise floor, but as a test and measurement example:
Say that signal levels of 10 uA@ 10 mV are presented at the actual physical connector and or switch, and that the the pre amp is capable of transferring or responding to that level.
Is this considered to be good low level resolution or are their figures better than this with examples of such pre amp designs.
What should the specs of a well designed pre amp be regarding low level quality signal reproduction?
I know that the smallest level of actual music signal that can be retrieved is relative to the system noise floor, but as a test and measurement example:
Say that signal levels of 10 uA@ 10 mV are presented at the actual physical connector and or switch, and that the the pre amp is capable of transferring or responding to that level.
Is this considered to be good low level resolution or are their figures better than this with examples of such pre amp designs.
What should the specs of a well designed pre amp be regarding low level quality signal reproduction?
Last edited:
Interestingly, you can often hear signals below the noise floor.
Analog circuits rarely have a "resolution" limit except in the imaginations of magazine writers, it's more a matter of how deep below the noise floor before you can't hear things. Digital circuits have resolution limits, but the whole purpose of dither is to circumvent that- which it does.
Analog circuits rarely have a "resolution" limit except in the imaginations of magazine writers, it's more a matter of how deep below the noise floor before you can't hear things. Digital circuits have resolution limits, but the whole purpose of dither is to circumvent that- which it does.
Yes, I do understand about ability to hear music signal within noise as in tape systems, vinyl, etc.
As for analog rarely having resolution limits. I think I see what you mean, but even if the noise floor were lets say even -150 dB below a micro-volt!!, what type of quality can we theoretically expect.
I mean, we are limited to the lowest level of the recording or transmission of source.
So, from my experience, musical signal levels of say minus 80 dB below 0 (0dB referred to .775 volts) are practically speaking, about at the limits depending, of course, on the level of monitoring and the total system noise right on through to the monitors.
I was just wondering on the actual quality of such signals, noise not withstanding, and what type of manufacture of pre amps behave well in this regard.
Specifically about the technical descriptions given at IRD Audio's LLC-P Purest pre amp page, Quoted in part:
... "To maintain short signal paths with a minimum of wires special relays are used for audio signal switching. High tech telecommunications relays from Japan have been selected for the job. All contacts are gold clad silver and signal levels as low as 10 uA@ 10 mV are easily accommodated for superb low level music detail."
Without getting into the gold-clad silver construction, I am curious about the 10 uA @ 10 mV wording.
What does this mean in practical terms please?
search for: Shannon Hartley Channel Capacity Theorem
in short there is a limit on information rate (in bits per second) that can be conveyed by a finite bandwidth Analog channel with additive noise
"hearing below the noise floor" is bandied about rather loosely in audiophile circles - usually people haven't adopted a sophisticated enough view - you need to include frequency dependence of "the noise" and some properties of our auditory system
auditory critical bands are measures of how far in frequency signals or noise have to be separated to be perceived simultaneously
there are some interesting correlation "unmasking" phenomena but I haven't seen papers showing that the effects really reach below the accepted frequency dependent hearing noise floor - which is only reached after minutes of accommodation in total silence - not the level we usually encounter in music venues
then you need to learn about thermal noise in resistors and shot noise in electronic amplifiers - there is no possibility of detecting/recording "-150 dB below a micro-volt" at room temperature, audio bandwidth
really extreme, expensive electronics could maybe give ~140 dB S/N over the audio band but would require Watts power dissipation in feedback resistors (assume signal electronics could use 10 rms max signal vs 1nV/sqrt(Hz)*sqrt(20KHz), ~ 50 Ohm max resistance)
once you have the electronics then you simply can't transduce audio sound levels into electrical signal at that dynamic range - microphones and their preamps are much noisier
in short there is a limit on information rate (in bits per second) that can be conveyed by a finite bandwidth Analog channel with additive noise
"hearing below the noise floor" is bandied about rather loosely in audiophile circles - usually people haven't adopted a sophisticated enough view - you need to include frequency dependence of "the noise" and some properties of our auditory system
auditory critical bands are measures of how far in frequency signals or noise have to be separated to be perceived simultaneously
there are some interesting correlation "unmasking" phenomena but I haven't seen papers showing that the effects really reach below the accepted frequency dependent hearing noise floor - which is only reached after minutes of accommodation in total silence - not the level we usually encounter in music venues
then you need to learn about thermal noise in resistors and shot noise in electronic amplifiers - there is no possibility of detecting/recording "-150 dB below a micro-volt" at room temperature, audio bandwidth
really extreme, expensive electronics could maybe give ~140 dB S/N over the audio band but would require Watts power dissipation in feedback resistors (assume signal electronics could use 10 rms max signal vs 1nV/sqrt(Hz)*sqrt(20KHz), ~ 50 Ohm max resistance)
once you have the electronics then you simply can't transduce audio sound levels into electrical signal at that dynamic range - microphones and their preamps are much noisier
Phil: First, stop trying to make sense out of high end audio marketing literature. Apply a little Ohm's Law to that meaningless bit of brochure-speak
Second, let's make sure you understand the distinction between dynamic range and resolution, two different things which you're conflating. "Resolution" has just as much meaning between (say) 1.00000V and 1.00001V as it does between 0.00000V and 0.00001V.
Second, let's make sure you understand the distinction between dynamic range and resolution, two different things which you're conflating. "Resolution" has just as much meaning between (say) 1.00000V and 1.00001V as it does between 0.00000V and 0.00001V.
Resolution and detail depend mostly on how carefully you are willing to listen to a system. And since your willingness to listen carefully more often than not depends on how well you like the sound of the system, you might say that detail and resolution are dependent more on distortion levels than noise floor issues, both for psycho-acoustic as well as technical reasons.
John
John
Ok SY, I'm trying.
Lets see. Dynamic range as understood by me, a violinist:
The least or absolutely quietest qualified note signal that can be distinguished as useful over that, which is purely bow noise at one end, and the range or amount of output power in musically acceptable terms representing the limit at which the instrument will begin to go no further by producing more noise than useful tone... Instruments dynamic range.
Resolution. the ability of the hearer or musician and the instrument to produce the smallest difference of level in output, (plus or minus) while also considering those level differences at various starting levels, such as resolution near the top extremities where the ability to produce differences is running out of head room, differences at the lowest extremes where resolution suffers from the same constraints as the former.
About the meaningless brochure-speak.
I can't say, so I'll take your word on it for now.
I believe that I did not make my self clear originally.
Also, jcx and jlsem brought out useful replies as well, and I do mostly understand their replies.
Let me ask in a different way.
Given a quality pre amp, in my terms, would mean that low level resolution would be at some level just above any distractions from noise and distortion.
At this condition does the response to the slightest dynamic, as well, any other useful measurement and listening quality - in that lowest range - have a different behavior when compared to the same conditions of dynamics/resolution when near the top of the pre amps range?
In other words, are these conditions - performance near the bottom scale versus the top end, characteristically different?
I would think so because they are all subject to basic laws of constraints.
As pointed out from jslem, it also depends on the listener, their resolution in hearing, their commitment towards that goal musicologically, etc.
Hi John:
You have a well documented history as a designer of quality audio devices.
I was really hoping for more comment on this matter, if you please.
About easy measurements:
Well do you have any experience with difficult measurements?
It's possible that some may not agree with what I will say, but even in audio electronics engineering, sometimes referred to a mastered and mature discipline, I see that we are not finished yet.
I 100% believe that new approaches, both in measurement and defining what a system should do and feature are not fully explored.
In simple terms, the dynamic behavior of a spring or speaker when subject to various applicable signals at small signal versus large, are not proportional.
This is what I am asking.
What type or make or pre amp is close to this ideal and or what should one look for - besides hearing - in design philosophy?
The problem is indeed definition- once you define what you mean, you can measure it. The problem is that "resolution" is tossed around colloquially. Let me give some quantitative examples:
Put 1V of DC into a circuit. You get 1V out. Put 1.000001V in. Does the output move? Is it distinguishable from the 1V case?
Put 1V at 1kHz into a circuit. You get 1V out. Put 1.000001V in. Does the output move? Is it distinguishable from the 1V case?
Put 1V of pink noise into a circuit. You get 1V out. Put 1.000001V in. Does the output move? Is it distinguishable from the 1V case?
Put 1V at 1kHz into a circuit at the same time that you have 1V of pink noise. You get 1V out. Put 1.000001V of 1kHz in. Does the output move? Is it distinguishable from the 1V case?
Four different definitions, four different measurements (all do-able), probably four different results.
Put 1V of DC into a circuit. You get 1V out. Put 1.000001V in. Does the output move? Is it distinguishable from the 1V case?
Put 1V at 1kHz into a circuit. You get 1V out. Put 1.000001V in. Does the output move? Is it distinguishable from the 1V case?
Put 1V of pink noise into a circuit. You get 1V out. Put 1.000001V in. Does the output move? Is it distinguishable from the 1V case?
Put 1V at 1kHz into a circuit at the same time that you have 1V of pink noise. You get 1V out. Put 1.000001V of 1kHz in. Does the output move? Is it distinguishable from the 1V case?
Four different definitions, four different measurements (all do-able), probably four different results.
"Specifically about the technical descriptions given at IRD Audio's LLC-P Purest pre amp page, Quoted in part:
... "To maintain short signal paths with a minimum of wires special relays are used for audio signal switching. High tech telecommunications relays from Japan have been selected for the job. All contacts are gold clad silver and signal levels as low as 10 uA@ 10 mV are easily accommodated for superb low level music detail."
Without getting into the gold-clad silver construction, I am curious about the 10 uA @ 10 mV wording.
What does this mean in practical terms please?"
It is true but also marketing **!
High tech means nothing.
All telecoms relays for signal switching are like that and most, if not all come (came!) from Japan.
Gold over silver is standard for contacts that don't switch a large enough DC current to clean the contacts.
The 10 uA @ 10 mV bit is because these relays are specced and hopefully tested to switch these levels reliably. Not a jot to do with audio.
Translation: We took a standard telecoms signal relay from the catalog(ue)s.
Edit added: They really screwed up! " ...as low as 10mV .." they lifted out of the relay datasheet, as indicated above. 10mV is actually several times LARGER than a cartridge average output!
Pure nonsense!
... "To maintain short signal paths with a minimum of wires special relays are used for audio signal switching. High tech telecommunications relays from Japan have been selected for the job. All contacts are gold clad silver and signal levels as low as 10 uA@ 10 mV are easily accommodated for superb low level music detail."
Without getting into the gold-clad silver construction, I am curious about the 10 uA @ 10 mV wording.
What does this mean in practical terms please?"
It is true but also marketing **!
High tech means nothing.
All telecoms relays for signal switching are like that and most, if not all come (came!) from Japan.
Gold over silver is standard for contacts that don't switch a large enough DC current to clean the contacts.
The 10 uA @ 10 mV bit is because these relays are specced and hopefully tested to switch these levels reliably. Not a jot to do with audio.
Translation: We took a standard telecoms signal relay from the catalog(ue)s.
Edit added: They really screwed up! " ...as low as 10mV .." they lifted out of the relay datasheet, as indicated above. 10mV is actually several times LARGER than a cartridge average output!
Pure nonsense!
Last edited:
There was actually a lead-up to that conclusion. I say that hearing better "resolution" is psycho-acoustic in nature since the "noise floor" in most music is the accompanying music. For example, if our violinist OP were listening to a violin concerto, he would want to be able to resolve the nuances of the soloist's playing within the "noise floor" of the orchestra.
John
John
Last edited:
Thanks Cliff:
Well, that explains a lot regarding marketing and the relays.
Although, I assume that this arrangement is not unusual and that it is possible that the IRD pre amp is a decent performer for the asking price-off-the-shelf, that is.
I received a few comments on the IRD in another post saying that it was 3 times over priced for the DIYer, but three times under priced for retail.
I will purchase the unit for 30 day review and we'll see.
=============================================
SY:
Understood, I see a semantic issue here, since the movement of 1.000001V over 1.000000V is an increase, it could be called a dynamic change - in level.
I believe you were also implying that the addition of pink noise, a more complex signal, with a DC or sine wave at whatever frequency would or should produce the difference in level, at least.
I can accept your examples as resolution, but when measuring various music signals, hearing evaluations not withstanding, the complexity of such signals will require a more detailed examination other than, and in addition to, level changes.
How fast can a change in level take place from in to out?
What is the envelope of those rates both rising and falling?
What type of envelope signature, on the DUT, and how does this vary with envelope of input signal?
At what frequencies and levels do those behaviors vary, or does the DUT have a generalized behavior except at the low and high end of operation, etc?
About accuracy, I would sum it all up with just that one word.
But back to original question:
In a good pre amp and in general, if the range of best performance in voltages is say between 1 volt and 10 volts, just to make it simple, can I assume that accuracy with complex music signals whose dynamic range over a give time is +/- 2 volts, referenced to 5 volts will be very close (accuracy) if the same is referenced to 7 or 3 volts?
Well, that explains a lot regarding marketing and the relays.
Although, I assume that this arrangement is not unusual and that it is possible that the IRD pre amp is a decent performer for the asking price-off-the-shelf, that is.
I received a few comments on the IRD in another post saying that it was 3 times over priced for the DIYer, but three times under priced for retail.
I will purchase the unit for 30 day review and we'll see.
=============================================
SY:
Understood, I see a semantic issue here, since the movement of 1.000001V over 1.000000V is an increase, it could be called a dynamic change - in level.
I believe you were also implying that the addition of pink noise, a more complex signal, with a DC or sine wave at whatever frequency would or should produce the difference in level, at least.
I can accept your examples as resolution, but when measuring various music signals, hearing evaluations not withstanding, the complexity of such signals will require a more detailed examination other than, and in addition to, level changes.
How fast can a change in level take place from in to out?
What is the envelope of those rates both rising and falling?
What type of envelope signature, on the DUT, and how does this vary with envelope of input signal?
At what frequencies and levels do those behaviors vary, or does the DUT have a generalized behavior except at the low and high end of operation, etc?
About accuracy, I would sum it all up with just that one word.
But back to original question:
In a good pre amp and in general, if the range of best performance in voltages is say between 1 volt and 10 volts, just to make it simple, can I assume that accuracy with complex music signals whose dynamic range over a give time is +/- 2 volts, referenced to 5 volts will be very close (accuracy) if the same is referenced to 7 or 3 volts?
John, I did not understand fully, your sentences:
"noise floor" in most music is the accompanying music".
Would you rephrase that please?
... "Noise floor of Orchestra"?
Do you not mean to say noise floor of the auditorium?
... " hearing better "resolution" is psycho-acoustic in nature..."
Provided that one's hearing is good, health wise, hearing better or hearing with great resolution is a matter of training. But, again it may be that I've misinterpreted that sentence.
I do want to understand what you are saying, so please consider re-writing it.
I think what John is saying is, "Can you hear if the third violinist in a symphony is playing a little scratchy?"
Except in pathological cases (or the usual bad high end designs, but I repeat myself), that's not a function of the preamp, more of the recording. It takes a special effort to screw that up at line level.
Several orders of magnitude faster than the fastest risetime of an audio signal.
Except in pathological cases (or the usual bad high end designs, but I repeat myself), that's not a function of the preamp, more of the recording. It takes a special effort to screw that up at line level.
Several orders of magnitude faster than the fastest risetime of an audio signal.
But a trained ear is only good as long as you are prepared to make use of it. If you don't like the way a system sounds overall, you may not have the desire to make full use of your ability. And often when something is changed or modified in a system, more careful attention is paid to listening and often results in the illusion of finer resolution. I've had the experience of someone adding a tweak and we all listen more intently and, lo and behold, yes, it seems more detailed. And then the tweak is removed later on and no difference is heard.
About the 3rd violinist playing a little scratchy.
It is doubtful from even optimum audience seating.
My inquires are about how the pre amp responds to the recording. If the scratchiness of the 3rd violinist can be heard at the record monitor and it's obvious to everyone at that position, we would expect to hear it on playback. Still not sure what John meant.
SY, what's not a function of the pre amp please?
I had asked how fast can a change in level (audio signal) take place from in to out? You replied... "Several orders of magnitude faster than the fastest risetime of an audio signal."
We already are talking about an audio signal, what is it you are referring to that is faster than the fastest rise of an audio signal.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.