Amount of power needed for dynamic peaks?

[Trdat]

I think I saw a conversation about this on the PS Audio amplifier review and it was suggested to move it to a new thread. You guys can go gung ho here if you like, but my question is simpler and most of the concepts were to technical for me anyway. I am more than happy to take Amirs take on the issue but the how do we work out how many watts we need if we do use Amir's theory that power demands must be met to get the best out your dynamic peaks.

The Harman calculator gives a reading of 11000 watts have I misunderstood something?

Example,

A Deltalite 2515 at 99db sensitivity 300 watts RMS 600 watts peak, I will be using an active crossover with DSP from the computer so I am guessing no power dissipation due to the lack of passive crossover. But it will be in a sealed box.

If I want the dynamics of large fluctuations of DB's how many watts would I need?

 

[hardisj]

Read this:
https://www.crownaudio.com/how-much-amplifier-power


Then use this:
https://www.crownaudio.com/en/tools/calculators



Example A: High Sensitivity speaker, Medium Listening Volume, Average Listening Distance

You have the following:

• Speaker Sensitivity = 90 dB @ 1w/1m
• Distance of speaker to MLP = 3m
• Desired listening level = 85dB @ MLP

Solve for desired listening level above for:
Highly compressed music (3dB dynamic range):
Needed power = 6 watts

If you listen to music that has high dynamic range, say 15dB (80's pop music typically has 10-15dB DR) then:
Needed power = 90 watts



So not much power. But when you have a speaker with more typical sensitivity such as 85dB @ 1w/1m and you want to listen loud, this is where things get hairy...

Example B: Medium Sensitivity speaker, Louder Listening Volume, Average Listening Distance

You have the following:

• Speaker Sensitivity = 85dB @ 1w/1m
• Distance of speaker to MLP = 3m
• Desired listening level = 93dB @ MLP

Solve for desired listening level above for:
Highly compressed music (3dB dynamic range):
Needed power = 113 watts

If you listen to music that has high dynamic range, say 15dB (80's pop music typically has 10-15dB DR) then:
Needed power = 1796 watts

 

[pozz]

If you're comfortable with the calculations, I attached a spreadsheet here: https://www.audiosciencereview.com/...ow-often-do-amplifiers-clip.12757/post-380179 It's not completely automated so you'll have to punch in some of the values.

Calculate for the amount of SPL at the listening position. Allow peak to be a certain amount, with the RMS some amount of dB below that.

 

[Duke]

A Deltalite 2515 at 99db sensitivity 300 watts RMS 600 watts peak, I will be using an active crossover with DSP from the computer...

That 99 dB figure includes the rising top-end and intentional 106 dB peak at 2.5 kHz, which make the Deltalite 2515 suitable for some electric guitar applications. The calculated efficiency based on T/S parameters is about 96 dB @ 1 watt, and is imo a more realistic figure for a home audio application.

Regarding calculating the power needed for unclipped peaks, I've found this online calculator to be useful, though of course it doesn't factor in thermal modulation which may occur if the driver(s) are being pushed hard:

http://myhometheater.homestead.com/splcalculator.html#anchor_13193

 

[Trdat]

If you listen to music that has high dynamic range, say 15dB (80's pop music typically has 10-15dB DR) then:
Needed power = 1796 watts

Is this answer following on from your examples or for my speaker of 99db sensitivity?

 

[Trdat]

If you're comfortable with the calculations, I attached a spreadsheet here: https://www.audiosciencereview.com/...ow-often-do-amplifiers-clip.12757/post-380179 It's not completely automated so you'll have to punch in some of the values.

Calculate for the amount of SPL at the listening position. Allow peak to be a certain amount, with the RMS some amount of dB below that.

I just checked it out, but its all for a 84db sensitive speaker I am perplexed on how to do the calculations for a 99db or 96db(according to Duke) driver.

 

[Trdat]

That 99 dB figure includes the rising top-end and intentional 106 dB peak at 2.5 kHz, which make the Deltalite 2515 suitable for some electric guitar applications. The calculated efficiency based on T/S parameters is about 96 dB @ 1 watt, and is imo a more realistic figure for a home audio application.

Regarding calculating the power needed for unclipped peaks, I've found this online calculator to be useful, though of course it doesn't factor in thermal modulation which may occur if the driver(s) are being pushed hard:

http://myhometheater.homestead.com/splcalculator.html#anchor_13193

So I am presuming that I need to play around with the watts till I get a decent SPL right? But does this factor in the peaks of dynamic range? That is what I am essentially looking for....

 

[Trdat]

https://www.crownaudio.com/how-much-amplifier-power

Then use this:
https://www.crownaudio.com/en/tools/calculators

https://www.crownaudio.com/en/tools/calculators

I dropped the SPL down to about a 100db and I got a normal calculation. With a 100 SPL I get 350 watts with a 99Db sensitivity. With 96 db sensitivity I get 715 watts, that's okay. Perhaps my SPL number was just to high for home use.

 

[Duke]

So I am presuming that I need to play around with the watts till I get a decent SPL right? But does this factor in the peaks of dynamic range? That is what I am essentially looking for....

Yes. But you have to figure out on your own what you want the peak SPL to be at your listening position. So if you want to listen at 85 dB average and have enough headroom for 20 dB uncompressed peaks, then add the two numbers, 85 + 20 = 105 dB. In this example you want enough amplifier power to hit 105 dB at the listening position.

Input your speaker's efficiency, number of speakers, and your listening distance. (I'd suggest using "away from walls" unless you have a good reason to include the walls, as reinforcement from room reflections is usually not very much in the midrange and treble regions, and if you want to you can mentally tack on another dB or two for room reflections. Note that it's not uncommon for speaker manufacturers to have already included room reflections in their efficiency claim, so you definitely don't want to include it twice.) Then trial-and-error your amplifier power to figure out how much you need to reach your unclipped peak target.

So if you have 96 dB efficient speakers (based on the Deltalite 2512) and you sit 10 feet from the speakers, after a little trial-and-error it looks like 37 watts will get you 105 dB peaks.

If you decide to assume that room reflections contribute 1.5 dB at the listening position, then your target can be 105 - 1.5 = 103.5 dB, and you can get there with 26 watts.

 

[pozz]

According to @Duke's spec of 96dB.

 

[Duke]

According to @Duke's spec of 96dB.

View attachment 77787
View attachment 77788

Pozz, is your chart for a single speaker? If so, then I think Trdat can assume 3 dB more output (or 1/2 the power requirement for a given SPL) from adding a second speaker.

 

[pozz]

Pozz, is your chart for a single speaker? If so, then I think Trdat can assume 3 dB more output (or 1/2 the power requirement for a given SPL) from adding a second speaker.

Yup, I should have added this:

• For two speakers, add 3dB.
• For 4 ohm impedance, add 3dB.
• For 16 ohm impedance, subtract 3dB.

 

[Trdat]

Yes. But you have to figure out on your own what you want the peak SPL to be at your listening position. So if you want to listen at 85 dB average and have enough headroom for 20 dB uncompressed peaks, then add the two numbers, 85 + 20 = 105 dB. In this example you want enough amplifier power to hit 105 dB at the listening position.

So if you have 96 dB efficient speakers (based on the Deltalite 2512) and you sit 10 feet from the speakers, after a little trial-and-error it looks like 37 watts will get you 105 dB peaks.

If you decide to assume that room reflections contribute 1.5 dB at the listening position, then your target can be 105 - 1.5 = 103.5 dB, and you can get there with 26 watts.

That is great help thanks. So even on the Crown calculator the SPL it seems is the average SPL plus the peaks or its the maximum SPL minus the peaks for the average listening SPL.

So my 75 watt Hypex amp should be plenty for 85db average listening with up to 15db peaks. That's great!

 

[Duke]

Yup, I should have added this:

• For two speakers, add 3dB.
• For 4 ohm impedance, add 3dB.
• For 16 ohm impedance, subtract 3dB.

Imo the latter two points are true if you are using sensitivity at 2.83 volts (which is equal to 1 watt into 8 ohms).

However if you are using efficiency at 1 watt (which would be 2 volts into 4 ohms, or 4 volts into 16 ohms), then imo your charts are correct.

 

[Thomas_A]

If you split two amps for sub and monitors with x-over at 80 Hz, what would a 100 Watt amp correspond to driving the monitors comparing to driving full range? Approximately?

 

[Sancus]

Read this:
https://www.crownaudio.com/how-much-amplifier-power


Then use this:
https://www.crownaudio.com/en/tools/calculators

The calculator above and most of the calculations in this thread are using the -6db/double distance formula, but that's only true in an anechoic chamber or outdoors. In real rooms you lose closer to 3dB/DD and this changes the power calculations substantially at 3m+ distances.

 

[pozz]

The calculator above and most of the calculations in this thread are using the -6db/double distance formula, but that's only true in an anechoic chamber or outdoors. In real rooms you lose closer to 3dB/DD and this changes the power calculations substantially at 3m+ distances.

For that adjustment subtract 3dB to find the required wattage.

 

[Duke]

In real rooms you lose closer to 3dB/DD

Minus 3 dB per doubling of distance has not been my experience with point-source-approximating speakers (as opposed to line-source-approximating speakers) in a home audio setting, but perhaps you have measured the falloff with distance in more rooms than I have. I agree that it's a bit greater than the theoretical -6 dB per doubling of distance, but in my (limited) experience it has been closer to that than to -3 dB per doubling of distance.

 

[North_Sky]

Which frequencies of the full audio spectrum (octaves) would require the most amount of dynamic energy from powerful music recordings...the lower frequencies?

Some subwoofers can deliver few thousand watts on peaks.
Some tweeters can handle 1,000 watts plus on peaks.

 

[Sancus]

Minus 3 dB per doubling of distance has not been my experience with point-source-approximating speakers (as opposed to line-source-approximating speakers) in a home audio setting, but perhaps you have measured the falloff with distance in more rooms than I have. I agree that it's a bit greater than the theoretical -6 dB per doubling of distance, but in my (limited) experience it has been closer to that than to -3 dB per doubling of distance.

I'm just quoting Toole's work (he has previously made the same post correcting me, and I took his advice). I've only personally measured a few rooms since(obviously) but in my experiments with mostly condo and apartment sized spaces, 3dB is pretty close. Of course, you're rarely going to get to distances more than 3m in those spaces anyways.

Obviously the simplest way to solve this is just to measure your own space.... or be happy with your extra overhead from the strict calculation