Power requirements for speakers to maximise performance.

tebbo65

Well-known Member
Having a bit of a re-config of my system and need advice on power requirements for speakers to maximise performance. What's got me thinking is a fellow AVF member suggested my LCR are under driven and would benefit from more power to improve dynamic range etc, even at lower volume? He also suggested XTZ's power rating is quite conservative.
See below more info on my present set up for initial advice. Note, I do need all channels driven by external power amps as I am about to introduce 2 x MiniDSP ddrc88BM's to try out Dirac.

So my question is, given that I may well be going to a 9.1.4 config (wides will be XTZ M6's), so will need an extra 2 channels of amplification what wattage rated power amps should I be looking at and should bi-amping the LCR be a consideration? Or am I over thinking it and what's in place is good enough?

Typical listening volume max -12dB.
All speakers crossed over with subs at 80Hz, so subs handle all the heavy lifting in the bass/sub bass region.
Distance from the LCR to the MLP is approx. 10ft.

Set up. (7.2.4)

  • LCR (XTZ M8's) 91,2dB 600w Peak 350w Cont 4-8 Ohm
  • Side Surrounds (XTZM6's) 89dB 300w Peak 150w Cont 4-8Ohm
Above currently powered by a Emotiva XPA-5 gen 2.
200w per channel; all channels driven; into 8 Ohms.
300w per channel; all channels driven; into 4 Ohms.
  • Rears (XTZ M6's) 89dB 300w Peak 150w Cont 4-8 Ohm
  • Over heads (XTZ S2's) 86dB 150w 75w Cont 8 Ohm
Above currently powered by a Nak AVP-1.
110w per channel; all channels driven; into 8 Ohms.
170w per channel; all channels driven; into 4 Ohms.

Thanks in advance.
 
Last edited:

tebbo65

Well-known Member
Emotiva, iota and nakamichi are all made by same OEM. They're pretty much same amp.

Why would you change one to anothrr?
Because at the mo I've got 12 channels of amplification feeding 11 speaker channels. If I go with wides I'll need 13 channels of amplification so I'll be one channel short. 2 x IOTA's/Naks will give me 14 channels.
 
Upvote 0

Mr Wolf

Well-known Member
At your -12dB listening you’d even be fine on AVR power but if you give me speaker distances to the MLP and their respective sensitivity ratings I can drop them into a spreadsheet model that estimates the peak power needs of all the speakers.
 
Upvote 0

tebbo65

Well-known Member
Bit of extra info and my forward planning;
  • I have a pair of ex-demo M6's I got a great deal on from XTZ arriving on Thursday, so I'll have the speakers for wides ready to go when the time comes.
  • I'm going to try the two MiniDSP DDR88-BM's + MiniDSP 2x4HD (to sum the subs), to try Dirac in my system as is. 7.1.4.
  • If I like, and I'm sure I will what Dirac does I will then look the buy a processor with at least 13 channels of processing and Dirac live built in. Monoprice HTP-1 or the like, maybe when a newer model comes out???
  • Or I may just go for a Denon 6700 (in pre-pro mode) and run the DDR88's.
  • In the meantime I would like to stick a power amp in my 2nd system to take some of the strain off the Denon 4500. So I thought I may as well get the IOTA now for the main system and it'll be ready to rock when the time comes amplification channel wise and put the Emotiva in No2 system now.
Does all that make sense? 🤯
 
Upvote 0

Mr Wolf

Well-known Member
I guess it does but as you’re going the power amp route anyway why not just get an AVP that supports Dirac and lose the DDR-88s or do they do things that an AVP can’t?
 
Upvote 0

tebbo65

Well-known Member
At your -12dB listening you’d even be fine on AVR power but if you give me speaker distances to the MLP and their respective sensitivity ratings I can drop them into a spreadsheet model that estimates the peak power needs of all the speakers.
That's great much appreciated, thank you.
Reason I need external power amps is to run the MiniDSP ddr88-BM's.

Centre - 10ft - 91.2dB
L&R - 10.7ft - 91.2dB
Side Surrounds - 6.3ft - 89dB
Rears - 5.3ft - 89dB
Top Fronts - 7.7ft - 86dB
Top Rears - 6.9ft - 86dB
Wides (If and when) approx 9ft - 89dB
 
Upvote 0

tebbo65

Well-known Member
I guess it does but as you’re going the power amp route anyway why not just get an AVP that supports Dirac and lose the DDR-88s or do they do things that an AVP can’t?
Absolutely right, that will be the route I go down and will make sense to get an AVP with Dirac and then offload the 88's. They are just a very cost effective (for what I paid) way to try Dirac with any AVR/C or P but do need to be hooked up via external power amps.
 
Upvote 0

Mr Wolf

Well-known Member
Here you go. The yellow boxes are the model's assumptions and the rest is formulae.

Key assumptions are:
  • All speakers are on-axis to the MLP apart from the heights so I've de-rated their sensitivity by 2dB
  • Mid-point for other speaker sensitivity assumptions
  • Emotiva XPA-5 powers LCRs/sides and an IOTA AVP1X or equivalent is on the rest so you'd need two of them
  • Minimum speaker impedance of 4-Ohms
  • Amplifier power output at 4-Ohms for High/Low = two/all channels driven

As expected, at -12dB volume you should only be using a tiny amount of power and have at least 15dB headroom in all channels so the system should be reference capable with 5dB headroom.

1647959094967.png
 
Upvote 0

tebbo65

Well-known Member
Here you go. The yellow boxes are the model's assumptions and the rest is formulae.

Key assumptions are:
  • All speakers are on-axis to the MLP apart from the heights so I've de-rated their sensitivity by 2dB
  • Mid-point for other speaker sensitivity assumptions
  • Emotiva XPA-5 powers LCRs/sides and an IOTA AVP1X or equivalent is on the rest so you'd need two of them
  • Minimum speaker impedance of 4-Ohms
  • Amplifier power output at 4-Ohms for High/Low = two/all channels driven

As expected, at -12dB volume you should only be using a tiny amount of power and have at least 15dB headroom in all channels so the system should be reference capable with 5dB headroom.

View attachment 1671581
Wow, great information. So to be a total pain I wonder if the numbers would be okay if I used IOTA's across the board and took out the EMO as I want to, (to give me the extra channel required via just two units).
If it really is no trouble and you have time would you mind running the numbers again but just with IOTA's? I wish I could electronically send you a couple of pints as appreciation🍻.
BTW the overheads are ON-axis.
 
Upvote 0

Mr Wolf

Well-known Member
No trouble at all, it took 30 seconds to make that change and copy it over.

You now get a bit less headroom in the LCR/side channels.


1647961393963.png


EDIT: have adjusted heights for being on-axis
 
Upvote 0
D

Deleted member 901590

Guest
With the power behind the x6700, you could consider part using its internal amps?
 
Upvote 0

Mr Wolf

Well-known Member
With the power behind the x6700, you could consider part using its internal amps?
No, you missed the part about using Dirac via the DDR-88s - power amps are essential.
 
Upvote 0
D

Deleted member 901590

Guest
No, you missed the part about using Dirac via the DDR-88s - power amps are essential.

Ah, yes, sorry. I'd remembered the bit about potentially using an x6700 but get that direction is decided now. Just trying to save cash 🤓
 
Upvote 0

tebbo65

Well-known Member
No trouble at all, it took 30 seconds to make that change and copy it over.

You now get a bit less headroom in the LCR/side channels.


View attachment 1671613

EDIT: have adjusted heights for being on-axis
You are a true gent Sir. Thank you I can now do a direct comparison, perfect.
If I'm ever down in Wiltshire or you are anywhere near Newport Pagnell, we'll hook up and I'll buy you a beer..........or several!
 
Upvote 0

tebbo65

Well-known Member
Ah, yes, sorry. I'd remembered the bit about potentially using an x6700 but get that direction is decided now. Just trying to save cash 🤓
You are spot on though buddy. I really need to get cracking and see if Dirac is a marked/worthwhile improvement.
At the mo I have the subs summed and REW Eq'd with a custom curve via MiniDSP 2x4HD. Auto EQ is limited to the schroeder freq of my room 213Hz via the Audyssey app. I don't use Dynamic EQ or any other Audyssey features. I've integrated (in order) the centre, mains and so on with the subs using the sub/speaker distance tweak. So all Audyssey is doing is distance setting, Eqing & bass managing all speakers and not touching the subs whatsoever as I limited the sub Audyssey EQ to zero Hz via the app.
It's a well treated room with properly positioned speakers and the general consensus is it sounds great, also the opinion of Dr Jules (aka AVF podcast member) and he's HAA certified an all 🤣.
So, if I don't get on with Dirac and revert back to what am running now then that's when the 6700 or equivalent may come into play for the extra wide channels and then I would use some of it's excellent internal amplification.
Must get those ddr88's into action 😵‍💫.
 
Last edited:
Upvote 0

dlaloum

Active Member
I know the theory and overall agree with what you're saying. In fact with certain speakers it can be even more demanding than you're suggesting due to added power inefficiency presented by large phase angles.

That said, this potential impedance/current swing issue really needs to be put into perspective where the overall power requirement is still relatively low and well within a PSU's current capabilities. For example, take a situation like the OPs where he may need 4W output into 8-Ohms. That's a peak voltage to the speaker of 5.66V at 0.71A. Even in a 3-Ohm impedance dip that's still only a maximum power draw of 10.7W on 1.89A of current. That's the same amount of current that would be drawn when outputting 29W into an 8-Ohm load so it shouldn't be particularly demanding. And all this is happening in transient peaks that last fractions of a second so that the PSU demands can be smoothed by large capacitors.

If you're driving a system pretty hard with difficult to drive speakers then I agree that choice of amplifier can become very critical. Personally I try and stay away from those sort of speakers.

Out of interest, the THX Ultra standard requires amps to be stable into 3.2-Ohm loads and be able to cover maximum current swings of up to 18A.
To a point I agree with you... but the system needs to be able to provide for +20db peaks - and those peaks must be well within the limits of the amp, otherwise it will clip - and once it clips the effect on the sound is distributed throughout the frequency range and is seriously NASTY.

So Amp sizing needs to be based on the worst case impedance point, and with +20db capability at that point.

As to speaker choices - I have always chosen speakers based on their sound - with my benchmark having been full range Electrostatics since the mid 80's (I have owned several different Quad ESL's over the years...) - the transparency and low distortion standards set by these benchmarks make it hard to find suitable speakers.... one day walking through a department store (of all places), I heard the Gallo Ref 3.2's playing.... unexpectedly, I though "they sound like electrostatics" (a high accolade) - they also provided me with a much more WAF friendly setup....

I have not heard speakers that could match them for the aspects that I value, within 3x or 4x their price....

They sound "good" with Quad current dumping amps (405, 606, 707, 909, etc...) or with most amps that are stable into loads below 2 ohm and have sufficient power at those impedances.
(you can pick up one of the older Quad amps for very reasonable $) - but they do sound better (albeit slightly) with the massively powerful Crown XLS2500 driving them....
In theory a Benchmark AHB2 would do fine as would a Purifi based amp with a robust PSU.... but all the alternate options would cost substantially more (2x to 3x the price) - for what is most likely a minor or inaudible improvement (high SINAD).

Right now I am satisfied with my system as a high performing, value for money setup.
 
Upvote 0

Mr Wolf

Well-known Member
To a point I agree with you... but the system needs to be able to provide for +20db peaks - and those peaks must be well within the limits of the amp, otherwise it will clip - and once it clips the effect on the sound is distributed throughout the frequency range and is seriously NASTY.
Actually it look likes you agree with me completely as all my power requirement calculations conservatively assume maximum cinematic 20dB peaks in the LCR channels off an 85dB reference (even though in practice on many consumer mixes they're pulled down to 17dB and true reference is 83dB).

In case you're not aware, in Dolby Atmos cinema mixes bed level surround peaks are capped at 17dB, Atmos channels at 14dB and the summed SPL output of all these (non-LCR) channels cannot exceed the maximum output of a single LCR channel. These SPL peak limits are imposed by the software (e.g. Pro Tools) at the mixing stage.

The upshot of all this is that no more than the equivalent of four amplifier channels are fully driven simultaneously at any point. It's been that way since the first days of Dolby Surround. All subsequent surround formats have done is change where the sound might come from, not increase the SPL hitting the viewer. That would be painful.
 
Upvote 0
D

Deleted member 901590

Guest
and the summed SPL output of all these (non-LCR) channels cannot exceed the maximum output of a single LCR channel.

A very important fact. And quite funny that this has been around for a long time, as the most appropriate power quote for and AVR for ACD should best be into 4 channels.

And, to quote something that you've said many times already on AVF, people need to think about this fact if they increase their system from (say) 5.x.0 up to 7.x.4. They've added six channels who's power demands are still summed into the two surrounds they already have.
 
Upvote 0

Mr Wolf

Well-known Member
A very important fact. And quite funny that this has been around for a long time, as the most appropriate power quote for and AVR for ACD should best be into 4 channels.
Actually I would still prefer to see 5-channel driven ratings rather than 4-channel ones for two reasons.

The main reason is running three amp channels at 100% power and say four surrounds equally at 25% power (so each surround is 6dB down but their summed SPL equals one LCR) would draw more power from a PSU than running four channels at 100% power. This is due to lower power efficiency that stems from a lower level of amplifier utilisation of each amp's potential. See chart below - a Class A/B amp driven at 25% is circa 30% energy efficient vs 55% when fully driven.

1648034991414.png

The other reason is that it provides some headroom in case the sound mixer didn't do his job properly and overcooked things a little.

Maybe this is why THX subject amps to simultaneous 1-5 channel driven testing (but no higher) as part of its certification process.

I think it's also no coincidence that 5-channels driven is the basis of D&M's 70% output guarantee. It's simply the most relevant one to consider for a worst case scenario.

Ideally we'd have a burst (not continuous) output rating though at 4-ohms (where most speakers dip to) to better reflect a real-world worst case scenario. As it stands, amplifier capacitance isn't factored into the numbers either which is misleading. The continuous ACD ratings we're used to seeing are effectively a PSU output stress test which doesn't tell the full story.
 
Last edited:
Upvote 0
D

Deleted member 901590

Guest
And also considering that the multiple non-LCR speakers could all have varying sensitivities, impedance, phase etc to one another.

But, yes, see the sense in 5 ACD but as you say no more. Would be interesting to see some dB maps of films, to see how close it can ever get to simultaneous peaks.

When you try to think of it, you think of explosions as a key example, when actually it's probably not at all because its probably the subs doing most of the hard work :) As said before, with crossovers, the subs remove huge chunks of power demand from the other speakers.
 
Upvote 0

Mr Wolf

Well-known Member
But, yes, see the sense in 5 ACD but as you say no more. Would be interesting to see some dB maps of films, to see how close it can ever get to simultaneous peaks.
According to Anthony Grimani (and he should know) THX developed its own software to read the SPL dynamics of digital soundtracks and fed 100s of them through it when developing its standards. Apparently they used this info to create a burst cycle algorithm to measure amp/speaker compression/distortion at different output levels. As a 20dB transient peak requires 100x the power and are very short their process primarily focuses on burst tests, not continuous ones, similar to CEA-2010 does on subs.

I've also seen an independent research paper that found (in the soundtracks they tested) that >99% of the dynamic peaks >90Hz lasted <40ms. Given this level of duration you can see the role capacitance must play. Conversely, >99% of peaks <90Hz lasted <300ms.
 
Upvote 0
D

Deleted member 901590

Guest
Conversely, >99% of peaks <90Hz lasted <300ms.

I'm surprised at that bit, I wouldn't thought most of them would've been longer. I'm sure, for example, there are a lot of very long LFE channel peaks. Or was that not including .1?
 
Upvote 0

Mr Wolf

Well-known Member
I'm surprised at that bit, I wouldn't thought most of them would've been longer. I'm sure, for example, there are a lot of very long LFE channel peaks. Or was that not including .1?
It might appear continuous but what you're hearing is going up and down all the time - that's what real world sounds do, they're not test tones.

Here's the research paper. It's on page 7.

 
Upvote 0

The latest video from AVForums

Samsung S95B QD-OLED Review - A Quantum Leap for OLED!
Subscribe to our YouTube channel

Latest News

Amazon Prime Day 2022: 12-13 July
  • By Ian Collen
  • Published
PSB Speakers unveils Passif 50 loudspeaker
  • By Ian Collen
  • Published
Samsung Gaming Hub goes live
  • By Ian Collen
  • Published
Huawei unveils FreeBuds Pro 2 true wireless earbuds
  • By Ian Collen
  • Published

Full fat HDMI teeshirts

Support AVForums with Patreon

Top Bottom