Amplifier power output - RMS vs Peak as it applies to home theater soundtracks - AVS Forum
Forum Jump: 
Reply
 
Thread Tools
post #1 of 42 Old 07-12-2013, 05:44 PM - Thread Starter
Advanced Member
 
Nyal Mellor's Avatar
 
Join Date: Aug 2010
Location: SF Bay Area, California, USA
Posts: 956
Mentioned: 5 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 75
Is there any difference in the RMS vs Peak power output when an amplifier is reproducing a home theater soundtrack?

Please consider I am thinking of amplifier behavior with program content, not amplifier behavior with a sine wave input. Also output should be low THD and not clipped.

The question originates from a consideration of amplifier power required to produce specification 105dB peaks at the listening position when a system is fed 0dB FS.

I believe THX do some program type measurements in their certification process but I do not know how the output figures with burst tones or the like compare with output figures when fed a sine wave.


To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.
. We specialize in the design and creation of high performance listening rooms, home theaters and project studios for discerning audio/video enthusiasts.
Nyal Mellor is online now  
Sponsored Links
Advertisement
 
post #2 of 42 Old 07-12-2013, 07:50 PM
AVS Special Member
 
lovinthehd's Avatar
 
Join Date: Dec 2007
Location: OROR
Posts: 6,664
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 227 Post(s)
Liked: 794
I was just looking at your other post on the subject. Subscribed now as interested....

ps Would hope that we're all working with rms as a constant.

lovinthehd is online now  
post #3 of 42 Old 07-13-2013, 05:51 AM
AVS Special Member
 
commsysman's Avatar
 
Join Date: Dec 2007
Location: Southern California
Posts: 5,285
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 131 Post(s)
Liked: 256
Power is normally measured as RMS, or continuous power.

Peak power is momentary, and is 1.414 times RMS power.

Some amplifiers, notably NAD, claim transient peak "Music" power capability greater than that because of power supply reserves that can be momentarily utilized (but not sustained).
commsysman is offline  
post #4 of 42 Old 07-13-2013, 06:32 AM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,381
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 748 Post(s)
Liked: 1162
Quote:
Originally Posted by Nyal Mellor View Post


Is there any difference in the RMS vs Peak power output when an amplifier is reproducing a home theater soundtrack?

Please consider I am thinking of amplifier behavior with program content, not amplifier behavior with a sine wave input. Also output should be low THD and not clipped.

The question originates from a consideration of amplifier power required to produce specification 105dB peaks at the listening position when a system is fed 0dB FS.

I believe THX do some program type measurements in their certification process but I do not know how the output figures with burst tones or the like compare with output figures when fed a sine wave.

Music is composed of a constantly changing collection of multiple tones.

Test bench power is based on a steady pure single tone.

In terms of demand and stress on amplifier components like power transformers and heat sinks, thesingle pure tone is a far bigger load and produces far more stress.

The most stressful music is music that has highly compressed dynamics. Musical instruments and voices just aren't much like pure tones. After examining a ton of recordings the worst case music I found contains about 8 dB less energy than a pure tone with the same peak level. If you do the math, this means that you should still be able to run 12 channels of music using the same power supply as it would take to run 2 channels of test tones. Of if you say my analysis is off and the difference is only 6 dB, then you can still run 8 channels of music using the same power supply as it would take to run 2 channels of test tones
arnyk is offline  
post #5 of 42 Old 07-13-2013, 09:38 AM - Thread Starter
Advanced Member
 
Nyal Mellor's Avatar
 
Join Date: Aug 2010
Location: SF Bay Area, California, USA
Posts: 956
Mentioned: 5 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 75
With reference to home theater soundtracks is it correct to say we have a program with a 20dB crest factor?

The faceplate power of amplifiers is tested and quantified using a sine wave which has a 3dB crest factor.

Regardless of test signal is it true that the peak output power of the amp, in Watts, is the same in both cases? i.e. that the amp cannot produce a higher peak output for music than it can for a sine wave without clipping.

The RMS power required will obviously be different between the two.


To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.
. We specialize in the design and creation of high performance listening rooms, home theaters and project studios for discerning audio/video enthusiasts.
Nyal Mellor is online now  
post #6 of 42 Old 07-13-2013, 09:14 PM
AVS Special Member
 
A9X-308's Avatar
 
Join Date: Mar 2008
Location: Australia; now run by adults.
Posts: 5,327
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 113 Post(s)
Liked: 84
Quote:
Originally Posted by Nyal Mellor View Post

With reference to home theater soundtracks is it correct to say we have a program with a 20dB crest factor?
20dB is not uncommon and may be higher in some cases. Play some tracks through Audacity or similar and you'll see for yourself. Likewise in the master list of bass tracks thread on this forum.
Quote:
Originally Posted by Nyal Mellor View Post

The faceplate power of amplifiers is tested and quantified using a sine wave which has a 3dB crest factor.

Regardless of test signal is it true that the peak output power of the amp, in Watts, is the same in both cases? i.e. that the amp cannot produce a higher peak output for music than it can for a sine wave without clipping.
The maximum clean power able to be delivered to a given load is determined by the supply rails of the amplifier until distortion reaches some maximum acceptable value.

For example, an amplifier specced as 100W/8R will need to be able to swing 28.28Vrms, or 40Vpk (80Vpp) at low distortion. Higher than this value and the waveform's shape will be changed by the hard upper and lower limits of the 40V voltage rails* and distortion will rise. The amp is now clipping. It does not matter what the shape of the waveform at any given instant, when the voltage of the waveform exceeds 40V (in this amp) it has reached maximum clean power.

SS amplifiers are a voltage source. They supply a voltage at the output, and the load determines the power delivered. I could put a sinewave into my hypothetical 100W/8R amp and have 28.28Vrms of undistorted voltage at the speaker terminals with no load attached. If I then switched in an 8R resistor as load, 3.535Arms (28.28/8) would flow through the resistor. If the load were 4R instead of 8, then 7.07Arms current would flow because there would still be 28.28Vrms across the speaker terminals. The 8R resistor would be absorbing 100W and the 4R, 200W. Note though that few real world amplifiers double down, delivering twice the rated power into half the impedance.
Quote:
Originally Posted by Nyal Mellor View Post

The RMS power required will obviously be different between the two.
There is no such thing as RMS power, however this is a convention used by many if only to distinguish between a real spec and BS numbers like PMPO.

* For simplicity's sake, I have ignored Vce/Vds and emitter/drain resistor losses. A real value for a rail voltage for this amp would be a few volts higher.
A9X-308 is offline  
post #7 of 42 Old 07-13-2013, 10:34 PM
AVS Special Member
 
DonH50's Avatar
 
Join Date: Feb 2010
Location: Monument CO
Posts: 6,120
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 171 Post(s)
Liked: 263
"Crest factor" has been redefined a couple of ways (leaving out the RF systems definitions):

1. The ratio of peak to average power in a musical (or movie) signal; and,
2. The ratio of peak to RMS voltage or current.

The former is the way I have seen it applied to music studies. As has been said, there is no such thing as "RMS power", only RMS voltage or current (or few few other things, but root-mean-square does not apply to/is not defined for derived parameters like power). The product of RMS voltage and RMS current is the average power. Peak power is usually the product of peak current times peak voltage, but for audio systems is the maximum output power for short peaks (before the rails sag). I am pretty sure the old IHF standard specified the duration but, despite having passed the certification a couple or three decades ago, I cannot recall the spec (way too long ago).

The peak to average value in music was assessed by the AES decades ago at ~17 dB (a power factor of 50:1). For movies, I am unaware of any specific studies but I am sure they are out there. I have read many times here and elsewhere that 20 - 30 dB (power ratios of 100x - 1000x) is not uncommon. Given the high bass content for explosions and such, coupled with Fletcher-Munson (or whatever) curves showing greatly reduced sensitivity to low frequencies, that does not surprise me.

I used to have the label ripped from a cheap SoundDesign all-in-one system that specified 2 W RMS and 100 W (or was it 1000 W?) "peak music power". We (myself and the other techs at the shop where I worked) decided that peak rating had to be when the power transistors shorted and dumped the rail to the load for a few ms before the thing blew up...

"After silence, that which best expresses the inexpressible, is music" - Aldous Huxley
DonH50 is online now  
post #8 of 42 Old 07-14-2013, 12:57 AM
 
goneten's Avatar
 
Join Date: Dec 2002
Location: South Africa
Posts: 3,681
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 17
Quote:
Originally Posted by DonH50 
For movies, I am unaware of any specific studies but I am sure they are out there. I have read many times here and elsewhere that 20 - 30 dB (power ratios of 100x - 1000x) is not uncommon. Given the high bass content for explosions and such, coupled with Fletcher-Munson (or whatever) curves showing greatly reduced sensitivity to low frequencies, that does not surprise me.

Assuming a movie is recorded with 105 dB peaks in the main channels when fed 0dB FS, it would seem that with a reasonably high sensitivity (90 dB) that the actual power required the majority of the time would be a handful of watts, but 100 watts, together with a subwoofer and high-passing the main speakers at some crossover frequency, like 60-80 Hz) shouldn't find it terribly difficult to hit those figures. So the idea that you need a separate power amplifier for dynamic peaks isn't really as critical as I once thought. At least the idea of needing 250-400 wpc amplifiers when sitting at reasonable distances, within 2-3 meters.
goneten is offline  
post #9 of 42 Old 07-14-2013, 02:27 AM
AVS Special Member
 
A9X-308's Avatar
 
Join Date: Mar 2008
Location: Australia; now run by adults.
Posts: 5,327
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 113 Post(s)
Liked: 84
Quote:
Originally Posted by goneten View Post

Assuming a movie is recorded with 105 dB peaks in the main channels when fed 0dB FS, it would seem that with a reasonably high sensitivity (90 dB) that the actual power required the majority of the time would be a handful of watts, but 100 watts, together with a subwoofer and high-passing the main speakers at some crossover frequency, like 60-80 Hz) shouldn't find it terribly difficult to hit those figures. So the idea that you need a separate power amplifier for dynamic peaks isn't really as critical as I once thought. At least the idea of needing 250-400 wpc amplifiers when sitting at reasonable distances, within 2-3 meters.
A 90dB/W/m* speaker and a 100W amp gets you 110dB at 1m and about 100.3dB at the listening position. [Calculator]. Boundary calculators are generally useless unless the frequency and the distances form the speaker to LP and boundaries are defined, especially as we're not talking about subs so I didn't use them and calc'd for 1 speaker.

At a similar listening distance, I have 1350W actively triamped and 99dB sensitivity. Should be enough.

*Some manufacturers like to use 2.83V (1W/8R) at 1m which for a 4R nominal speaker is 2W. Depending on the amp/speaker this could result in a couple of dB lower output.
A9X-308 is offline  
post #10 of 42 Old 07-14-2013, 04:07 AM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,381
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 748 Post(s)
Liked: 1162
Quote:
Originally Posted by Nyal Mellor View Post

With reference to home theater soundtracks is it correct to say we have a program with a 20dB crest factor?

A 20 dB crest factor would be an extreme but possible case.
Quote:
The faceplate power of amplifiers is tested and quantified using a sine wave which has a 3dB crest factor.

Agreed.
Quote:
Regardless of test signal is it true that the peak output power of the amp, in Watts, is the same in both cases? i.e. that the amp cannot produce a higher peak output for music than it can for a sine wave without clipping.

When I've tested amps on the bench with musical signals and a loudspeaker-like load, I've generally obtained far more voltage output than I've obtained in tests with sine waves and resistive loads. High-pass filtering such as is used with bass management for use with a subwoofer further increases this difference..

Observed differences have ranged from 1-3 dB.
arnyk is offline  
post #11 of 42 Old 07-14-2013, 10:06 AM
AVS Special Member
 
JHAz's Avatar
 
Join Date: Mar 2009
Posts: 3,959
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 96 Post(s)
Liked: 154
FWIW, with respect to relative power across the audio band, for purposes of "thinking out loud" I fall back on octave-to-octave power in pink noise. As I understand it, pink noise is typically used for testing because it more nearly mimics the real frequency content of real material.

Pink noise has equal power in each octave. So the octave from 40 to 80 Hz has the same power in a pink noise signal as the octave from 10,000 to 20,000 Hz. (white noise averages equal power at every frequency, so it would have equal power from 40 to 80 Hz as it has from 10,000 to 10,080 Hz). Since most music has little sub-40 Hz content, I tend to think in terms of the power in the three octaves from 40 Hz to 320 Hz being equal to the three octaves from 320 to 2560 Hz, which is equal to the three octaves from 2560 Hz to 20,480 Hz. At least roughly.

In terms of "peak power" meaning VI at the top of a sine wave, from a time perspective, the power is at the peak for the exact same amount of time as power is literally zero in a sine wave. Neither the power at the top of teh wave nor the power at the zero crossing reasonably characterizes the amp's power output, IMO. At least if I have a (real) peak number I can assess RMS.

In terms of peak power as a very brief clean power power output in excess of the amp's long term clean power, it's a real enough phenomenon, and for a while Stereophile tested it for every power amp they reviewed. It was interesting to see . . . Of course whether one's speakers can cleanly and without compression reproduce that very brief transient, even if it's present on the recording (i.e. not compressed dynamically) and cleanly reproduced by the amp, will determine whether it gets reproduced by the system . . . Certainly I expect my smallish Paradigms would struggle to reproduce such transients cleanly and without compression at highish volume. Lucky for me I don't listen all that loud generally speaking.

And you just have to wonder whose backside some marketing departments pull their obviously outrageous "peak" power numbers from. Surely they are unconstrained by any distortion limit, but even so there are just limits to what any given device can do . . .
JHAz is offline  
post #12 of 42 Old 07-14-2013, 12:22 PM
AVS Special Member
 
Ethan Winer's Avatar
 
Join Date: Apr 2003
Location: New Milford, CT, USA
Posts: 5,748
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 7 Post(s)
Liked: 133
Quote:
Originally Posted by Nyal Mellor View Post

Is there any difference in the RMS vs Peak power output when an amplifier is reproducing a home theater soundtrack?

You're getting good, technically accurate info, so I'll add just one point:

Most of the energy in music, including the music in movie soundtracks, is in the bass range. Further, the bass instrument in a lot of music is compressed to sustain more and sound more solid. Or, if not pop music, the bass may be played with a bow which is also mainly continuous. So the available steady state power as measured with sine waves is a realistic metric when applied to music and movies.

--Ethan

RealTraps - The acoustic treatment experts

To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

Ethan Winer is offline  
post #13 of 42 Old 07-14-2013, 02:17 PM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,381
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 748 Post(s)
Liked: 1162
Quote:
Originally Posted by Ethan Winer View Post


Most of the energy in music, including the music in movie soundtracks, is in the bass range.

Definition of terms. Music varies all over the map, so which music are we talking about? Bass is not an exact quantity, either. I think we know the approximate frequency range of bass, but probably won't agree within 10 Hz.

I picked the bridge from what I believe is an original recording of "Mustang Sally".

Using a brick wall filter, brick wall low pass filtering at 150 Hz removed half of its energy. Back in the days of analog I read an article that put this point at a much higher frequency. I suspect the dynamic range and bass extension of digital recording over classic analog recording has changed music. Vinyl was often cut after the music had been high-pass filtered at 50 or 80 Hz.

Getting back to "Mustang" the peak to average ratio of the resulting wave was about 12 dB. Prior to filtering (IOW the entire musical signal) had a peak to average ratio that was more like 15 dB, so the non-bass portion of the wave was a little more dynamic or impulsive which seems to make intuitive sense.

The bass line did not seem to be compressed. This is a legacy recording that was probably played by a highly experienced player with lots of control.
arnyk is offline  
post #14 of 42 Old 07-15-2013, 09:38 AM - Thread Starter
Advanced Member
 
Nyal Mellor's Avatar
 
Join Date: Aug 2010
Location: SF Bay Area, California, USA
Posts: 956
Mentioned: 5 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 75
Thanks for the replies...

So generally speaking it is true that the peak output power of the amp, in Watts, can be considered the same regardless of the input signal. There may be some small observed differences with some very brief transient signals on the order of a couple of dB (is this what is called dynamic headroom?), although this varies from amp to amp.

With respect to crest factor for movies. Again generally speaking if the average movie soundtrack level is 85dB SPL* with peaks of 105dB then by definition we have a 20dB crest factor. *I believe the 85dB SPL actually refers to the average dialog level when set up as per THX guidelines so the program level could be more or less than this and will vary from movie to movie.


To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.
. We specialize in the design and creation of high performance listening rooms, home theaters and project studios for discerning audio/video enthusiasts.
Nyal Mellor is online now  
post #15 of 42 Old 07-15-2013, 11:43 AM
AVS Special Member
 
Ethan Winer's Avatar
 
Join Date: Apr 2003
Location: New Milford, CT, USA
Posts: 5,748
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 7 Post(s)
Liked: 133
Quote:
Originally Posted by arnyk View Post

Music varies all over the map, so which music are we talking about? Bass is not an exact quantity, either.

Of course. Further, music on vinyl is typically thinner sounding than modern CDs, often much thinner. The less bass you put on an LP, the more of everything else you can fit, including total program length.
Quote:
The bass line did not seem to be compressed. This is a legacy recording that was probably played by a highly experienced player with lots of control.

Right, a good bass player doesn't need a compressor. biggrin.gif

--Ethan

RealTraps - The acoustic treatment experts

To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

Ethan Winer is offline  
post #16 of 42 Old 07-15-2013, 11:44 AM
AVS Special Member
 
Ethan Winer's Avatar
 
Join Date: Apr 2003
Location: New Milford, CT, USA
Posts: 5,748
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 7 Post(s)
Liked: 133
Quote:
Originally Posted by Nyal Mellor View Post

So generally speaking it is true that the peak output power of the amp, in Watts, can be considered the same regardless of the input signal. There may be some small observed differences with some very brief transient signals on the order of a couple of dB (is this what is called dynamic headroom?), although this varies from amp to amp.

Yes, I think so.

--Ethan

RealTraps - The acoustic treatment experts

To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.

Ethan Winer is offline  
post #17 of 42 Old 07-15-2013, 03:24 PM - Thread Starter
Advanced Member
 
Nyal Mellor's Avatar
 
Join Date: Aug 2010
Location: SF Bay Area, California, USA
Posts: 956
Mentioned: 5 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 75
Finally then when specifying an amplifier to reproduce 105dB peaks (85dB program with 20dB crest factor) it seems logical to specify the amplifier such that it's RMS rated power reproduces 102dB continuous (105dB peaks, 3dB crest factor). Put another way, we can derate the RMS power by 17dB to get the equivalent power output with movie content.

Most seem to be simply taking the Watts necessary to reproduce 105dB peaks and picking the an amp with the equivalent RMS Watts, which would be double the amplifier size required (3dB).


To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.
. We specialize in the design and creation of high performance listening rooms, home theaters and project studios for discerning audio/video enthusiasts.
Nyal Mellor is online now  
post #18 of 42 Old 07-15-2013, 06:33 PM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,381
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 748 Post(s)
Liked: 1162
Quote:
Originally Posted by Nyal Mellor View Post

Thanks for the replies...

So generally speaking it is true that the peak output power of the amp, in Watts, can be considered the same regardless of the input signal.

Probably not. The maximum output voltage without clipping that you can get out of a power amp will depend on the signal. Pure tones demand more power supply and heat sinks than music. The less extreme low frequency content, the less extreme highs, and the more impulsive nature of the music, the more voltage the amp can deliver.
Quote:
There may be some small observed differences with some very brief transient signals on the order of a couple of dB (is this what is called dynamic headroom?), although this varies from amp to amp.

Yes. But the bigger difference is amplifying pure tones as compared to music. That can be an 8 dB or more difference. The dynamics of the music can give you a 1-3 dB difference, more or less.
Quote:
With respect to crest factor for movies. Again generally speaking if the average movie soundtrack level is 85dB SPL* with peaks of 105dB then by definition we have a 20dB crest factor. *I believe the 85dB SPL actually refers to the average dialog level when set up as per THX guidelines so the program level could be more or less than this and will vary from movie to movie.

Music can have a comparable crest factor to movies. The crest factor of a recording can be varied during recording and production.
arnyk is offline  
post #19 of 42 Old 07-15-2013, 09:06 PM - Thread Starter
Advanced Member
 
Nyal Mellor's Avatar
 
Join Date: Aug 2010
Location: SF Bay Area, California, USA
Posts: 956
Mentioned: 5 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 75
Quote:
Originally Posted by arnyk View Post

Probably not. The maximum output voltage without clipping that you can get out of a power amp will depend on the signal. Pure tones demand more power supply and heat sinks than music. The less extreme low frequency content, the less extreme highs, and the more impulsive nature of the music, the more voltage the amp can deliver.
.

Really, so you can get a significantly more peak output voltage with a music signal as opposed to a sine wave? From what A9X-308 said above the supply rails determine peak output voltage and I imagine they are reasonably steady in a well designed amp.

I would 100% understand what you say if you were only talking about average power or whatever the correct term is for that. With low crest factor signals the average power required is much higher than with high crest factor signals.


To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.
. We specialize in the design and creation of high performance listening rooms, home theaters and project studios for discerning audio/video enthusiasts.
Nyal Mellor is online now  
post #20 of 42 Old 07-16-2013, 01:38 AM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,381
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 748 Post(s)
Liked: 1162
Quote:
Originally Posted by Nyal Mellor View Post

Quote:
Originally Posted by arnyk View Post

Probably not. The maximum output voltage without clipping that you can get out of a power amp will depend on the signal. Pure tones demand more power supply and heat sinks than music. The less extreme low frequency content, the less extreme highs, and the more impulsive nature of the music, the more voltage the amp can deliver.
.

Really, so you can get a significantly more peak output voltage with a music signal as opposed to a sine wave?

Yes. For a middle-of-the-road AVR running all channels, yes and by quite a bit. For a power amp with a very stiff power supply like a Krell, yes but less so.
Quote:
From what A9X-308 said above the supply rails determine peak output voltage and I imagine they are reasonably steady in a well designed amp.

He appears to be using a power amp like a Krell, with a very stiff power supply, as his standard.
Quote:
I would 100% understand what you say if you were only talking about average power or whatever the correct term is for that. With low crest factor signals the average power required is much higher than with high crest factor signals.

Yes .But as everybody seems agree, music has a large crest factor. Sine wave = 3 dB, dynamic music or movie = 20 dB. Music never has a crest factor of less than something like 8 dB. 3 dB is double the power, 6 dB is 4 times the power, 10 dB is 10 times the power and 20 dB is 100 times the power.

Here is a relevant graphic from a thread in the stereo forum:



And here is a conservative estimate of what the same average AVR would look like for low crest factor music:



So, the first graph looks pretty dire, and that is what the high end ragazines publish. Coincidentally, it looks like something that sells expensive amplifiers. Whooda thunk! ;-)

The second graph is much more like what people actually hear in their listening rooms.
arnyk is offline  
post #21 of 42 Old 07-16-2013, 07:22 AM
AVS Special Member
 
DonH50's Avatar
 
Join Date: Feb 2010
Location: Monument CO
Posts: 6,120
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 171 Post(s)
Liked: 263
Quote:
Originally Posted by Nyal Mellor View Post

With low crest factor signals the average power required is much higher than with high crest factor signals.

Depends on how loud the average power level sounds for each case...

"After silence, that which best expresses the inexpressible, is music" - Aldous Huxley
DonH50 is online now  
post #22 of 42 Old 07-16-2013, 09:22 AM - Thread Starter
Advanced Member
 
Nyal Mellor's Avatar
 
Join Date: Aug 2010
Location: SF Bay Area, California, USA
Posts: 956
Mentioned: 5 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 75
Quote:
Originally Posted by arnyk View Post

Yes. For a middle-of-the-road AVR running all channels, yes and by quite a bit. For a power amp with a very stiff power supply like a Krell, yes but less so.

Thanks for the info smile.gif We may be talking about two slightly different things here. Not sure.

Are you referring to the peak power from a single channel or are you referring to the sum of the peak output power of the whole multichannel amp?

For a single channel your graph seems to indicate that peak power is the same (or almost the same) whether one is playing a sine wave or music, yet the sum of the peak power of the individual amp channels is greater for music than for movies. I guess this makes sense because the limiting factor for the multichannel amps you refer to is the power supply whereas for a single channel it is it's ability to swing voltage (??).


To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.
. We specialize in the design and creation of high performance listening rooms, home theaters and project studios for discerning audio/video enthusiasts.
Nyal Mellor is online now  
post #23 of 42 Old 07-16-2013, 10:17 AM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,381
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 748 Post(s)
Liked: 1162
Quote:
Originally Posted by Nyal Mellor View Post

Quote:
Originally Posted by arnyk View Post

Yes. For a middle-of-the-road AVR running all channels, yes and by quite a bit. For a power amp with a very stiff power supply like a Krell, yes but less so.

Thanks for the info smile.gif We may be talking about two slightly different things here. Not sure.

Are you referring to the peak power from a single channel or are you referring to the sum of the peak output power of the whole multichannel amp?

Either
Quote:
For a single channel your graph seems to indicate that peak power is the same (or almost the same) whether one is playing a sine wave or music,

More true for a multichannel ampliifer.
Quote:
yet the sum of the peak power of the individual amp channels is greater for music than for movies.

Got data to back that up? IME it can go either way.
Quote:
I guess this makes sense because the limiting factor for the multichannel amps you refer to is the power supply whereas for a single channel it is it's ability to swing voltage (??).

Don't try to read too much precision into my example. My example has "approximate" written all over it!

The point I was trying to make is that music and movies draw enough less current than a pure sine wave that you don't lose nearly as many wpc as you run signals through more and more channels.
arnyk is offline  
post #24 of 42 Old 07-16-2013, 10:59 AM - Thread Starter
Advanced Member
 
Nyal Mellor's Avatar
 
Join Date: Aug 2010
Location: SF Bay Area, California, USA
Posts: 956
Mentioned: 5 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 75
Do you have or can you point me to more data on peak power testing for amplifiers especially the point about peak power output being higher for music than for sine wave? I'd be interested in seeing the data if you have any.


To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.
. We specialize in the design and creation of high performance listening rooms, home theaters and project studios for discerning audio/video enthusiasts.
Nyal Mellor is online now  
post #25 of 42 Old 07-16-2013, 12:44 PM
AVS Special Member
 
DonH50's Avatar
 
Join Date: Feb 2010
Location: Monument CO
Posts: 6,120
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 171 Post(s)
Liked: 263
The issue is music, typically having a much higher peak-to-average ratio than a sine wave, does not stress the power supply as hard. That is, a pure sine wave has a pretty constant average power and low crest factor (peak-to-average ratio). Music with the same average power ratio has much higher peaks. The power supply does not sag as much for short peaks as it does for steady-state sinusoidal signals (often used during testing). As a result, typical receivers do not lose as much power in the real world as steady-state multichannel testing would indicate.

This of course led to things like "peak music power" ratings that, without standardization, became the target of many jokes and grossly inflated specifications.

This Wikipedia article has some info: http://en.wikipedia.org/wiki/Audio_power It suggests 20 dB. Arny, do you have the AES article showing 17 dB? I think it was in the late 70's or early 80's... They studied a bunch of music and determined the peak-to-average ratio averaged around 17 dB. Compressed pop a little less; some orchestral etc. a little more.

"Dynamic headroom" was a popular term for some time and designated the amount of power an amplifier could produce before the rails sagged, typically 10 - 100 ms (I do not recall the spec). I think Bryston specified 3 dB (do not know if they still use that rating). Amplifiers with heavily-filtered or regulated supplies (Krell, ML, Sanders come to mind) had very little "dynamic headroom" because their power supplies were stable. When I see an AVR rating of X W/ch in stereo that drops to X/2 or lower with multiple channels, that is an indication the power supply (and perhaps thermal management) is not up to supplying all channels at once. Whether that matters in the real world, well, probably not for the vast majority of cases.

"After silence, that which best expresses the inexpressible, is music" - Aldous Huxley
DonH50 is online now  
post #26 of 42 Old 07-16-2013, 01:42 PM - Thread Starter
Advanced Member
 
Nyal Mellor's Avatar
 
Join Date: Aug 2010
Location: SF Bay Area, California, USA
Posts: 956
Mentioned: 5 Post(s)
Tagged: 0 Thread(s)
Quoted: 52 Post(s)
Liked: 75
Quote:
Originally Posted by DonH50 View Post

The issue is music, typically having a much higher peak-to-average ratio than a sine wave, does not stress the power supply as hard. That is, a pure sine wave has a pretty constant average power and low crest factor (peak-to-average ratio). Music with the same average power ratio has much higher peaks. The power supply does not sag as much for short peaks as it does for steady-state sinusoidal signals (often used during testing). As a result, typical receivers do not lose as much power in the real world as steady-state multichannel testing would indicate.

Bingo, thanks Don, that makes a ton of sense.

It kind of makes it difficult to specify amplifier power required to meet SPL requirement when using AVRs and power amps with saggy voltage rails, THOUGH I guess if one was always conservative and used published RMS test figures then that would be a safe way to do it, even if you ended up with more amp power than strictly required.

Do all AVRs exhibit the same power supply performance or does it tend to be the cheaper ones? Do any outboard power amps have saggy rails?


To view links or images in signatures your post count must be 0 or greater. You currently have 0 posts.
. We specialize in the design and creation of high performance listening rooms, home theaters and project studios for discerning audio/video enthusiasts.
Nyal Mellor is online now  
post #27 of 42 Old 07-16-2013, 02:45 PM
AVS Special Member
 
JorgeLopez11's Avatar
 
Join Date: Oct 2003
Location: MEXICO CITY
Posts: 1,887
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 18
Quote:
So, the first graph looks pretty dire, and that is what the high end ragazines publish. Coincidentally, it looks like something that sells expensive amplifiers. Whooda thunk! ;-)

The second graph is much more like what people actually hear in their listening rooms.
 

 

Arny, this was exactly wat we were discussing at the other thread. This case is not surprising at all. 

 

I will always prefer a modestly rated amp with generous headroom instead of a supposedly powerful amp with anemic PSU than can not sustain peaks at all. IMO AVRs other than flagships are good examples. 

JorgeLopez11 is offline  
post #28 of 42 Old 07-16-2013, 02:50 PM
AVS Special Member
 
JorgeLopez11's Avatar
 
Join Date: Oct 2003
Location: MEXICO CITY
Posts: 1,887
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 18
Quote:
Originally Posted by Nyal Mellor View Post


Bingo, thanks Don, that makes a ton of sense.

It kind of makes it difficult to specify amplifier power required to meet SPL requirement when using AVRs and power amps with saggy voltage rails, THOUGH I guess if one was always conservative and used published RMS test figures then that would be a safe way to do it, even if you ended up with more amp power than strictly required.

Do all AVRs exhibit the same power supply performance or does it tend to be the cheaper ones? Do any outboard power amps have saggy rails?

Nyal, IME cheaper AVRs do not have exemplary performance. Or at least not the models that hit the market since some 10 years ago. Flagship AVRs and power amplifiers are usually better.

JorgeLopez11 is offline  
post #29 of 42 Old 07-16-2013, 05:13 PM
AVS Addicted Member
 
arnyk's Avatar
 
Join Date: Oct 2002
Location: Grosse Pointe Woods, MI
Posts: 14,381
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 748 Post(s)
Liked: 1162
Quote:
Originally Posted by Nyal Mellor View Post


It kind of makes it difficult to specify amplifier power required to meet SPL requirement when using AVRs and power amps with saggy voltage rails,

Not really, because even the low end but mainstream brand AVRs with the saggiest rails are still adequate.
Quote:
Do all AVRs exhibit the same power supply performance or does it tend to be the cheaper ones?

No, some high end AVRs are overbuilt.
Quote:
Do any outboard power amps have saggy rails?

Yes.
arnyk is offline  
post #30 of 42 Old 07-17-2013, 02:25 AM
 
OllieS's Avatar
 
Join Date: Apr 2013
Posts: 251
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 16
Quote:
Originally Posted by DonH50 
The power supply does not sag as much for short peaks as it does for steady-state sinusoidal signals (often used during testing). As a result, typical receivers do not lose as much power in the real world as steady-state multichannel testing would indicate.

Do you have any evidence that suggests this or are you just making assumptions? I would love to see some cold hard data on this.
OllieS is offline  
Reply Audio theory, Setup and Chat

User Tag List

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off