• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Amplifier Power: How much do you really need and your system’s dynamic range.

Can i ask what amps do you have Ammir?
I have the Purifi reference as mentioned. Besides that, I have the Mark Levinson No 532 and dual No 53 monoblocks.
 
Am I misunderstanding... I thought room correction software drops the relative output to ensure that any corrective peaks don't clip. I would've thought that before you got clipping, you'd have a complaint about not getting loud enough...

Dirac often boosts the bass with varying success.

Hi

Would a contributing factor also be the thickness of the speaker wire? To thin and it would not pass sufficient power not sure if to thick would be a problem, except what would to thick look like? Not sure about the thickness of the cables internal to the amp as they are generally very short in comparison.

Just my thoughts

Regards

William

the thickness of the cable only matters when the cable distances are long. other than that if they can handle the current flow it's totally fine.
 
Cable gauge has little impact in the real world.
Look at a 25' run of 18 AWG (50' total L). We can ignore the reactance of L and C for this discussion since they don't absorb real power.

Total cable R ~ 0.4 Ohm for 50'
If the speaker is 8 Ohm, cable loss ~ 5%
High but all you need to do is turn the volume up a bit.
If an amp is putting out 10 V, pretty loud, ~ 12 W.
Loss is 0.57 W, ~ -0.2 dB W

Changes volume ~ 1 dB SPL, inaudible to most, but it's relative change you notice, not if you listen one day at 69, then next day 70 after putting shorter fatter wires in and setting volume the same.
Moot
 
I have the Purifi reference as mentioned. Besides that, I have the Mark Levinson No 532 and dual No 53 monoblocks.
nice, i didnt even knew that brand. is it better than benchmark?
 
thats all i needed to hear.
 
Gotta get your hands on the insanely expensive amps to see how good they are. Stereophile measured some $100,000 Boulder amps and I recall them measuring pretty well.
yeah right
 
Indeed. Dirac aggressively attempts to fill in the dips and causes the amp to run out of juice.
Filling in dips is a foolish journey to nowhere IMHO if the dips are due to nodes in a resonance, which they are quite likely to be.
One of the reasons some of the room correction software looks lacking in common sense to me, though it is quite usual for humans to use static thinking for dynamic systems.
 
The recording is the driver of power levels, look to the waveform for the initial clues.

The waveform represents the request for voltage from the amplifier by the recording to be applied to the speaker terminals.

As for "bass transients", uh, there aren't any. Bass is slow slow slow.

In the organ piece above, here is the highest voltage request. Ooh... looks like an abrupt change, and hard to reproduce... The clip is about 2 milliseconds in length.

View attachment 126309

Let's call that peak 50V. Loud. 625Watts into 4 ohms load loud.

Now zoom in on that "transient". The dots are sample values, and are 0.000023 seconds apart - 23 microseconds.

View attachment 126310

The slew rate - rate at which an amplifier can change voltage - "Most well designed amplifiers are flat out to 100kHz so in that case you would need a minimum slew rate of about 32V/usec in order to achieve such a wide bandwidth". - https://www.audioholics.com/audio-amplifier/amplifier-slew-rate

So, for this example, assuming 50V peak of the music waveform, might need to move 2 or 3 volts in 23 uS, well below its capability.

There are faster transients, none should really tax a decent amplifier's ability to follow the waveform, if the amplifier's bandwidth exceeds the bandwidth of the musical signal, and it can flow current as demanded by the load at the requested voltage.

Thanks a million. It all makes sense to me now. Even though I knew about SR before, it was not intuitive to me that following the voltage would be so easy. So it's more about providing enough current to sustain the voltage then.

Funny though, I was even able to find an article about my trusty old Yamaha DSP-a1000 in Audio Magazine from '91, written by Leonard Feldman (was AES member) and containing his AP measurements of it back then. A very interesting review to compare with modern day ones. For people who may care I thought I might put the link to it:

https://docplayer.net/90110308-Proj...er-happened-sony-ta-e1000-digital-preamp.html (the review starts at page 54)
 
I had 3 x 18W8545K00 in parallel for mid-woofer duty. When going from this:
http://groundsound.com/pa3cc.php
to this:
http://groundsound.com/hpa2k.php

There was a clear tightening in the bass. Not that the PA3CC did a bad job.... but the added power, made it sound deeper, more controlled and with some added "smack". I already use the HPA2K for subwoofers, so I live perfectly fine with the "smaller" amp for mid-woofers. But when trying to let the 18W's play all the way down with no subwoofer-support. The bigger amp, was simply a bit better.
 
I took a slightly different tack to computing the power requirements for my listening preferences. The concept is to determine the maximum volume settings on the processor that will never clip with a 0 DBFS signal. Once determined, if you are listening below this level, content does not matter. If listening above this volume level, then there may be clipping. This is important because, as many have mentioned here, the power levels can be surprisingly low.

I have had many amplifiers, that ranged from 200 to 300 WPC per channel. Prior to the AHB2s, I had ATI AT6000 amps that are conservatively 300 WPC into 8 ohms and 450 WPC into 4 Ohms. Even with these amps, I can clip (clip indicators illuminating) driving the Salon2s when testing the limits. However, I suspect I could clip 1 kW amp as well. But that does not matter, since this is not my use case as I am well into the years where preserving my hearing is valued :)

After a friend brought over an AHB2, I was impressed by the sound of the this amplifier so bought 1 and used it to drive the mains.
Watching Aquaman with the fronts set to large, there was a lot of bass and the 3 woofers were clearly moving.
The levels were very loud, louder than I liked (and certainly louder than my family likes), the AHB2s were not clipping.
This seemed odd, since into 8 Ohms, this amp was 1/3rd the power of the ATI and into 4 Ohms, about 1/2 the power (180 v 450).

Switching over to digitally recorded music, I found the same, I can clip this amp but not when at my preferred listening levels.
So rather than computing on dynamic range of the source, it seemed reasonable to determine the maximum power requirements a a digital recording. Popular music, as discussed in the video, are often at 0 DB, maximum volume so this may be representative.

To begin, power usage in my room must be determined, in room, at the listening position for 2 speakers (the minimum really).
The baseline power computed by measuring 0 DBFS sinewaves with both speakers playing at my 10 foot listening position at 2.83 volts.
220 Hz, 440 Hz, 1K kHz, and 2 kHz tones were measured and all produced consistent SPL using REW and a calibrated UMIK-1.
The Salon2 are basically 4 Ohm speakers, so 2.83 volts this requires to 2 watts.
The Salon2s are about 86 dB efficient and surprisingly, measured 86 dB playing at my listening position. Room renforcement, 2 speakers making up for the distance loss. I suspect that the average user will find similar results.
The processor volume at -31 dB produced 86 dB which requires 2 watts.
With these measurements, all that was left is to populate as spreadsheet to show processor volume, dB at the listening position (for stereo music), and the maximum power required for digital stereo music.

Measured SPL at Listening Position.jpg


So irrespective of the digital source, the system volume at -13 will not clip but at -10 it may. I tested this with known tracks and the amplifier did clip show occasional clipping at -10. This was a surprise to me but, logically, it makes sense that moving from 200 WPC to 300 is 1.5 dB, not earth shattering.

This is a bit more than most are willing to do, so I went back to this online power calculator:
Salon2BenchmarkSPLEstimate8Ohms.jpg


When "Near a wall" is selected, DB SPL drops to 102, and "Away from walls drops to 99.
So this one parameter changes the power requires up to 6 dB or 2 to 4 times the power.

It is possible to establish the maximum volume setting in room at the listening position below which the amplifier will never clip. Once known, content and dynamic range factor in. Content can now be factored in. If the content, for example dynamic classical music, requires more than -10, there is a possibility the amp will clip. Fortunately, the AHB2 have excellent clip indicators and protection, so these are perfect for me. I do not clip these amps. Since producing my typical listening chart, Streaming services have become my primary source for movies. And I do turn the volume to -10 because the average level of the source is all over the place. Still, the AHB2 are not clipping, so I am good. If that changes, there is always the option to bridge to provide 500 WPC, which is on par with the ATI6000s.

This is a different way to look that the capabilities of the system and map them to the listeners listening preferences that anyone with an SPL and voltmeter can do at home that removes some of the guesswork made by SPL calculators.

- Rich
 
Last edited:
Hi
I don’t pretend to understand all of this but how does the 4-8watts square with a speaker that the manufacturer recommends that the amplifier should be 60watts ? Or is it just a matter of dynamic range?

Thanks

Regards

William
 
Well, i follow with open mouth.

I have used a SS amp from Cyrus, rated with 60 Watt, for a long long time. This time i have changed to a tube amp, rated 8 Watt.
My speakers are Klipsch La Scala or the small R 51.

Speakers with high efficency makes life easier.

In both cases there is no clipping audiable, and both configureations are loud enough.

Yes, a tube amp is different in clipping, this is mostly not so audiable.

Do you think i hear the wrong music? What is reason for this hurry?

Best regards Winnie.
 
I finally decided to make the effort to see how speaker voltage input relates to listening levels at the listening position.

Speakers are Duntech Black Knights (PCL400) - these are a five driver, three-way design from 1990 and come in at 154 pounds each, 90 dB @ 2.83V sensitivity, nominally 4 Ohms (minimum 2.8 Ohms.), over six feet in height, representing an earlier John Dunlavy design while he was in Australia.

Amps are from 1999 - Musical Fidelity X-A200 monoblocs - 200W/ch into 8 ohms, 400W/ch into 4 Ohms.

Room, about 45 square meters (484 sq ft), listening position about 3.5 meters from speakers.

Now, I've always suspected that the amplifier power was overkill, but I didn't know that in 1999, so I dragged out my DSO (digital storage oscilloscope) and actually measured the output of one of my amps at the speaker terminals for a variety of music and levels.

I was looking for voltages at musical peaks.

For ordinary music listening (e.g. Jazz) at comfortable levels, maybe half to one watt (average) was more than enough, with peaks of a few watts.

When choosing music with a significant bass component and turning up the volume to the limits of what I could stand for short periods, the music was peaking infrequently at 20V (~100W). This corresponds to approximately 104 dB at the listening position.

The above is roughly in line with the sorts of measurements that @RichB was getting if one adjusts for the differences in the speaker sensitivities.

For me, that leaves quite a bit of superfluous amp headroom, meaning that I could have survived quite well with an amp rated at just 100W @4 ohms, all other things being equal. However, it's good to know that amp power will never likely be the limitation of the system .

Perhaps an AHB2 would be a good option then should my MFs eventually die .
 
I finally decided to make the effort to see how speaker voltage input relates to listening levels at the listening position.

Speakers are Duntech Black Knights (PCL400) - these are a five driver, three-way design from 1990 and come in at 154 pounds each, 90 dB @ 2.83V sensitivity, nominally 4 Ohms (minimum 2.8 Ohms.), over six feet in height, representing an earlier John Dunlavy design while he was in Australia.

Amps are from 1999 - Musical Fidelity X-A200 monoblocs - 200W/ch into 8 ohms, 400W/ch into 4 Ohms.

Room, about 45 square meters (484 sq ft), listening position about 3.5 meters from speakers.

Now, I've always suspected that the amplifier power was overkill, but I didn't know that in 1999, so I dragged out my DSO (digital storage oscilloscope) and actually measured the output of one of my amps at the speaker terminals for a variety of music and levels.

I was looking for voltages at musical peaks.

For ordinary music listening (e.g. Jazz) at comfortable levels, maybe half to one watt (average) was more than enough, with peaks of a few watts.

When choosing music with a significant bass component and turning up the volume to the limits of what I could stand for short periods, the music was peaking infrequently at 20V (~100W). This corresponds to approximately 104 dB at the listening position.

The above is roughly in line with the sorts of measurements that @RichB was getting if one adjusts for the differences in the speaker sensitivities.

For me, that leaves quite a bit of superfluous amp headroom, meaning that I could have survived quite well with an amp rated at just 100W @4 ohms, all other things being equal. However, it's good to know that amp power will never likely be the limitation of the system .

Perhaps an AHB2 would be a good option then should my MFs eventually die .
As you found out, there is a reason why most amplifiers in receivers and other equipment is around 75 to 100 watts per channel. That amount of power easily allows for 90% or more of peoples volume level. In special cases where someone actually needs extra power, they can buy more but mostly 100 watts will do the trick. But since it is a hobby, people need to find (which means making up a reason) to buy a more expensive more powerful amp. So there is need and want. Most of the stuff any hobbyist is looking at is pure want and then silly reasons to justify the want. Most all of it is non-essential to very good sound reproduction with todays products. Excluding all snake oil products of course......
 
As you found out, there is a reason why most amplifiers in receivers and other equipment is around 75 to 100 watts per channel. That amount of power easily allows for 90% or more of peoples volume level. In special cases where someone actually needs extra power, they can buy more but mostly 100 watts will do the trick. But since it is a hobby, people need to find (which means making up a reason) to buy a more expensive more powerful amp. So there is need and want. Most of the stuff any hobbyist is looking at is pure want and then silly reasons to justify the want. Most all of it is non-essential to very good sound reproduction with todays products. Excluding all snake oil products of course......
Indeed. In the distant past I have lived with my speakers being run with amps as low as 70W (4 Ohms). I do wonder however whether my existing amps have lasted so long - at least, in part - because most of the time they're not being stressed and therefore never become overly warm.
 
...if nothing else I have learned here that my amplifier is not overpowered although the power rating of my speakers might indicate that to be the case...

...I find it exquisitely ironic that well recorded classical music would require more amplification than heavy metal music...
 
During my chat with the The Daily HiFi guys this past Monday, the topic of higher sensitivity speakers having more dynamic range came up. I gave my condensed spiel about why I think it isn't the speakers but rather the system as a whole that when integrating higher sensitivity speakers that provide more dynamics.

In this video I address that a little bit more but with a focus on how to obtain a realistic number for how much amplifier power you need using three factors:
1) Speaker sensitivity
2) Distance from MLP to speaker
3) Dynamic range of the content you listen to

The third is key but never discussed when the topic of "how much power do I need" comes up on the forums. So, I spend some time on that aspect. If you're wanting to kill about 20 minutes listening to audio chat, I have provided the link below.




--------------------------------

To address a comment I already had from @MZKM:
* doubling of speakers = +3dB
* doubling of power = +3dB
* net gain of adding a speaker & its own amp channel = +6dB (theoretical)
* speaker in a room vs anechoic = +3-6dB (+3dB is the norm above the transition frequency)

I tested the +6dB doubling of speakers/power by simply using a single powered speaker and measured the FR via RTA. Then I connected the other powered speaker and measured the response. I got about +5dB from about 300Hz+. In the HF+ I noticed less overall gain, which I believe aligns with what Toole discusses in his book regarding the point where a speaker becomes more directional. In the LF the gain was anywhere from +3 to +6dB.


--------------------------------

I created a poll on AVSforum a year or two ago asking what the listening distance is for people from the MLP to the speakers. The poll showed 3-4m average listening distance.

I created a poll on my YT channel and my FB group asking who pays attention to the mfg spec for "recommended power". Over 80% said they pay it no mind.


--------------------------------


Links used:
https://dr.loudness-war.info/
https://www.crownaudio.com/how-much-amplifier-power
https://www.crownaudio.com/en-US/tools/calculators#amp_power_required
Nice video!
 
Back
Top Bottom