measurement bollocks.

So when we measure a signal we look for distortion vs the source, so I'm thinking what is the correlation between taking a stereo signal and modifying it to place one element of the music in a different location spatially. You know just nudge it half way across the soundstage via mixing tools. What level of distortion would that change represent compared to the original signal.

If we are always going on about stage depth, width, height, space round instruments and all that bollocks why has no one ever done something really obvious like relocating one of the lead instruments in an orchestra and measure the difference in terms of the commonly used measurement sets?

I'm sure we can all hear the difference in moving a player within the sound-field, but I'm wondering if our ability to home into a narrow range of sounds within a huge sound-field and apply our unique processing power to temporal/spatial cues to define a 'position' for soemthing in the mix isn't actually beyond the capability of our measuring tools.

kind of like. "Measurement tells us everything", yeh well measure me the bit that tells us whether the guy on the trumpet is using an aluminium or rubber mute.

 

You can alter the relative positioning of an instrument within the mix by isolating it and panning it left or right, plus you can alter phase to alter the perceived positioning.
You can alter relative levels to to bring the performer closer, move him further away etc.

All of that can be measured either in isolation or as part of the whole.

Your guy on the trumpet will produce a different sound depending on the material from which the mute is made. The sound will have a different harmonic content, and that can be measured.

I don't really see your point.
If something is obvious enough to be audible it is easily obvious enough to be measured.

 

I also don't quite get your question. Most measurements have very specific input signals so we can look for specific output changes. Orchestral music is not normal a suitable input signal for any specific measurement.

 

If you move a performer within the orchestra, his position with the soundstage on playback will also have moved. You can certainly measure this.
So if the question is can a measurement detect that a performer has moved, well yes it can but only if you do the right measurement.
Run an overall spectral analysis and you won't get any difference. Run a spectral comparison on the two channels and you have your answer as it will reveal the change.

Much the same argument exists with THD.
It is a useful indicator but doesn't tell the full story. Look below the headline figures and all is revealed.

 

Increasingly one has to look so far below the headline figure into the small print (if any were supplied) that I basically disregard %THD these days as a descriptor beyong a very wideopen window... requiring an arbitrary 'low enough' is a perhaps a necessary, but nowhere-near-sufficient condition for recommendantion. A full description why so an immense task in itself.

%IMD , to a stated standard, is probably more helpful description of 'whats wrong with the sound' but much, much less commonly quoted, let alone understood widely. A shame.

- And of course both measures represents a different qualitative value judgement of the very same bent transfer function; which is also the root problem.

 

What I'm getting at is that there is almost zero correlation between the measured performance of an amp and how it actually sounds. Reviews talk in terms of meaningless numbers like 0.003% distortion, but these bare no relevance whatsoever to how an amp sounds.

Listeners can sit with two amps and describe one as providing more space around instruments in the mix, yet we can't point to the measurement for these qualities that we look for an identify as listeners.

What is the point of measurements that don't reflect anything about the human interaction with our equipment.

 

I'd probably argue that 'very low' say <0.01% means whatever is making up that figure can be disregarded. It can be exclusively unpleasant sounding high order stuff, or much less intrusive 2nd order. Once it gets below audibility it doesn't matter.
Better than 0.01 THD is easy to hit with even budget kit these days.

Back when any decent reviewer commissioned a full technical assessment to accompany the listening sessions, IMD was always measured. As was other important things such as performance below 1w and at full power.

Standards have slipped.

 

I just think that you are completely wrong Simon and don't accept the zero correlation argument at all.

In one sense you are correct, that comparing 0.005 with 0.003 is utterly pointless but only because both figures are better than required to remove the distortion as a factor in the comparison.

Try listening to an amplifier with lots of crossover distortion, or a SET with several % THD and can certainly hear it. But those are not competent amplifiers IMO.

Other things to look at are bandwidth. Is it curtailed within the audible range and if so by how much. That might be audible.
Noise performance matters, certainly if you run sensitive speakers.

Output impedance matters a great deal. Not just because of electrical damping but because it alters the response of the loudspeaker system connected. Martin has a simulation on his web site which is eye-opening but otherwise check the Stereophile amplifier reviews where response is measured into a range of simulated loads.

The classic case of subjectivists not seeing the wood for the trees is when comparing that smallest of amplifier, the phono stage.
Take a bunch of them, say six, any make and cost will do and then look at the accuracy of the RIAA correction. They'll all differ and often substantially in places. Some will roll sharply at LF, others will have mid humps, some have rising/falling top etc. Several dB across a broad section of audio.
Yet you still see people routinely describing the bright ones as detailed, the bass light ones as nimble and fast, and the humped ones as lively and tactile.
Overload margins need to be measured at a range of spot frequencies. With SS stages it often falls quickly at HF. IMD can get very poor if a SS phono stage is driven heavily. Cue those liking high output MC driving stages designed for low output MCs claiming them so sound more dynamic. Dynamic, or lively due to distortion? - no measurements and you're in the dark.
In other words, degrees of design goodness are being assigned to varying degrees of technical incompetency!
Measurement bollox indeed

These are all real, measurable effects that help explain amplifier differences when they occur.

Fortunately for most users, but unfortunately for some audiophiles there are hundreds of amplifiers out there that satisfy basic performance competency such that you can forget amplifier sound unless driving a very few unusual speaker loads.

We need more measurements, not fewer and that is part of the problem. A few basic figures might not reveal the true picture.
Pick up a copy of an early 80s A5 sized Hi-Fi Choice and take a look at the complexity of the testing. I guess it costs money to be that thorough though.

 

That's kinda my point Rob, magazine technical testing is currently pointless it is so gross in form to tell us nothing other than 'the distortion on this amp is not typically apparent'.

It's pointless numbers rolled out every single review that give no real indication to how amps sound. I'm far from saying measurement is pointless, we need more of it, but giving peak power into static loads, THD and FR is pretty meaningless.

I love the reviews in whichever US mag it is that prints pages of graphs, that's useful stuff. In the UK we fall so far behind this standard- we used to lead.

 

'Standards have slipped' - Yes. And consumers are undiscerning. And there are lies, damn lies, and statistics.

Just for an example as to why, after consideration, I cannot agree that an arbitrary measurment such that '0.01%' easily attained by budget amps' isn't anywhere near good enough: declaring and comparing measurement bandwidths.

Generic example of this - Class-D amps (and variants on this analogue switching approach) are in vogue; some of which even appear to measure reasonably on a %THD basis. The ones that apparently measure well usually use AP's own 6 -pole measurement filter at 20Khz, which basically means that spurs beyond 20Khz are not measured. It's done to deal with the (horrendous) output noise of the such switching amps on the measurement gear (which is telling, if you think about it) but the unintended consequnce is this: if you were interested in dissonant harmonics generated, say the 7th and above, well, its basically not measured beyond the low midrange, and certainly not representative of the effect on the critical presence range (7*3khz... utterly dissappears). So '%THD' will appear very low!

'%THD+N' would be a more honest measure, but again, measurement bandwidth needs to be stated at the very least.

Note I'm not suggesting stuff beyond 18-20Khz is directly audible. But it's effect in terms of 'mixing' and resultant sidebands (IMD)on inherent linearity in analogue systems very definitely is.

A quick experiment with RC roll-offs in front of preamps inputs can convince most of this - taking an amp that is notionally 'wideopen' and limiting it to say 50-70Khz usually sounds much cleaner. And conversely, removing such sensible BW limiting from an amp input is never an improvement (unless it was grossly ill-chosen, which is very rare). *

This aspect is a particular concern when, as with switching amps, the audio is on a medium -wave frequency carrier and therefore very, very likely to mixing with other stages due to simple proximity and poor EMI practices (which many audio-only designer are traditionally terrible at). A more common source of such problems these days would be streaming devices with shit RF feed-through performance; and universal cheap SMPS,; and a few other such candidates.

Quite how budget , or even rather expensive amps really audibly behave then is a crapshoot which is unlikely to accord with a simplistic reported %THD measure.

Sufficient measurement can yield all, I agree. But then...we're chasing dragons with a scope and not listening to music - which is what I'd rather do

 

Very true, but what you say - lies, damn lies, and statistics - is pretty much confined to the manufacturer seeking to show the product in the best light by selective quoting of test results, or arranging the test to produce the required result.
So I go back to slipping standards, it is up to the press to ensure that products are properly assessed using more than just a few basic tests.
Some do that but nowhere near enough.

So what we should be saying, and perhaps Simon as OP would also agree, is 'Specification Bollox', because proper and thorough measurements are never bollox.

 

> So I go back to slipping standards, it is up to the press to ensure that products are properly
> assessed using more than just a few basic tests.

Why? It is up to the press to make a profit and they do this primarily by being an attractive vehicle to the audiophile industry because the vast majority of their income, perks, products to test, etc... come from the industry and not the consumer. The content is designed to be attractive to those consumers that believe in the marketed properties of audiophile hardware such as most of the posters to this forum. It is not designed to be attractive to those with scientific/technical knowledge of sound, sound perception and audio hardware or, frankly, the non-technical with a bit of common sense.

 

It's certainly not set in stone, but having a full set of measurements for each product, one can usually find the reason, if it isn't just imagined.

Simon (actually anyone), you might find this an interesting read if you didn't see it already. Read from Distortion Metrics onwards: http://www.gedlee.com/downloads/Distortion_AES_I.pdf

Part II: http://www.gedlee.com/downloads/Distortion_AES_II.pdf

 

I wholeheartedly endorse ''stereophile style'' technical reviews of equipment.

I would swap pages of magazine speak for a mere 1/2 page of such information on equipment reviews.

I don't think tech specs tell the whole story, but I do think they are invaluable and sorely missed.

On this it looks like we have a first - - - a zerogain agreement

 

Chasing Dragons with a scope--that says it all for me what about those two things
we all have on the sides of our head .I think they are the best sound detector of all
and we get them for free. Noel W.

 

I believe we have arrived at a consensus. HTF did that happen?

 

> I wholeheartedly endorse ''stereophile style'' technical reviews of equipment.

Shouldn't that read stereophile style "technical" reviews of equipment?

> I would swap pages of magazine speak for a mere 1/2 page of such information on equipment reviews.

Why? Your attraction to valve amplifiers would seem to suggest the opposite.

> I don't think tech specs tell the whole story, but I do think they are invaluable and sorely missed.

The story of what? Specifications (assuming they are correct) are simply factual statements about performance. Those that argue against being provided with facts about what is being discussed are making a rather strong statement about themselves.

Of course, the information/meaning that one can extract from these facts relies on a knowledge about how things work. In order to be a passively consuming audiophile it is pretty much essential this knowledge is absent so that other beliefs can be held instead.

> On this it looks like we have a first - - - a zerogain agreement

What points of agreement can you see in this thread?

 

> Chasing Dragons with a scope--that says it all for me what about those two things
> we all have on the sides of our head .I think they are the best sound detector of all
> and we get them for free. Noel W.

Why do you think they are they best sound detectors of all?

 

Why?

Well, because if they'd don't they are as useful as a chocolate teapot.
The audio press used to be more thorough, more rigorous.

The alternative is for others to fill the void - forums & blogs, which is fine but they seldom have access to the required test gear.

 

Shouldn't that read stereophile style "technical" reviews of equipment?


who gives a rats ass?

or is it ... who gives a rat's ass?



> I would swap pages of magazine speak for a mere 1/2 page of such information on equipment reviews.

Why? Your attraction to valve amplifiers would seem to suggest the opposite.



YAWN..