Why do we hold amplifiers to such high standards?

[ampetrosillo]

Hi, I'm new to the forum and I'm a prospective builder (not really of hi-fi systems, although I would like to try my hand at a couple of speakers; I'm more interested in MI amplification). I'm *not* an audiophile by any means. I do care about fidelity in the right places (studio monitors, for instance) but I'm not the kind of person to get obsessed over it, and I realise that expediency and/or subjective taste come into play most of the time.

Now, the golden standard for a hi-fi amplifier worthy of its name is that it is capable of x watts at @0.1% THD+N (60dBs down). Some people expect more (for example 0.01% THD, or lower still). I accept this standard since it is so easy to achieve with "modern" electronics (really, anything built in the last 30 years or so; you can buy chip amps that are well capable of such figures) and I reject any notion of "pleasant low-order harmonic distortion" as complete bollocks for a hi-fi system (if you find IMD bad, and most people do, then you find harmonic distortion bad because that entails IMD). What I question though is how much it actually matters, since speakers (even if you cross them over actively, with separate amps, etc.) are always going to introduce far more THD than even Nelson Pass's amps

This is not to say that, now that I mentioned him, Nelson Pass is "right" (he is, of course, entitled to his opinion, and I suppose to his market niche). I'm not justifying the insanity that happens in the audiophile world, which to me is populated by people with more money than knowledge (I wouldn't even say "sense", because they actually stop much earlier), and I do know that the DIY crowd is often happy about (expensive!) subpar results just because they did it themselves. But of course I would like to understand the reasoning behind certain criteria. Also, how do you notice 0.1% THD if even the best speaker is going to produce, realistically speaking, 0.5-1% THD (that is, 14-20dB higher distortion) at normal listening volumes? What I'm asking is, I suppose, is that 0.1% figure just a conservative engineering target (based on the real world expectation that your listening levels are going to be around 60dB over the background noise, and that if you have a perfect speaker system, you need to achieve this figure)?

 

[DVDdoug]

Who is "we"?

I EXPECT most amplifiers (and other electronics) to have distortion and frequency response that are better than human hearing.

I've never heard distortion from ANYTHING that wasn't broken or overdriven. If I'm buying an amplifier, I'm mostly going to look at output power (if I can get actual measurements or if I believe the specs) and maybe the noise. But with noise, there are different ways of measuring it so you really need independent-comparable measurements so published specs aren't usually usable.

On his reviews here, Amir combines noise and distortion as SINAD and for amplifiers he graphs distortion vs output power so you can see where clipping kicks-in.

 

[Multicore]

0.1% THD+N doesn't seem like a very conservative engineering target. Take a look at the amplifier reviews here on ASR, as summarized by this chart. The highlighted one is reviewed here.

Nobody knows for sure what qualifies as the engineering specs for "transparent audio" from the point of view of subjective audibility. So some of us use other premises for our performance targets. For example, I would like my amp to be as clean as 16 bit-per-sample 44.1/48 kHz PCM. Other here claim to be able to hear more than 16 bits so maybe they want even cleaner amps. It's not very expensive these days to make clean amps so why not? It's comforting to know that the amp is not part of the problem, whatever the problem may be.

 

[IAtaman]

I am not sure if there is a standard or is it just convention to quote amplifier power where distortion reaches 0.1% at 1kHz for a certain load. In any case, that does not mean the said amplifier never produces distortion over -60dB.

Take Denon PMA-600NE for example. Its distortion at 1KHz at 5W is -75dB. At 10kHz that is a good 10dB higher.

This is a bit like the 0-60 of the car - it tells you how much power the car can pass on to the road and is a rough short hand for how well the car is engineered overall. 99% of us will never need to reach 100km/h in 3.9 seconds ever, but we like that in a car, because a car that can do that will most likely be responsive in day to day driving. That and the fact that we are all aged children.

 

[RayDunzl]

I have an old Amp vs Speakers measure...

Details:

Krell FPB 350 mcx

A quickee look and see... Bought 9 years ago used, no rebuild, no repairs, no faults. Serial number indicates 2005 build date. Left monoblock, approximately 5W/4Ohm tone, loaded by the speakers: Not great... Left Channel. 695 more rated watts available, though. Let's see: THD 0.012%...

www.audiosciencereview.com

Summary:

Amp measured driving speakers at about 5 watts via Focusrite Clarette 4Pre USB

Speakers simultaneously measured via UMIK1

Not a "clean room" here, so whatever could interfere in the power supply range simply is what it is. Cabling, components not isolated, etc.

 

[Zapper]

What I'm asking is, I suppose, is that 0.1% figure just a conservative engineering target (based on the real world expectation that your listening levels are going to be around 60dB over the background noise, and that if you have a perfect speaker system, you need to achieve this figure)?

In general these expectations are governed by what is readily achievable with current technology. 0.1% THD was an achievable spec in the 1970's in mass market solid state audio amplifiers. It turns out most people can't hear any improvement past that level, so it has remained a benchmark.

Now that very inexpensive amps can produce THD values orders of magnitude smaller, expectations are higher. Even though most of us still couldn't hear the improvement compared to those 1970's amps.

Some people get sucked into the "objective audiophoolery" of believing that huge SINAD numbers are essential to their listening enjoyment, or their OCD won't let them buy less than the best even though they suspect the differences are inaudible. I think a practical reason for avoiding the amplifiers on the red end of Amir's charts is that any competent designer can do better than that with technology available for 50 years. If the designer screwed that up, what else did he get wrong?

 

[kemmler3D]

Some people get sucked into the "objective audiophoolery" of believing that huge SINAD numbers are essential to their listening enjoyment, or their OCD won't let them buy less than the best even though they suspect the differences are inaudible

I keep hearing about these people, but I think there's just a few of them who pop into review threads to say stuff is bad. You don't see people in these threads coming in to defend 110dB SINAD as a necessity.

 

[ampetrosillo]

In general these expectations are governed by what is readily achievable with current technology. 0.1% THD was an achievable spec in the 1970's in mass market solid state audio amplifiers. It turns out most people can't hear any improvement past that level, so it has remained a benchmark.

Now that very inexpensive amps can produce THD values orders of magnitude smaller, expectations are higher. Even though most of us still couldn't hear the improvement compared to those 1970's amps.

Some people get sucked into the "objective audiophoolery" of believing that huge SINAD numbers are essential to their listening enjoyment, or their OCD won't let them buy less than the best even though they suspect the differences are inaudible. I think a practical reason for avoiding the amplifiers on the red end of Amir's charts is that any competent designer can do better than that with technology available for 50 years. If the designer screwed that up, what else did he get wrong?

The thing is, sometimes designers actually aim at higher distortion levels in expensive ways. Single ended class A operation which need huge heatsinks, with specialist devices that have little market, for maybe one watt of "cleanish" power.

 

[phantompowered]

As a shameless Benchmark fanboy, I retort:

< 0.0003 % THD+N at full rated power, 20 Hz to 20 kHz

Now that's a target value you can set your watch to. (I'm being glib, of course.)

I guess it "matters" because your amplifier, being at the end of the signal chain, is really what is determining the maximum theoretical dynamic range of your system - either bottlenecking or providing headroom for whatever comes before it - and physically controlling the drivers in your speakers. So it should do this as well as it possibly can.

We are of course dealing with theoretical performance limits that are far, far beyond the realm of human hearing to discern... but that doesn't stop engineers from trying, I suppose!

 

[Multicore]

The thing is, sometimes designers actually aim at higher distortion levels in expensive ways. Single ended class A operation which need huge heatsinks, with specialist devices that have little market, for maybe one watt of "cleanish" power.

I was overall rather confused by your topic post but then you summed up with "What I'm asking is..." so I focused on that.

If you want to have a rant about very expensive "audiophile" products with questionable features and benefits, be my guest. I find those threads a bit of a snooze but some people seem to like them.

 

[wwenze]

Do we? I'm fine with a $60 amp producing 0.03% THD+N

But a $600 amp better produce at worst 0.003% at a decently high wattage.

In a similar vein, we use $2 cables, but a $2000 cable better find a cure to cancer.

 

[ampetrosillo]

I was overall rather confused by your topic post but then you summed up with "What I'm asking is..." so I focused on that.

If you want to have a rant about very expensive "audiophile" products with questionable features and benefits, be my guest. I find those threads a bit of a snooze but some people seem to like them.

I was actually thinking more of "First Watt" products when I wrote this post. My original post was inspired by the insistence of a certain "audiophile" demographic to pay through the nose for stuff that would be only passable for an (expensive) electric guitar practice amp (or maybe into a dummy load as a studio effect, like a "bus processor", which is another corner of the industry where people overpay for ancient technology that could well be replicated by a couple of free plugins). But then I thought about the reason why we have certain standards, and the 0.1% THD figure *always* comes up when talking about fidelity (also in the DSP world; keeping aliasing products below -60dB is considered to be a basic requirement).

 

[ampetrosillo]

Do we? I'm fine with a $60 amp producing 0.03% THD+N

But a $600 amp better produce at worst 0.003% at a decently high wattage.

In a similar vein, we use $2 cables, but a $2000 cable better find a cure to cancer.

Yeah, and why are you fine with 0.03% THD and not, say, 0.5% or 1%? Because you actually care and can hear higher figures or because it feels stupid to settle for less (which would be a perfectly valid reason)?

 

[wwenze]

It's a $60 amp, amps don't go much cheaper than that, and you'd had to go to aliexpress for that.

I'd probably be fine with 0.1% too. But even the cheapest of amps already has performance better than what would bother a person, and also better than the speakers.

And a lot of audiophilia scams make use of this fact. Since we cannot actually hear the difference, they will hand-produce crap that is measurably even worse and sell it to you as exotic food.

 

[ampetrosillo]

It's a $60 amp, amps don't go much cheaper than that, and you'd had to go to aliexpress for that.

I'd probably be fine with 0.1% too. But even the cheapest of amps already has performance better than what would bother a person, and also better than the speakers.

And a lot of audiophilia scams make use of this fact. Since we cannot actually hear the difference, they will hand-produce crap that is measurably even worse and sell it to you as exotic food.

I suspect we may be able to hear a bit of the difference. Drive hard something that has 3-5% THD at that power level and it will possibly sound a bit more compressed and "glued" in a way that is hard to tell straight away but might be "subconsciously" perceptible. It could even be (probably "is", subjectively, to many, especially with music that hasn't been heavily smashed in mix/mastering) a bit pleasing, in the same way that when you do audio production you sometimes "warm things up" deliberately with distortion of some kind (like another topic here mentioned, reel-to-reel tape emulations are all the rage). I use those tools myself when I mix; but do I actually always want them? Usually when I finalise a mix (and that's BEFORE mastering, which will then compress and smash the thing to the Moon and back) I give it a try and, like, 50% of the time I turn it off. Imagine having that as a sound system.

 

[Descartes]

0.1% THD+N doesn't seem like a very conservative engineering target. Take a look at the amplifier reviews here on ASR, as summarized by this chart. The highlighted one is reviewed here.

View attachment 351729

Nobody knows for sure what qualifies as the engineering specs for "transparent audio" from the point of view of subjective audibility. So some of us use other premises for our performance targets. For example, I would like my amp to be as clean as 16 bit-per-sample 44.1/48 kHz PCM. Other here claim to be able to hear more than 16 bits so maybe they want even cleaner amps. It's not very expensive these days to make clean amps so why not? It's comforting to know that the amp is not part of the problem, whatever the problem may be.

To me unless they are in the blue I would not touch it!

 

[tmtomh]

Yeah, and why are you fine with 0.03% THD and not, say, 0.5% or 1%? Because you actually care and can hear higher figures or because it feels stupid to settle for less (which would be a perfectly valid reason)?

For me it's a combination of wanting an amplifier that can (1) match the quality capabilities of a typical source, e.g. for me CD-quality 16/44.1k digital audio and (2) provide some "buffer" or "wiggle room" in its specs to account for differing distortion levels at different volume/power levels and different frequencies, and also unit-to-unit variations.

I will say that I have noticed that my aspirations for amplifier distortion/SINAD performance increased when I saw that relatively affordable Class D options were available that provide excellent performance in this regard. And then my give-a-sh*t level promptly plummeted when I switched to Genelec actives, which contain good but not great amp modules in terms of their SINAD. From using a couple different passive speakers with my Purifi-based amp, and then listening to the Genelecs, it's blindingly obvious to me that I can't hear the difference in the amplifier modules' SINAD performance - and I'm confident that no one else could either.

 

[Brian Hall]

I want my amplifiers to do a great job of reproducing the music I am listening to without adding distortion or screwing up anything.

So for me, there is absolutely no use for tube amps or other "dirty" amps. It is not logical to use an amp that distorts or "colors" the sound.

 

[JSmith]

I'll just pop this here...

IMD - Something New

ESP Audio Pages - Intermodulation distortion - things you didn't know.

sound-au.com

Intermodulation Distortion

Intermodulation Distortion (IMD)

sound-au.com

JSmith

 

[radix]

0.1% THD was SOTA in the late 70s in an amp. And 0.005% was late 80s, early 90s. At least for McIntosh gear.

If a manufacturer is not at least 0.005% THD today, what are they doing? To me, it's a measure of competency.