• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. 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.

OP
hardisj

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
And efficiency of speaker is related also to the impedance of speaker. And both amplitude and phase of impedance is varied with frequency. So, you need to calculate the needed power when your speaker approaches the lowest level of impedance, not the vague "nominal impedance"

I talk about this in the video as well.
 

Ingenieur

Addicted to Fun and Learning
Joined
Apr 23, 2021
Messages
938
Likes
747
Location
PA
And efficiency of speaker is related also to the impedance of speaker. And both amplitude and phase of impedance is varied with frequency. So, you need to calculate the needed power when your speaker approaches the lowest level of impedance, not the vague "nominal impedance"

I think that is why most standards use a frequency, 1 kHz iirc. Those numbers are only order of magnitude imo.

I think they should rate them at the highest cos (phase) x |Z|, below 100 Hz, ie highest real power demand to drive to say 85 dBSPL.

Mine looks like 8.5 Ohm at ~ 32 Hz
Upper range 13 Ohm at 2500 Hz
 

Attachments

  • 37EC1C90-0C1C-48F1-AB7C-16790968952E.jpeg
    37EC1C90-0C1C-48F1-AB7C-16790968952E.jpeg
    602.7 KB · Views: 91

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,389
Location
Seattle Area
Unfortunately this type of analysis while simple to do, is not correct with respect to the amplification requirement. Or true dynamic range needed to avoid distortion.

The main issue is measurements of dynamic range. The meter used in these databases which gets its data from Foobar plug in, is based on a plug-in that PMF designed years ago (now defunct). You can read about it in the internet archives: http://web.archive.org/web/20130317013741/http://pleasurizemusic.com/en/our-aim

Here is the key aspect:

1619297962498.png


In plain English the meter is designed to find very short-term dynamic compression in music with a real-time meter. It does not at all as it indicates, give you the *max* dynamic range for an entire song. As a trivial example, if I have 30 seconds of silence and then 30 seconds of sine wave at full value, then I have as massive of a dynamic range as the format allows. In 16 bit format, that would be 96 dB! A short-term peak to average meter would show 3 dB which is the peak to average of the sine wave.

A much more accurate analysis is what @solderdude has done with grabbing different segments of a track and plotting the amplitudes in a storage scope. Then you know the real dynamic range for that segment of music. You would then have to repeat that for the whole track and of course for your entire library for that matter!

One catch is any digital silence. Such like above example, huge exaggerates dynamic range so you need statistical analysis to find and remove these.

I cover most of this in my recent video on dynamic range:


So while I admire the effort Erin put in to outline the whole process of a quick, back of the envelop computation, it unfortunately does not produce the proper answers for this application.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,389
Location
Seattle Area
Makes sense
So 70 dB SPL 0.02 W or so
It does? That is the amount of power your cell phone produces on its headphone jack. Give it a try and see if you can get satisfactory performance. Pay special attention to bass as at 70 dB, you are at or below threshold of detection at 20 Hz!
 

Harmonie

Major Contributor
Forum Donor
Joined
May 30, 2020
Messages
1,927
Likes
2,084
Location
France
The Harbeth video demonstrates this very well between the bass heavy drum song and the string quartet, completely different power requirements for the same speaker in the same room at the same volume.



Nice video.
Why didn't they make the demo with the horn speakers in that other room ?
The amp's power requirement is X times different between a 103dB and a 87dB speaker ....
 

ShadowFiend

Member
Joined
Dec 20, 2020
Messages
69
Likes
89
I think that is why most standards use a frequency, 1 kHz iirc. Those numbers are only order of magnitude imo.

I think they should rate them at the highest cos (phase) x |Z|, below 100 Hz, ie highest real power demand to drive to say 85 dBSPL.

Mine looks like 8.5 Ohm at ~ 32 Hz
Upper range 13 Ohm at 2500 Hz
I think you misunderstand. You should rate speaker at the lowest impedance, not the highest impedance cos (phase) x |Z|. Because amplifier needs to provide much more current at lower impedance than higher impedance, result in higher power needed
 
OP
hardisj

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
So while I admire the effort Erin put in to outline the whole process of a quick, back of the envelop computation, it unfortunately does not produce the proper answers for this application.

I agree. It isn’t a one size fits all, one stop solution. However, for the general consumer it is a step in the right direction of applying some thought in to what is rational when it comes to purchasing power. Much like what is bolded in the graphic you used.

Small battles. If we can get people to accept a little bit of science at a time and understand the usefulness, I consider it a win. :)
 

Ingenieur

Addicted to Fun and Learning
Joined
Apr 23, 2021
Messages
938
Likes
747
Location
PA
It does? That is the amount of power your cell phone produces on its headphone jack. Give it a try and see if you can get satisfactory performance. Pay special attention to bass as at 70 dB, you are at or below threshold of detection at 20 Hz!

That is an average dBA SPL over a minute compared to an efficiency of 87 dB 1 W/1000 Hz. Obviously the bass will be much higher and the silence//background close to 35 dBA in my home, 0 W.
 

Ingenieur

Addicted to Fun and Learning
Joined
Apr 23, 2021
Messages
938
Likes
747
Location
PA
I think you misunderstand. You should rate speaker at the lowest impedance, not the highest impedance cos (phase) x |Z|. Because amplifier needs to provide much more current at lower impedance than higher impedance.
The lowest will draw the most current so yes, P would be highest.

So perhaps a fixed power is the way to go.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,424
Likes
7,940
Location
Brussels, Belgium
Unfortunately this type of analysis while simple to do, is not correct with respect to the amplification requirement. Or true dynamic range needed to avoid distortion.

The main issue is measurements of dynamic range. The meter used in these databases which gets its data from Foobar plug in, is based on a plug-in that PMF designed years ago (now defunct). You can read about it in the internet archives: http://web.archive.org/web/20130317013741/http://pleasurizemusic.com/en/our-aim

Here is the key aspect:

View attachment 126038

In plain English the meter is designed to find very short-term dynamic compression in music with a real-time meter. It does not at all as it indicates, give you the *max* dynamic range for an entire song. As a trivial example, if I have 30 seconds of silence and then 30 seconds of sine wave at full value, then I have as massive of a dynamic range as the format allows. In 16 bit format, that would be 96 dB! A short-term peak to average meter would show 3 dB which is the peak to average of the sine wave.

A much more accurate analysis is what @solderdude has done with grabbing different segments of a track and plotting the amplitudes in a storage scope. Then you know the real dynamic range for that segment of music. You would then have to repeat that for the whole track and of course for your entire library for that matter!

One catch is any digital silence. Such like above example, huge exaggerates dynamic range so you need statistical analysis to find and remove these.

I cover most of this in my recent video on dynamic range:


So while I admire the effort Erin put in to outline the whole process of a quick, back of the envelop computation, it unfortunately does not produce the proper answers for this application.

Actually the newest loudness meter and loudness algorithims have come up with 'Loudness range'. It is mostly for statistical analysis as it excludes 'outliers' by removing the 5% loudest parts and the 10% quietist parts of a recording.

since the score is in loudness units, which is already kind of A-weighted, it's not very easy to find an 'audiophile' track that would be considered dynamic under this system. I have only been successful with orchestras which have hundreds of instruments, thus having no outliers and very smooth spectral profile.

1619299066999.png
 

Ingenieur

Addicted to Fun and Learning
Joined
Apr 23, 2021
Messages
938
Likes
747
Location
PA
Imo an important factor is how quiet the room is. 70 dBA with 90 dB peak is a different range in a room with 30 dBA background vs. one with 40, is it not?
 

Attachments

  • B75957A7-902F-462F-A049-E9CD05DBDBF0.jpeg
    B75957A7-902F-462F-A049-E9CD05DBDBF0.jpeg
    259.5 KB · Views: 69

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,389
Location
Seattle Area
OP
hardisj

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
Actually the newest loudness meter and loudness algorithims have come up with 'Loudness range'.

I read about this recently, I believe. If this is the EBU L128 (???) standard. Is that right?
 
OP
hardisj

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
Imo an important factor is how quiet the room is. 70 dBA with 90 dB peak is a different range in a room with 30 dBA background vs. one with 40, is it not?

I have seen this stated in other places and it never really made much sense to me wrt a playback system. Unless you're talking about car audio or you live in a very noisy area, you're always going to be listening above the ambient noise floor.
 

Ingenieur

Addicted to Fun and Learning
Joined
Apr 23, 2021
Messages
938
Likes
747
Location
PA
Mathematically though, dynamic range is a ratio. It is the loudest signal divided by “root mean squared” (essentially average) value of noise in the system expressed in decibels. Therefore the dynamic range is reduced if the noise floor of the system is increased.

In a noisy room, say 50 dB it takes less power to get to 70 than in a quiet 30 room.
So more power from quiet/no music to peak.

important or significant? Not really below a certain level.
 

levimax

Major Contributor
Joined
Dec 28, 2018
Messages
2,348
Likes
3,462
Location
San Diego
There are so many variables with speakers and amps and rooms and preferred listening levels and now DSP room/ speaker EQ that trying to use a "calculator" where you feed it "average values" based on possibly marketing affected manufactures specs when you are trying to figure out "peak values" is not going to be reliably accurate. While a little extreme the two methods described in this article https://sound-au.com/project191.htm either directly measuring the peak voltage and current while playing music or building a precision digital DSP "watt meter" seems like 2 of the only reliable ways to answer the question of how much power you really needs for your unique system / room /DSP EQ / listening levels.
 

eddantes

Addicted to Fun and Learning
Joined
May 15, 2020
Messages
706
Likes
1,382
Ummm ... I know this is ASR and "pedantic" is kinda in the motto of the place - but the question was "How much power does one need?" And I believe that the 9/10ths answer was adequately provided... Was it not? If not, can we get another explanation/formula for us plebes to use.
 
Top Bottom