'Headroom' is a measure of the badness of an amplifier. The bigger the number, the worse the amplifier.

[solderdude]

Certain NAD amps from a certain time period had a special circuit that could momentarily only let the amp provide a higher output voltage quickly sagging to the normal voltage rail so continuous power was much lower than the short peak power only. This circuit had a slowly charging 'bootstrap' alike increase in the power supply voltage above the normal voltage rail of the output voltage.
Sort of a class G design but with a higher power rails that only could supply short bursts. Enough to deliver short and very high power peaks but not continuous.
This was by design and AFAIK only shortly used in some NAD amps.

These amps had other issues (thermal runaway caused by the omission of emitter resistors in the output stage. These amps, when pushed too hard were prone to blow their output devices. This was all in an attempt to increase momentary increase of output power.

 

[Vladimir Filevski]

Lesson on (lack of) logic:

First - the premise:

Headroom measurements tell use how BAD and amplifier (mostly power supply performance) is. It is not a measure of goodness.

OK, then what do you think - which amplifier will sound better:

amplifiers of identical circuit but different power supplies. They all provided 100 watts into 8 ohms continuous, but were configured as follows:-

A) Conventional Sagging Supply, providing 100 watts continuous, but 150 watts on 'short-term' peaks.
B) Stabilised supply providing 100 watts continuous, but no extra on peaks.

Answer:

Amplifier A MIGHT sound better than amplifier B.

 

[AnalogSteph]

I would expect deep sustained organ notes to have a crest factor approaching that of a sine, so if you absolutely do need to blast those while driving a tough load, there will be "no replacement for displacement" or rather sustained sinewave power à la FTC. Too bad for your wallet and your back, as large mains transformers are neither lightweight nor cheap, unless you want to open the can of worms of SMPS (which as mentioned have a habit of shutting down abruptly if overloaded and as such need to be derated accordingly).

With most kinds of music, however, crest factor is high enough that an amplifier that can deliver 50 W rms and 200 W peak is going to be more useful than one limited to 100 W rms / peak. (Driving funny loads is related but another discussion.)

I think you can only see images from Facebook when you're logged in there, this shows up blank (with just the image file name) for me.

 

[MRC01]

An amp having peak power well above its continuous power, could in some cases be nothing more than a side effect of having a well regulated power supply with lots of capacitance.

A better measure of a power amp's power supply is how it increases its power into lower impedance loads. If it is not power supply or current limited, output power will double as load impedance drops from 8 to 4, then again from 4 to 2 ohms. This can be desirable because it makes the amp flexible, able to drive any speaker.

Yet even this doesn't necessarily indicate the amplifier's sound quality or dynamics. So long as your peak power demand never exceeds the amp's continuous power into your speaker's lowest impedance (you get a big enough amp), this won't matter.

 

[mhardy6647]

...
I think you can only see images from Facebook when you're logged in there, this shows up blank (with just the image file name) for me.

 

[EJ3]

I would expect deep sustained organ notes to have a crest factor approaching that of a sine, so if you absolutely do need to blast those while driving a tough load, there will be "no replacement for displacement" or rather sustained sinewave power à la FTC. Too bad for your wallet and your back, as large mains transformers are neither lightweight nor cheap, unless you want to open the can of worms of SMPS (which as mentioned have a habit of shutting down abruptly if overloaded and as such need to be derated accordingly).

With most kinds of music, however, crest factor is high enough that an amplifier that can deliver 50 W rms and 200 W peak is going to be more useful than one limited to 100 W rms / peak. (Driving funny loads is related but another discussion.)

I think you can only see images from Facebook when you're logged in there, this shows up blank (with just the image file name) for me.

Shows up just fine for me.
And I do not even have Facebook.

 

[kemmler3D]

What happens if I turn the volume up to just peak the music on one of these amps and it runs out of 'dynamic power' to maintain that peak?

I think "maintaining a peak" is an oxymoron. Peaks are transient (short, like measured in ms short) by nature, if it's not transient, it's not a peak, it's continuous. Uncompressed music is characterized by a really low RMS value with really high peak values that typically only last tens or maybe hundreds of ms. Even compressed music has a really low RMS compared to a sine.

 

[Keith_W]

@Keith_W above seems to be suggesting you might be technically competent. Unfortunately that competence doesn't seem to extend to making your case in a reasoned written argument. Based on your posts here, any competence is (to this reader at least) masked.

I have no desire to have my reputation tied to him All I will say is - he has gotten into some debates with amplifier designers in the past which were quite entertaining to read. At least he measures, he is honest, and he can back up his statements. He is not a troll, I know him well enough to know that he genuinely believes in what he is saying (whether it is credible or not, you be the judge). And above all, I am not defending his ideas or his argument. He can do that on his own.

 

[coonmanx]

I think "maintaining a peak" is an oxymoron. Peaks are transient (short, like measured in ms short) by nature, if it's not transient, it's not a peak, it's continuous. Uncompressed music is characterized by a really low RMS value with really high peak values that typically only last tens or maybe hundreds of ms. Even compressed music has a really low RMS compared to a sine.

If you have trouble "maintaining a peak" then get some Cialis... Otherwise if your peak lasts longer than six hours... please see a doctor...

 

[Chr1]

Nowt worse than a sagging power supply...

 

[Deleted member 48726]

Nowt worse than a sagging power supply...

Well. Gravity and age does that..

 

[Chr1]

Excess alcohol too

(Just add Viagra for the youthful vigour of an over-specced linear toroid and supply caps.)

Max headroom.
(temporary/transient)

 

[kemmler3D]

If you have trouble "maintaining a peak" then get some Cialis... Otherwise if your peak lasts longer than six hours... please see a doctor...

With all the talk of "stiff" and "soft" power supplies in this thread I think maybe I'm just missing some important subtext and this thread isn't about audio at all.

 

[jooc]

@Zaphod​

All Equal Else Equal:
100 watts average power and burst power of 100 watts. > 80 watts average power and burst power of 100 watts.
100 watts average power and burst power of 100 watts. < 100 watts average power and burst power of 120 watts.

the only controversial question would be:
100 watts average power and burst power of 100 watts. Or 90 watts average power and burst power of 120 watts.
In this case it depends on your needs. your speakers and your music.

After reading this thread I'm not entirely sure everyone's working with the same exact definition of "headroom."

Aside from that I suspect the above illustrates the bulk of what OP is arguing against, even if he doesn't realize it, and he may not realize it because he's applying some older class a/b stereo selling experience from the 1980s to current class D products and also wildly misunderstanding the audible thresholds involved.

 

[Cbdb2]

When did headroom become the same as peak power. Headroom was always how far over nominal level (0VU) a continuous signal could go before clipping, as in mixing consoles with 20db of headroom. It had nothing to do with peaks. And amps (at least pro amps) never had a headroom/peak power rating till the marketers and reviewers got involved. If it was rated 100 watts that what it would deliver no matter what kind of signal you fed it.

Amplifiers – Bryston

bryston.com

Not one mention of headroom or peak power.

(And the signal in signal to noise was at 0VU while dynamic range was noise to headroom, ie if a component had 100db of S/N and 20db of headroom it had a dynamic range of 120db, going to digital with 0 dbfs confused the issue (try recording a reference tone at 0dbfs on a digital tape machine, it was usually -18,16,14 depending on how much headroom you wanted) and the marketers got involved so now the same device is said to have 120db of S/N.)

 

[MAB]

I am unclear on the source of all of these new maximum power/SPL manifestos that keep popping up on ASR lately.
No way any of this thread's assertions are audible unless you compare absolute junk to competent amps (FYI, I consider a Niles multi-zone amp competent). Maybe at peak output this matters, but then suggestions of critical listening above 100dB with gear that has no cost-constraints on ASR are odd. And any rational person who wants 100+dB should look to speaker efficiency rather than kiloton amps, since the cost is less and speakers actually make a difference.
Amps are just not the defining part of a HiFi, despite this thread's attempt to generate amp-relevance. They are boring, and the artifacts are not audible unless comparing a broken or really crappy amp to one that works properly, or an amp with really high outpuit impedance in appropriately paired with speakers with low impedance. So many things matter, maybe in a distant future where speakers approach perfection this matters.

Also, comparing a Fosi to an A/B with 5kVA and 0.31F of capacitance is really odd. Do you think you can discern the two in a blind test? Reasonable volume? Even clipping, can you hear the difference?

Otherwise, this is a standard discussion of power supply cost and complaining about standard tradeoffs, and would be better if attended by a discussion of audibility, which you really inelegantly stumbled through in these two posts:

an amplifier with a higher headroom figure will always sound worse than one with a lower headroom figure.

Amplifier A MIGHT sound better than amplifier B. That would depend on a huge number of factors. However, that was not what I was referring to. Amplifier C is the preferred one.

Also, good practice that when you edit a post, you indicate what was edited, especially if it is more than a simple typo.

Welcome to ASR. I look forward to your follow-on measurements to demonstrate your point since appeals to reason and appeals to experience are not useful.

edit: added a word for clarity

 

[antcollinet]

I have no desire to have my reputation tied to him

I thought I'd used enough weasel words (suggested, might) to leave you out of jeopardy.

 

[Zaphod]

You still have not defined "better"
More energy efficient. more cost efficient, more space efficient, more weight efficient...

Whats the point if you can't even define how its better

An amplifier with a stiff power supply, when compared with an otherwise identical amplifier with a soft power supply will always measure better and possibly sound better.

 

[jooc]

What definition of 'headroom' is Crown Audio using in their calculator below, and is it the same thing the OP is talking about? Is 'peak power' and 'headroom' being treated as the same thing consistently?

 

[voodooless]

Welcome to ASR

He’s longer a member than you