Do different amps make a difference?

andy1249, as a percentage, THD+N in solid state amplifiers rises dramatically as the volume and therefore power levels are reduced. Check out the measurements page on any of the online Stereophile amplifier reviews.

The situation is exactly the opposite of what you say. Measuring THD+N at power levels just below clipping (and therefore very high volumes) gives misleadingly favourable results.

Total Distortion is made up of many distortion types.
Two main components are clipping type distortion ( main concern) which increases with level, and crossover distortion ( lesser concern ) which decreases with level.

Clipping distortion being the major contributor and least desirable (Most destructive), measurements are made for worst case scenario of this type.
 
Last edited:
andy1249, yes indeed there are many types of distortion that amplifiers can produce. In posts #21 to #23 we were talking about total harmonic distortion and noise (distortion). Here's the Stereophile THD+N vs power graph for the Boulder 2150 amplifier.

217B2150fig4.jpg


A side effect of clipping is that this graph goes up nearly vertically at the onset of clipping.
This graph cuts off at 100 milliwatts. IE a long way away from the one tenth of a micro-watt that I will be listening at late at night. The X axis of the graph would need to be about two and a half times bigger in order to extend it down to one tenth of a microwatt! You can also see that THD+N as a percentage increases dramatically as the volume is decreased from 500 watts to 100 milliwatts.
 
The biggest variation of the sound from a system is the room itself, (It dwarfs everything else) with the second being how the amp interreacts with the speakers, (Therefore active speakers tend to give a more accurate sound, as the amp and speaker drivers can be optimised for each other) so unless the measurements are done in the room that the system is in, it is not possible to determine the final sound via measurement alone. (Hence you must always try before you buy)

While there are many variables in amps, providing the designer follows the principle of straight wire with gain, (As most pro amps do) then providing it is not overdriven, it minimises the effect it has on the sound.

Hope this helps

Bill
Would you rate how the amp interacts with the speakers as being of more importance to the variation of sound in systems than the selection of the speaker drivers and the cabinet design?


And even if you do all measurements in room are you aware of any set of measurements that have ever been made on amplifiers that co-relates to how well that amplifier will recreate, for example, the sound of a grand piano? Amplifier frequency response and THD+N measurements do not come remotely close to doing that - in my experience.
 
So in laymans terms does this mean thd and signal to noise ratio matter even at small figures and that what you are saying is that with even very small differences under less than 0.05% on thd, it could matter in terms of sound quality and recreating realism as with piano. Therefore it's something to look out for if an amp is 0.01 or 0.05 in reword to thd.

That's how I've always understood it, that thd measurements do make a difference even though small. Am I Right?
 
andy1249, yes indeed there are many types of distortion that amplifiers can produce. In posts #21 to #23 we were talking about total harmonic distortion and noise (distortion). Here's the Stereophile THD+N vs power graph for the Boulder 2150 amplifier.

217B2150fig4.jpg


A side effect of clipping is that this graph goes up nearly vertically at the onset of clipping.
This graph cuts off at 100 milliwatts. IE a long way away from the one tenth of a micro-watt that I will be listening at late at night. The X axis of the graph would need to be about two and a half times bigger in order to extend it down to one tenth of a microwatt! You can also see that THD+N as a percentage increases dramatically as the volume is decreased from 500 watts to 100 milliwatts.

All amps have some distortion plus noise.
Clearly, as the measurement is a ratio of noise in output signal vs noise in input signal, the lower the level of signal the more the inherent noise in the amp is an issue.

Indeed it is possible to measure any amp with a spectrum analyser right down to the level where ratio becomes 1 ( THD +N = 100% ) and signal is indistinguishable from the amps baseline noise level.
Every amp has this level.
This level " Should" be vanishingly small.

The standard measurement is mainly concerned with clipping distortion, the assumption being that low level noise should be negligible at usable levels.

This is not always the case.
Indeed Many amps / receivers available at present have significant noise and crossover distortion at lower levels.
One of the many reasons it can be proven that not all amps are the same.
 
Last edited:
At one time, I got into trouble on here for naming and shaming an amp that had 10% ....yes 10% ....distortion at the 100w level and got worse from there.

It was popular at the time and a big seller.
A truly shocking piece of junk that seemed to be well reviewed everywhere you looked.
 
THD+N at one tenth of a micro-watt would be much more relevant to me than THD+N at 150 watts.

Its been bugging me...
What is your basis for 1 tenth of a microwatt? (100 nano watts )
Where are you getting this figure from?

You would be right in that no amp gets measured this low.
However Watts with regard to amps specify continuous output power into a specified load, i.e. An electrical measurement.

The only place I can find sound referenced at micro -nanowatt level is using the SI unit of sound power.
Of course , not the same thing at all!
Sound power - Wikipedia

Edit: for example 0.1 watt in SI sound power units is 100 milliwatts = to 110 db or Rock Concert level, whereas 100mw RMS continuous output power into 8 Ohm load in an amp is barely audible, if at all!
 
Last edited:
At one time, I got into trouble on here for naming and shaming an amp that had 10% ....yes 10% ....distortion at the 100w level and got worse from there.

It was popular at the time and a big seller.
A truly shocking piece of junk that seemed to be well reviewed everywhere you looked.

What was that amp Andy? I'm intrigued.
 
Would you rate how the amp interacts with the speakers as being of more importance to the variation of sound in systems than the selection of the speaker drivers and the cabinet design?


And even if you do all measurements in room are you aware of any set of measurements that have ever been made on amplifiers that co-relates to how well that amplifier will recreate, for example, the sound of a grand piano? Amplifier frequency response and THD+N measurements do not come remotely close to doing that - in my experience.

Clarification

The thread relates to amplifier sound, which is why I put the integration with the speakers 2nd, however, if you are talking a complete system, then the speakers have a far greater influence on the sound than any amplifier.

Measurement of sound differences heard

Record a Grand Piano in a typical room and store the measurements on a computer, then have a listen to the sound, next move the piano 12” and measure it again, (Using the same music and with everything in the room (Bar the piano) in the same place) then have a listen, you will notice the sound has changed, and this has been caused by the room acoustics.

Compare the 2 measurements, and you can identify the differences in the measurements for the difference you heard. (It’s no longer just a subjective change, but a measured change as well)

As to changing the sound of a piano, (Which occurred between the 2 amps) then studio engineers do this all the time, and with the resources available to them, can make it sound exactly as you want. (One of the biggest changes can be made with the use of EQ which alters the frequency response, to make it warmer, more full bodied, brighter and pretty much anything you want)

While the layman cannot explain the differences in the sound of amps using just measurements, a professional engineer will tell you straight away, for the simple reason that they have been professionally trained to.

Download the free Audacity program, load some music into it, (Piano for example) and play about with its features, (Start off with EQ) then after a time you will find what does what, and you will realise that buying an amp with a specific sound is pointless, as you can achieve it yourself via a simple software program. (Hence professional studios go for amps based on the straight wire with gain approach)

Hope this helps

Bill
 
Its been bugging me...
What is your basis for 1 tenth of a microwatt? (100 nano watts )
Where are you getting this figure from?

You would be right in that no amp gets measured this low.
However Watts with regard to amps specify continuous output power into a specified load, i.e. An electrical measurement.

The only place I can find sound referenced at micro -nanowatt level is using the SI unit of sound power.
Of course , not the same thing at all!
Sound power - Wikipedia

Edit: for example 0.1 watt in SI sound power units is 100 milliwatts = to 110 db or Rock Concert level, whereas 100mw RMS continuous output power into 8 Ohm load in an amp is barely audible, if at all!
One tenth of a microwatt produces about 33dbs at 1 m from my speakers. That's about the volume of the quieter instants when listening late at night.

100 milliwatts fed into my speakers would give me about 93 dbs. That's not barely audible, it's very LOUD!
 
Begins with O , ends with O

But that shows you shouldn't really buy these 'double o' home cinema amps for quality 2 channel audio.
 
One tenth of a microwatt produces about 33dbs at 1 m from my speakers. That's about the volume of the quieter instants when listening late at night.

100 milliwatts fed into my speakers would give me about 93 dbs. That's not barely audible, it's very LOUD!

Again , how are you getting your watt figure and what unit are you using?

Sounds very much to me that you are using the definition of watt as in SI unit of sound power.
Your figures are roughly in agreement with the SI Sound power chart.
Sound power - Wikipedia
This would be power as measured by an acoustic db/w meter at a Point in space from the source.

This is Not the same watt unit or definition that is commonly used for amps.
This measurement would be an electrical measurement taken across the load in the circuit.
Derived from RMS value of Voltage.
Audio power - Wikipedia

Completely different.
In which case, your reading/understanding of the stereophile charts is wrong in that all useable levels "are" covered in these charts.

Edit: added links to how these differing watt units are calculated.
Due to the different definitions it would be tedious to calculate a correlation, but the two definitions are at least 10^6 apart in value.
 
Last edited:
But that shows you shouldn't really buy these 'double o' home cinema amps for quality 2 channel audio.

That shows you shouldnt buy this amp at all.
 
Is 10% even possible in a modern amp? Was it not a buggered one?

I don't doubt your info' Andy, but I'm staggered that anything that bad could leave a factory as the finished thing.
 
Is 10% even possible in a modern amp? Was it not a buggered one?

I don't doubt your info' Andy, but I'm staggered that anything that bad could leave a factory as the finished thing.

It was 10% , no doubt about it.
Funny thing was, the amp was advertised as " feature loaded and great value" and when you actually re-read most of the reviews they never actually said one way or the other how it sounded, which was basically shit at any volume.
 
Glad I've never been tempted by an Onkyo.
 
One tenth of a microwatt produces about 33dbs at 1 m from my speakers. That's about the volume of the quieter instants when listening late at night.

100 milliwatts fed into my speakers would give me about 93 dbs. That's not barely audible, it's very LOUD!

Well following convention, if your speakers produce 93db with 100mW at 1m, (The standard measurement distance) then they have an efficiency around 103db at 1m for a 1W input, which is high even if they are horn speakers. (Most Hi-Fi speakers average 87db at 1m for 1W input)

I think we need clarification on the type of speakers used here.

Bill
 
Again , how are you getting your watt figure and what unit are you using?

Sounds very much to me that you are using the definition of watt as in SI unit of sound power.
Your figures are roughly in agreement with the SI Sound power chart.
Sound power - Wikipedia
This would be power as measured by an acoustic db/w meter at a Point in space from the source.

This is Not the same watt unit or definition that is commonly used for amps.
This measurement would be an electrical measurement taken across the load in the circuit.
Derived from RMS value of Voltage.
Audio power - Wikipedia

Completely different.
In which case, your reading/understanding of the stereophile charts is wrong in that all useable levels "are" covered in these charts.

Edit: added links to how these differing watt units are calculated.
Due to the different definitions it would be tedious to calculate a correlation, but the two definitions are at least 10^6 apart in value.
I am using the watts that the power amplifier is sending to my speakers.

Using power = voltage squared divided by impedance.


I know that 2.83 volts produces 103 dbs at 1 metre in front of my speakers. IE 1 watt from the amplifier to the speaker produces a measurement of 103 dbs on my sound meter when standing 1 meter in front of a single speaker. To my ears that sounds VERY LOUD at my listening position.

You could nit pick by saying the power sent to the speakers will depend on the impedance which varies with the frequency. My speakers have a nominal impedance of 16 ohms, dropping down to a minimum of 5.5 ohms in the bass region. I'll keep the maths simple by using an 8 ohm impedance figure.

2.83 volts squared divided by 8 ohms = 1 watt fed from the amplifier to the speakers.


Every decrease in power by 10 times gives us a reduction in volume of 10 dbs. That's the nature of the logarithmic db sound scale.

Therefore 0.1 watts produces 93 dbs at 1 metre in front of my speakers

10 milliwatts = 83 dbs
1 milliwatt = 73 dbs
100 microwatts = 63 dbs
10 microwats = 53 dbs
1 microwatt = 43 dbs
0.1 microwatt = 33 dbs

The Stereophile graph cuts off at 100 milliwatts - that's 73 dbs - a fairly generous volume. It does not come remotely close to measuring how much THD+N I'll be getting at lower volumes.

I have measured how many dbs I listen at (with the meter 1 m in front of my speakers) with a calibrated sound meter. Music is not a steady volume. For late night listening the peaks on a good recording will hit 60 something dbs with the troughs being down at 30 something dbs (or lower - we're hitting background noise levels here).


I don't think I can put it any simpler than that. If you're still struggling to understand you'll have to get give me more details of what you don't understand.
 
I know that 2.83 volts produces 103 dbs at 1 metre in front of my speakers. IE 1 watt from the amplifier to the speaker produces a measurement of 103 dbs on my sound meter when standing 1 meter in front of a single speaker.

I understand perfectly.
You are using the wrong units ( sound power )as I laid out in post 39.
As such , your calculations do not correlate with the Stereophile charts.

Stereophile are using electrical watt units and not sound power watt units.
You cannot mix and match these two fundamentally different units in calculations.
100mw electrical does not produce anywhere near 73dbs from a speaker.
You need to study the difference between the two definitions.
 
Last edited:
Well following convention, if your speakers produce 93db with 100mW at 1m, (The standard measurement distance) then they have an efficiency around 103db at 1m for a 1W input, which is high even if they are horn speakers. (Most Hi-Fi speakers average 87db at 1m for 1W input)

I think we need clarification on the type of speakers used here.

Bill
EV Patrician 800's.

Yes of course, you can do some simple calculations to see how much power your amplifier is feeding your speakers at 33 dbs. If you have 87db efficient speakers it will be 40 times more than my speakers. IE 33 dbs requires 4 microwatts from your amplifier. This is still well off the scale of the Stereophile graph. IE we'd need to extend the X axis of the graph by two and a quarter times to give us more meaningful results.
 
I understand perfectly.
You are using the wrong units as I laid out in post 39.
As such , your calculations do not correlate with the Sterophile charts.
Stereophile are using electrical watt units not sound power watt units.
No I'm not. I'm using the electrical watts output by the amplifier to the speakers!
I am using the correct units!

My speakers can handle 50 watts continuosly output from the amplifier. If I fed 50 watts output from the Boulder amplifier I'd measure a volume of 120 dbs from my speakers. If I fed 600 watts (it is a very powerful amplifier) to my speakers the voice coils would burn out like a filament lightbulb.


1 nominal watt of electrical power output from the amplifier produces 103 dbs from my speakers. As already mentioned, my speakers are 40 times more efficient at turning electrical energy into sound than many speakers. There are speakers - eg Klipshorns that are 1.5 times more efficient than my Patricians.



Edit: using this: Efficiency and sensitivity conversion - loudspeaker percent and dB per watt and meter loudspeaker efficiency versus sensitivity vs speaker sensitivity 1 watt = 2,83 volt box chart - sengpielaudio Sengpiel Berlin my speakers turn 12.5 % of the electrical energy fed to them into acoustic energy.
 
Last edited:
Dont want to turn this into a pantomime , as in yes you are no you arnt kind of thing.
You are mistaken , you have some studying to do.
 
IE we'd need to extend the X axis of the graph by two and a quarter times to give us more meaningful results.

No you dont , this mistake comes from mixing your units.

Take any point on the Stereophile electrical watt measurement chart.
Feed that to your speaker.
Take a db measurement and derive Sound Power watt units.
Result will be much much lower.

They are not the same units
 
Last edited:
Dont want to turn this into a pantomime , as in yes you are no you arnt kind of thing.
You are mistaken , you have some studying to do.
OK, let's take this 1 step at a time.

Do you agree with me or disagree that 1 watt output from an amplifier would produce a sound level of 103 dbs - as measured by a calibrated sound meter 1 m in front of my EV Patrician speakers?




Do you agree with me or disagree that 1 watt output from an amplifier would produce a sound level of 87.5 dbs - as measured by a calibrated sound meter 1 m in front of Magico S5 Mk II speakers?
Ref: Magico S5 Mk.II loudspeaker Measurements
 

The latest video from AVForums

Is 4K Blu-ray Worth It?
Subscribe to our YouTube channel
Back
Top Bottom