does it make sense to test how amps behave when they clip?

[Sir Sanders Zingmore]

I'm a big fan of Roger Sanders. He has a very measurement-based approach to his products and is of the view that amps do sometimes sound different but not for the reason that most audiophiles think.

to quote:
"Most audiophiles simply don't recognize when their amps are clipping. This is because the clipping usually only occurs on musical peaks where it is very transient, and does not occur at the average power level. Transient clipping is not recognized as clipping by most listeners because the average levels are relatively much longer than the peaks. Since the average levels aren't obviously distorted, the listeners think the amp is performing within its design parameters -- even when it is not.

Peak clipping really messes up the performance of the amplifier as its power supply voltages and circuits take several milliseconds to recover from clipping. During that time, the amp is operating far outside its design parameters, has massive distortion, and it will not sound good, even though it doesn't sound grossly distorted to the listener.

Instead of distortion, the listener will describe an amp that is clipping peaks as sounding "dull" (due to compressed dynamics), muddy (due to high transient distortion and compressed dynamics), "congested", "harsh", "strained", etc. In other words, the listener will recognize that the amp doesn't sound good, but he won't recognize the cause as simple amplifier clipping. Instead, he will likely assume that the differences in sound he hears is due to some minor feature like feedback, capacitors, type of tubes, bias level, class of operation, etc. rather than simply lack of power.

But his opinion would be just that -- an assumption that is totally unsupported and unproven by any evidence. Most likely his guess would not be the actual cause of the problem. "

so, assuming the above makes sense (it seems to, at least to me) I'm wondering if it's possible or useful to measure this "transient" clipping?
It might provide a data point for readers. But I guess on the other hand, it may be pointless as we should probably just advise people to buy amps that are powerful enough not to clip

 

[ninetylol]

Interesting. Curious to read what others think about this.

 

[Blur]

To me it makes sense to measure amps and DACs for that matter where the distortion is the worst and where the user would nominally use it. This way we could get a better picture of the overall performance over a range of distortions / output voltages.

In most cases unity gain works well as the voltage is pretty high for headphones.

 

[amirm]

Transient clipping is no different than regular clipping. The difference is in the fact that with music the transients don't look like plain sine waves and don't last as long.

In addition to clipping, transients can get distorted due to power supply voltage droop, feedback going crazy, and the two fighting with each other. As a result, the waveforms don't have the classic clipping we imagine. Here are some examples I have post before:

This is a Pioneer SC-61 class D AVR Amplifier:

Notice the oscillations prior to clipping.

This is another AVR with one channel barely clipping and the other not:

Notice that it is not a flat line but rather crooked.

This is an Onkyo with severe clipping:

Again, it is not a flat line.

Each one of these would obviously sound different.

 

[restorer-john]

I'm a big fan of Roger Sanders. He has a very measurement-based approach to his products and is of the view that amps do sometimes sound different but not for the reason that most audiophiles think.

to quote:
"Most audiophiles simply don't recognize when their amps are clipping. This is because the clipping usually only occurs on musical peaks where it is very transient, and does not occur at the average power level. Transient clipping is not recognized as clipping by most listeners because the average levels are relatively much longer than the peaks. Since the average levels aren't obviously distorted, the listeners think the amp is performing within its design parameters -- even when it is not.

Peak clipping really messes up the performance of the amplifier as its power supply voltages and circuits take several milliseconds to recover from clipping. During that time, the amp is operating far outside its design parameters, has massive distortion, and it will not sound good, even though it doesn't sound grossly distorted to the listener.

Instead of distortion, the listener will describe an amp that is clipping peaks as sounding "dull" (due to compressed dynamics), muddy (due to high transient distortion and compressed dynamics), "congested", "harsh", "strained", etc. In other words, the listener will recognize that the amp doesn't sound good, but he won't recognize the cause as simple amplifier clipping. Instead, he will likely assume that the differences in sound he hears is due to some minor feature like feedback, capacitors, type of tubes, bias level, class of operation, etc. rather than simply lack of power.

But his opinion would be just that -- an assumption that is totally unsupported and unproven by any evidence. Most likely his guess would not be the actual cause of the problem. "

so, assuming the above makes sense (it seems to, at least to me) I'm wondering if it's possible or useful to measure this "transient" clipping?
It might provide a data point for readers. But I guess on the other hand, it may be pointless as we should probably just advise people to buy amps that are powerful enough not to clip

Back in the old days when magazine reviewers actually tested amplifiers properly, all these issues were tackled, procedures were standardized and the testing was comprehensive. There is nothing new here, people have just forgotten what is important and argue against tough tests being reintroduced.

I can dig out numerous 1970s reviews where stability, performance at clipping, overload etc were not just talked about- they also presented oscillograms of the behaviour- in every review. Those tests are still valid today and I use them all the time.

At medium to high volumes, amplifiers clip often, regardless of how much power you have. I have numerous 200W+/ch power amplifiers and they all sound different at clipping. Transient clipping is the easiest to audibly differentiate them (and the most dangerous to tweeters). Anyone who thinks all similarly rated amplifiers will sound the same at their limits is mistaken.

 

[pma]

I have clipping like this in my amplifier (Y-axis not calibrated)
http://pmacura.cz/pmacfa_en.htm

 

[Sir Sanders Zingmore]

It would be interesting to see how valves clip compared to solid state (hard to generalise I know).
Isn’t the conventional wisdom that valves clip more “gently”. Or is that just another audiophile myth?

 

[sergeauckland]

For a long time now, I've been convinced that much if not all of audiophiles' all amps sound different, comes from uncontrolled listening sessions where amps are allowed to clip, and consequently what's being assessed isn't the amplifiers' inherent sound, but their different behaviour into clipping. Those examples Amir gave above are a case in point. Some amps clip cleanly and come out of clipping equally cleanly, others latch up, go unstable and oscillate or otherwise behave badly. This is especially the case with amps of modest power, say 30-100 Watts, where an uncontrolled and enthusiastic listening session will almost certainly result in clipping.

As John said above, reviews in the past did assess clipping behaviour and commented on anything other than a clean clip. ASR amplifier reviews should similarly show a clipping waveform into a resistive and reactive load so making it clear if the amp is likely to have problems in overload.

S

 

[Ron Texas]

I doubt that my Crown XLS 1502 ever clips. The -20 db lights blink occasionaly. It takes a big room before you can crank that amp up.

 

[sergeauckland]

It would be interesting to see how valves clip compared to solid state (hard to generalise I know).
Isn’t the conventional wisdom that valves clip more “gently”. Or is that just another audiophile myth?

I've seen valve amps that don't clip, they just get increasingly distorted. I've put this down to low feedback, but it could equally be down to the output transformer giving up before the output valves run out of volts.

Certainly, valve amps don't have the sharp clipping levels lf SS amps, so it could be thought of as soft clipping.

S

 

[solderdude]

I tested some opamps used in the typical C'Moy config.
Most clip differently also under loads they aren't really suited for (like low impedance headphones)

 

[restorer-john]

Not sure if I've already posted this. A little Rotel RA-930ax wound up to clipping and you can see the power supply rail modulation. This is 0.05%-0.06% (-66dB) THD.

(channels offset for clarity)

Clipping may look benign in this example, but take a typical usage scenario where the amp is playing at 50% of rated power and along comes a transient that pushes it way past clipping- that's what sorts the sheep from the goats.

 

[0bs3rv3r]

It would be interesting to see how valves clip compared to solid state (hard to generalise I know).
Isn’t the conventional wisdom that valves clip more “gently”. Or is that just another audiophile myth?

I think it applies more to "older" designs using power supplies, such as unregulated, tube rectifier, minimal capacitance, which are relatively high impedance. That available voltage tends to sag under load and ameliorate the hardness of the clipping. Of course, many modern implementations use just such a power supply - you've probably read about tube rolling of the rectifier tubes and changes in sound claimed.

 

[oivavoi]

The simple solution to this is to use over-powered amps which never clip. I really don't understand why there aren't more audiophiles taking this approach. Before the onset of clipping, almost all blind tests which have been conducted show that amps are more or less indistinguishable from each other. If there are differences, they are very subtle. But when amps clip, all bets are off: they start sounding different from each other, and do bad things to tweeters.

So why not just use amps which are very powerful?
(ah yes, maybe I know why, because then one wouldn't be able to use esoteric "puristic" technology, or put vanishingly low levels of THD on the spec sheet)

 

[0bs3rv3r]

The simple solution to this is to use over-powered amps which never clip. I really don't understand why there aren't more audiophiles taking this approach. Before the onset of clipping, almost all blind tests which have been conducted show that amps are more or less indistinguishable from each other. If there are differences, they are very subtle. But when amps clip, all bets are off: they start sounding different from each other, and do bad things to tweeters.

So why not just use amps which are very powerful?
(ah yes, maybe I know why, because then one wouldn't be able to use esoteric "puristic" technology, or put vanishingly low levels of THD on the spec sheet)

Well, it would exclude my 2 watt SET amp, the best sounding amp I have.

 

[restorer-john]

So why not just use amps which are very powerful?

That's what most audiophiles do.

They employ as much unclipped power as they can, but also have to consider the other issues that come with that approach. The more potential power you have, the larger the power supply, it associated radiated field and absolute noise floor (hiss and hum etc) often suffers. With greater power comes considerably more paralleled output devices. More devices, more noise.

I've never heard a big arc-welder power amplifier that was completely silent electrically or physically.

So, the sweet spot tends to be 100-200 watts in my opinion, with the quietest amplifiers typically around 120W@8R. Otherwise, what you gain (pun intended) with more power at the top end (of power scale), you can lose in low level noise in the bottom end (of the power scale).

That's why we had a rush of amplifiers in the 80s with high short term capability without the compromises of (noisy) massive power supplies. They were quiet and powerful (for music).

Class D is giving us those benefits again, a high short term power output capability, but unfortunately with added noise outside the audible bandwidth and (sometimes) issues at high frequencies. On the whole, I think quality amplification has been solved for many decades, but that doesn't mean we cannot or should not subject all amplifiers to the same types of testing at the limits.

 

[March Audio]

Purifi do have some clipping plots imn their datasheet. I will try and grab some shots of this at more subtle levels and of Hypex modules.

 

[restorer-john]

Purifi do have some clipping plots imn their datasheet. I will try and grab some shots of this at more subtle levels and of Hypex modules.

Those plots are a long way into clipping, but they do show the amplifiers' behaviour when grossly over-driven. Very stable at 1KHz into hard clipping.

Whether that is a steady 1KHz tone or a 1KHz toneburst, I don't know. Tonebursts are good as they don't stress the amplifier too much and you can uncover some real sleepers in the dynamic department.

 

[trl]

Most manufacturers try to create clipping protections or soft-clipping inside their amps because clippings will induce DC into the speakers while putting lot of stress in the output stage and the power supply too. I would never test an amp when clipping, instead I'll try to put some markings on the volume knob when clipping starts occurring with 2V RMS at input and never try to pass volume knob below that mark.

Instead of testing amplifiers while they clip I suggest purchasing amps that have two times the output power of the speakers (audioholics guys and others suggest the same). This way we ensure that speakers will never see DC due to clipping and the amp will always provide a low THD+N. Of course, how everyone protects its speakers for overloading is another story.

McIntosh MC240 with old tubes clipping around at 25 watts per channel.
(Source: https://www.kenrockwell.com/audio/why-tubes-sound-better.htm)​

"When a solid state amplifier clips, there is a sharp edge where it looks like someone simply took a pair of scissors to the tips of the waveforms. The sharp edges of solid state amplifier's waveforms at clipping give rise to insane levels of very high order ultrasonic harmonics, which are what blow out tweeters."

Figure 1 - Comparison of Transistor (Red) and Valve (Green) Clipping
​

"The 'soft' characteristic (Green) shown in Figure 1 has few high order harmonics. The harmonic content is predominantly third harmonic, with a smaller amount of fifth, and lesser amounts of each additional higher odd-order harmonic. Because the waveform is symmetrical, even order harmonics are typically at vanishingly small amplitudes. Figure 2 shows the harmonic structure of each waveform. Note that 'hard' clipping produces high levels of eleventh, fifteenth and nineteenth harmonics compared to the soft clip circuit. However, both signals will sound objectionable with full range music and with the amount of clipping shown."

​

Worth checking https://www.diyaudio.com/forums/ins...-powered-princeton-reverb-18.html#post5455791 as well.

L.E.:

http://www.gmarts.org/index.php?go=217 - Hard clipping vs. soft clipping vs. no-clipping​

 

[trl]

When a tube starts distorting it will create a rounded-corners shape instead of sharp-edges clipping on both positive and negative sides, so a tube distortion will always increase the compression of sound (hence lowering the dynamics) making it sound louder to our ears, while instead it actually does distorts, but without the listeners being able to feel the distortions (due to the soft-clipping induced by those rounded-shaped waves instead of harsh ones).

Figure 1 - Waveforms as Clipping Point is Approached and Exceeded
(Source: https://sound-au.com/valves/clipping.html)​

Of course, tubes make the sound more pleasant while distorting than transistors, but below their distortions level the harmonic profile of transistored amps is much more close to reality that the harmonic profile of the tubes (well, at least when speaking about output stage).

Figure 2 - Spectrum of Transistor (Red) and Valve (Green) Clipping Distortion
(Source: https://sound-au.com/articles/soft-clip.htm)
​