So why are clipping indicators not standard on amplifiers in 2026?

[john61ct]

Limiting isn't necessarily better than clipping. As you turn-up the volume the peaks are limited (or clipped) but the overall average power still goes-up and it's the short-term average which overheats voice coils.

Limiting might be worse since the distortion is harder to hear.

So why are clipping indicators not standard on amplifiers in 2026?

Limiting isn't necessarily better than clipping. As you turn-up the volume the peaks are limited (or clipped) but the overall average power still goes-up and it's the short-term average which overheats voice coils. Limiting might be worse since the distortion is harder to hear. WOW good to...

www.audiosciencereview.com

...

On the topic of just using gain / level matching / attenuation, not just preventing clipping but specifically for the purpose of speaker protection

They have the same impedance matching issues and potential interaction with cable capacitance as a 'passive preamp' so need to be used accordingly.

OK then, I want to use Aphex 124b units on each channel, as a balanced <> SE level interface anyway, and it also does gain level matching as a buffer amplifier.

Looking for audio interface to use for audio measurement

To this end I need precise and repeatable gain settings in and out, accessible from the unit itself flat response down to well below 5 Hz and up to at least 40-50 kHz both for inputs and outputs. 32 bit float and all integer bit formats. preferably multiple microphone/Line inputs inputs...

www.audiosciencereview.com

If that is the final stage line level component, feeding directly into each power amp (which have no clipping indicator)

I set it to the max voltage level I want at that point for speaker protection

Then the relative-between-all channels gain adjustment is handled upstream with crossovers (maybe via DSP if delays / timing / phase alignment tuning is needed)

First Aphex unit that shows input clipping, that then provides the lowest common denominator setpoint for tuning my adjustment of the upstream Wiim Ultra (master source hub) output voltage limit setting.

All constructive feedback, from anyone, is welcome

 

[pablolie]

(Preferably with the option to turn them off for aesthetic reasons.)

So I don't get it. They can't cost much to implement, and have been standard on most pro-audio amps for many years.
So, have we been duped by the home hifi industry?
Obviously I can understand why manufacturers may not want us to know their product is struggling, but it seems like a fairly basic and important functionality/feature to me.

Or no?

Many class D designs stop at their 0dB max power without going into clipping. Hence unnecessary.

 

[MRC01]

Many class D designs stop at their 0dB max power without going into clipping. Hence unnecessary.

How does that work? Suppose you've cranked it up so a -12 dB input signal is driving the amp at -2 dB. Suddenly the signal jumps to full scale - 12 dB louder. But the amp can only do 2 dB louder. It *must* clip or soft limit the signal.

 

[Chrispy]

Many class D designs stop at their 0dB max power without going into clipping. Hence unnecessary.

Yeah, like MRC01 would like a lot more info on how this works (or at least with what specific gear you've got info on).

 

[pablolie]

How does that work? Suppose you've cranked it up so a -12 dB input signal is driving the amp at -2 dB. Suddenly the signal jumps to full scale - 12 dB louder. But the amp can only do 2 dB louder. It *must* clip or soft limit the signal.

It doesn't clip. As you deduced, it smartly limits the signal, easy enough to do in digital. You really don't have to worry about driving any of the good Class D designs into clipping. Many of them implement stuff like hard limiters, DSP-based protection, rail monitoring, current limiting and soft clipping algorithms... pretty visible in many of the amp measurements on this site.

But by all means do your due diligence... as should always be the case when matching components. If that seems complicated and daunting, just get active speakers.

 

[Chrispy]

It doesn't clip. As you deduced, it smartly limits the signal, easy enough to do in digital. You really don't have to worry about driving any of the good Class D designs into clipping.
But by all means do your due diligence... as should always be the case when matching components. If that seems complicated and daunting, just get active speakers.

Are you implying that the class Ds somehow digitally step in and limit that?

 

[pablolie]

Are you implying that the class Ds somehow digitally step in and limit that?

Yes I am. There's nothing magic about it. Many modern Class D amps are computer-controlled enough that they simply refuse to output more power once they hit their safe limit, instead of brutally flattening the waveform into distortion. It is quite visible in some of the measurements here.

 

[Chrispy]

Yes I am. There's nothing magic about it. Many modern Class D amps are computer-controlled enough that they simply refuse to output more power once they hit their safe limit, instead of brutally flattening the waveform into distortion. It is quite visible in some of the measurements here.

Not that I'm aware of. Cite some examples of class d amps that work that way please.

 

[john61ct]

If any at all exist, that is not a feature I've seen, of Class D power amps as a category.

It would be a feature explicitly implemented by the DSP as a "computer".

More commonly, there are thermal protection features that trip when things overheat to prevent permanent damage

but those are rarely based on the clipping itself, not what I would call a computer control, but very common going back way before Class D even came along

 

[MRC01]

... There's nothing magic about it. Many modern Class D amps are computer-controlled enough that they simply refuse to output more power once they hit their safe limit, instead of brutally flattening the waveform into distortion. It is quite visible in some of the measurements here.

Whether it is "clipping" depends on how quickly they sense the overload and reduce their gain. Any sufficiently quick reduction in gain is clipping or soft limiting. This adds odd order harmonics. The limiting case is a square wave, shutting down "instantly" which creates odd order harmonics going to infinity (or whatever the system bandwidth is). The simplest/softest case is 3rd harmonic only, which rounds off the peaks (soft limiting). The only way a device can avoid added distortion like that is to reduce gain more slowly, which requires looking ahead in the signal.

Is that what they do? They take a digital input, buffer it by some fraction of a second to look ahead? If so they must introduce significant latency. That might not work if it's part of a video system. One would need to delay the video too to keep them in sync.

 

[KSTR]

Many class D designs stop at their 0dB max power without going into clipping. Hence unnecessary.

Yes I am. There's nothing magic about it. Many modern Class D amps are computer-controlled enough that they simply refuse to output more power once they hit their safe limit, instead of brutally flattening the waveform into distortion. It is quite visible in some of the measurements here.

I am familiar with a great many of Class-D amplfier modules from ICEpower, Hypex, Purifi, Pascal, and some others. All of them are fully analog designs, nothing digital in there. Most of them try to clip softly, some have limiters, but that is not achieved by digital audio processing in any way.

There are exceptions like Axign which have a digital (PCM) input and a true digital modulator (and some even read back the analog output voltage with an ADC to close the feedback loop).

And there are some more integrated plate amps and such which have digital input front-ends with DSP, but often the amp itself is analog and could be replaced by a linear amp. This also applies to some highly integrated chips (for sound bars, automotive etc) which combine the DSP, DAC and class-D amp (and ADCs to read back voltage and current).
But the Class-D amp proper the majority of the designs is pure analog.

 

[EJ3]

My active speakers (K&H O300D) indicate clipping of the power amps by flashing the iluminated Logo, and I think the Neumann KH310 does the same.

My 6 NAD 2200's have them, as do my 2 PROTON D1200's (which also have meters). But, it's a design choice and I chose to buy my amps with that design.
Also that was pre-class D, for which I believe is un-needed IF they are properly designed.
BUT I am not sure on that point.
At any rate, what I have is easily repairable and I fully expect it to last beyond my lifetime just fine. And, after that, I suspect that I won't care about it anyway.

 

[pablolie]

I am familiar with a great many of Class-D amplfier modules from ICEpower, Hypex, Purifi, Pascal, and some others. All of them are fully analog designs, nothing digital in there. Most of them try to clip softly, some have limiters, but that is not achieved by digital audio processing in any way.

There are exceptions like Axign which have a digital (PCM) input and a true digital modulator (and some even read back the analog output voltage with an ADC to close the feedback loop).

And there are some more integrated plate amps and such which have digital input front-ends with DSP, but often the amp itself is analog and could be replaced by a linear amp. This also applies to some highly integrated chips (for sound bars, automotive etc) which combine the DSP, DAC and class-D amp (and ADCs to read back voltage and current).
But the Class-D amp proper the majority of the designs is pure analog.

NAD implements soft clipping. So does Crown. So do a zillion others. PWM (control) is a digital technology.

 

[pjug]

PWM (control) is a digital technology.

I am not saying this is wrong because I am sure you can find a respected engineer who would agree that Class D and PWM are digital. But I would defer to Bruno Putzeys:

https://www.researchgate.net/publication/242013327_All_amplifiers_are_analogue_but_some_amplifiers_are_more_analogue_than_others

 

[Roland68]

NAD implements soft clipping. So does Crown. So do a zillion others. PWM (control) is a digital technology.

That is not entirely accurate.
PWM is not a digital technology; this assumption usually stems from the misconception that the signal is merely being switched on and off.
In the same vein, Class D does not equate to "digital."
Whether a system is digital or analog depends entirely on the specific application.

Example: TPA3255.
The input is analog, and the device accepts only analog signals. Internally, the analog signal is converted into a pulse-width-modulated signal by a pulse-width modulator (PWM). This is an analog process—executed via operational amplifiers (OPAs)—that drives the switching transistors within the TPA3255. In this context, the PWM signal at the amplifier's output must be regarded as an analog signal, which—following low-pass filtering—can be used directly to drive a loudspeaker.
There is no digital signal, no digital signal processing, and no digital component involved in this process.

Example: Axign AX5689 Digital Amplifier IC.
The Axign AX5689 chip processes only digital signals (or converts analog signals into digital form for subsequent processing). At the output stage, the Axign chip generates a PWM signal from the digital input, which is then used to drive the transistors. If one can speak of a D/A conversion process, it takes place here.
In this case, too, the PWM signal at the amplifier's output must be regarded as an analog signal, which—following low-pass filtering—can be used directly to drive a loudspeaker.

 

[EJ3]

I know in the States it was a broader church but over here we had retailers who didn't stock any Japanese brands at all. Probably we still do.

Not even NAD?
Which were definitely available in Austria in the mid 1970's, as I that is the first place that I saw any NAD gear.
I had friends that had NAD gear there.

 

[pablolie]

I am not saying this is wrong because I am sure you can find a respected engineer who would agree that Class D and PWM are digital. But I would defer to Bruno Putzeys:

https://www.researchgate.net/publication/242013327_All_amplifiers_are_analogue_but_some_amplifiers_are_more_analogue_than_others

We are splitting hairs. Fact is Class D can be easily made to soft clip. Do we agree? And I'd argue if something has a modulator in there, it is most definitely a digital component... and afaik Ncore and Purifi (and all of them...?) use modulators these days.

 

[boXem]

Fact is Class D can be easily made to soft clip. Do we agree?

No. Fact is that all modern class D hard clip.

And I'd argue if something has a modulator in there, it is most definitely a digital component

This is most definitely wrong.

 

[Roland68]

We are splitting hairs. Fact is Class D can be easily made to soft clip. Do we agree? And I'd argue if something has a modulator in there, it is most definitely a digital component... and afaik Ncore and Purifi (and all of them...?) use modulators these days.

That, too, is incorrect.
You will find the following statement in many technical textbooks:
The modulation method can be either analog or digital, regardless of the nature of the payload signal.

Take an FM radio broadcast signal with stereo transmission—how was it generated? With a modulator.
So, was FM radio digital?
In color television, brightness and color information were additionally modulated onto the signal—was that digital, too?
Both utilized analog amplitude modulation (AM) and frequency modulation (FM).

In analog signal transmission, the PWM signal generates arbitrary (time-varying) values—not digital ones. There are no ones and zeros involved. Consequently—among other things—variations in the rise and fall times (edge steepness) of the PWM signal will distort the signal.
With purely digital signals, this issue does not arise.

You are free to view it differently, but that does not make it correct.

 

[Geert]

And I'd argue if something has a modulator in there, it is most definitely a digital component

An entirely analog circuit can generate a PWM signal with truly continuous, infinitely variable pulse widths. That’s how most class D amplifiers operate. The information is encoded in a continuous time domain (this is key), even though the output voltage hard-switches between two levels.

A PWM signal can also be generated digitally, using a clock and counters, resulting in a pulse width with discrete values. Now it’s a digital PWM signal