SPL targets for speaker design

[Pan]

Graaf,

nice post from you there.

I have not read that particular work from KTH but I've read others. BTW Svante (Edge, Basta, Tombstone, Tone..) is the principal at one department of the Swedish Royal Institute of Technology.

Just want to add that in the end it's very unlikely to experience peaks above 110dB in a typical seating as far as I understand. Obviously the musicians and conductor are subject to higher levels than the visitor. Now that said I realise that classic concerts is not everyones reference but it's an interesting starting point for a music rig IMO.

Also I assume (without having measured myself or seen anything about it) that the peaks in a concert hall seat will be in the upper bass to lower midrange, meaning that the requierments are different for the different drivers in a multiway loudspeaker.

Regarding the dynamic range of music, I was hesitating to write anything at all regarding the headroom that we need from the system since it' hard to get anything down in numbers as you showed in your post. More interesting is to simply look at the total dynamic range of the music that you will use in your stereo rig. Wether a record has 10dB or 40dB of peaks above average is not very interesting but the ability to reproduce the peaks are (if hifi is the goal).


/Peter

 

[auplater]

calculations vs. measurements

In my rather largish listening space, the inverse square law seems closer to Salas's 5dB when listening to monopoles as measured, whereas with my linesource dipoles, there is, at most ~3dB/doubling out to maybe 15' or so (listening position in a arc ~15' wide) at most frequencies (at least until room modal effects come into play)

So the theoretical calculations need to be accepted with some regard to actual measurements. At 102dB average(C weighted) at my listening position, I find the levels about all I can take with classical, jazz, HT, rock, etc... so I'm more than satisfied with my setup, since this is ~-5dB (average, not peak measured) below the level where my system starts to noticeably distort.

If you want an excellent low bass jazz piece to test, try Gato Barbieri "Que' Pasa" cd, first 2 cuts rattle the rafters with upwelling synth bass below 30 Hz...

I use Holst as a reference as well, along with Mussoursky's "Gates of Kiev"... guaranteed to wake the neighbors...

John L.

 

[gedlee]

Re: Re: Re: SPL targets for speaker design

The inverse square is truncated to -5db per doubling of distance in my example in a domestic environment bcs that is all I have measured. That and the +3dB for stereo, it was calculated in my paradigm.

Dr. Earl Geddes would have protested immediately!😀

To be fair, I get annoyed by all the disinformation spread on these sites. Someone has to stand up and say "NO WAY!"

Floyd Toole reports about -3 dB per double distance. His measurements too are based more on a low directivity index. For my speakers I would estimate that -4 to -5 dB is more realistic. For a monopole, I would use -2 to -3 dB.

One of my hypothesis about the sound of my speakers is the lack of early refections AND that fact that the direct field is extended much further out into the room, hence putting the listener much further into the direct field.

 

[Salas]

Re: Re: Re: Re: SPL targets for speaker design

For a monopole, I would use -2 to -3 dB.

I confirm. I measured -2.5dB per distance doubling in a 30m2 furnitured and carpeted living room for true extended range monopoles that I have built for experimental purposes. These are full range (push push) located front and back on mini cabinet with a tweeter on top. You go around them, power does not change, all over their bandwidth.

In the picture, as copied by a Dutch DIYer who's driving them with a Hiraga amp. The originals have a slot port, no tw sensitivity selector, and an Audax 0.75 inch tw inset with a router.

 

[graaf]

Your +7 dB for the reverberant field is a bit high and your calculations are all based on monopole excitation. With my speakers I have almost 9 dB of DI which moves the reverberant distance out almost to the listener and so the reverberant addition would be 3 dB at the most and could even be less. This would push even your numbers up closer to what I was claiming.

I agree
the amount of this reverberant field gain depends on the directivity index of the speakers and on the volume and RT60 of the room and also is frequency-dependent
the case of my speakers is quite the opposite to Yours - they are very close to ideal monopole and there is practically no direct field at all in my listening room

Floyd Toole reports about -3 dB per double distance. His measurements too are based more on a low directivity index. For my speakers I would estimate that -4 to -5 dB is more realistic. For a monopole, I would use -2 to -3 dB.

In a living room you will never get +7dB. About +1.5dB from 1k to 3k for stereo sources is all I have measured in 25-40m2 living rooms.
(…)
I measured -2.5dB per distance doubling in a 30m2 furnitured and carpeted living room for true extended range monopoles that I have built for experimental purposes.

It all depends on relevant factors and differences in those factors explain all differences in the results
The sound level in a room is determined by the strength of the source, the volume of the room and the absorption at the walls.
Outside the reverberation distance the sound level is constant (independent of the distance to the sound source)
According to KHT handbook the typical values of reverberation distances for real rooms are around 0.5 m in case of small room (of volume of around 36 m3 as an example), 0.8 m in case of big room (of volume of around 36 m3 as an example) and 5 m in case of a concert hall (of volume of around 11 000 m3 as an example)
the reverberation distance is dependent on the volume of the room and its RT60

second question - what is this constant sound level outside the reverberation distance?
the people at KHT give an example of bass singer that "produces" 94 dB at 1 m and this what they found:

We use a bass singer and the three rooms with reverberation times of 0.5, 0.8 and 1.5 s, respectively. A little calculation shows that the sound levels (outside the reverberance distance) become 102 dB SPL, 98 dB SPL, and 86 dB SPL, respectively.

+8 dB gain re 94 dB on a small listening room and +4 dB gain in a really big listening room (=small auditorium)
they seem to be quite sure about this and thus they conclude:

Thus it seems unwise to sing loudly in a small room

the source of discrepancy between their statements and the results of measurements by Dr. Toole and by Salas is difference in RT60 of the rooms
the volume factor seems to be much less important
(in case of measurements by Dr. Geddes the difference can be mainly attributed to unusually high directivity index of His speakers)
the people at KHT assume specific RT60 for their three model rooms:

Typical values of the reverberation time to aim at in the design are 0.5, 0.8 and 1.5 s, respectively for our three rooms

0.5 for a small room and 0.8 for a big room is probably much more than in case of a typical listening room, not to mention a dedicated audiophile listening room
these rooms are typically much more absorptive
I read that RT60 of 0.2 is typical for a furnished living room
I suppose that in an audiophile listening room typical RT60 may be even lower, perhaps even below 0.1
for a medium sized listening room of volume of around 100 m3 change in RT60 from 0.5 to 0.08 means a change in reverberation distance from 0.6 m to 4.2 m - very significant, and at RT60 of 0.05 it would be 6.5 m

the "SPL target" issue seems to be a part of "speaker-room interaction" issue
of "loudspeakers and room as a system"

best,
graaf

 

[gedlee]

I think that the complexity of your comments is exactly the reason that I ignore this level of detail and did not even mention it when I first posted. If you want to go through this analysis for a specific room you might be able to change the values by as much as 10 dB, but if you want to give numbers that are likely to be valid in maost rooms then ignoring the details makes the most sense. Which leaves us back to my original numbers as reasonable - unless you want to do the detailed calculations.

Most people working in small rooms believe that RT60 is not a useful measure in such rooms. When its approaches the integration time of our hearing (20 - 50 ms) it becomes completely meaningless.

 

[graaf]

Re: Re: Re: SPL targets for speaker design

In a living room you will never get +7dB

I have responded:

the people at KHT give an example of bass singer that "produces" 94 dB at 1 m and this what they found :

+8 dB gain re 94 dB on a small listening room and +4 dB gain in a really big listening room (=small auditorium)
they seem to be quite sure about this

(...)

the source of discrepancy between their statements and the results of measurements by Dr. Toole and by Salas is difference in RT60 of the rooms

the people at KHT assume specific RT60 for their three model rooms:

0.5 for a small room and 0.8 for a big room is probably much more than in case of a typical listening room, not to mention a dedicated audiophile listening room
these rooms are typically much more absorptive
I read that RT60 of 0.2 is typical for a furnished living room
I suppose that in an audiophile listening room typical RT60 may be even lower, perhaps even below 0.1
for a medium sized listening room of volume of around 100 m3 change in RT60 from 0.5 to 0.08 means a change in reverberation distance from 0.6 m to 4.2 m - very significant, and at RT60 of 0.05 it would be 6.5 m

the "SPL target" issue seems to be a part of "speaker-room interaction" issue
of "loudspeakers and room as a system"

now I have found this: http://www.vividaudio.com/downloads/Acoustics_an_introduction.pdf

there Laurence Dickie, designer of the original Nautilus of B&W, asserts that:

Reverberation can increase sound levels within a room by up to 15 dBA

and that "optimum midfrequency RT60" for a "broadcast studio" is 0.5, that is the same as recommended for small rooms by KHT handbook quoted above

this seems to confirm my statement that to have "110 dB max peak headroom at 3m" we only need about 100 dB "per stereo speaker capability"

comments welcome

best,
graaf

 

[Salas]

Re: Re: Re: Re: SPL targets for speaker design

this seems to confirm my statement that to have "110 dB max peak headroom at 3m" we only need about 100 dB "per stereo speaker capability"

comments welcome

best,
graaf

Its not my practical experience. If you get such a level bonus in your room, you are quite well.

 

[graaf]

Hello again!

I have found an interesting statement in the brochure of 1973 JBL L200 Studio Master loudspeaker:

75-80 dB is a comfortable listening level

it was a note to the information concerning the efficiency of the L200 (84 dB/1W @15 feet)

see:
http://www.lansingheritage.org/html/jbl/specs/home-speakers/1973-l200.htm
at page 4

I have found the statement in other JBL product information brochures from that period as well
so this was a kind of their "official standpoint" with regard to "SPL target for home listening"

I wonder what is their standpoint today?

best,
graaf

 

[graaf]

I think that Dr Geddes knows JBL and their products very well
so perhaps He can answer?

 

[AndrewT]

I'm going to take a guess.
75dB @ the seating position sounds like comfortable.
+10dB sounds like it might be loud.
+20dB sounds like it might be too loud verging on the uncomfortable.

Now add 20dB of peak transient ratio to those average levels and we end up with guesstimates of:
95dB = comfortable peak
105dB = loud peak
115dB = too loud towards uncomfortable peak.

Does this sound familiar or logical?

 

[gedlee]

I think that Dr Geddes knows JBL and their products very well
so perhaps He can answer?

First, you can bet that those numbers are dBA, add 10 dB or so for dBC, which is what all of my data is in. Then add 10-20 for "louder" levels and another 10-20 for peaks, so 30-50 dB onto those numbers seems reasonable and we are right back to what I said in the begining.

 

[graaf]

I'm going to take a guess.
75dB @ the seating position sounds like comfortable.
+10dB sounds like it might be loud.
+20dB sounds like it might be too loud verging on the uncomfortable.

Now add 20dB of peak transient ratio to those average levels and we end up with guesstimates of:
95dB = comfortable peak
105dB = loud peak
115dB = too loud towards uncomfortable peak.

Does this sound familiar or logical?

this statement in JBL L200 brochure is a note to the information concerning efficiency
it is supposed to be a part of an answer to the customer question - "how loud can it play? will it be loud enough?"
therefore we can assume that the word "comfortable" is used in the brochure to mean "comfortably loud", "loud enough", "You won't be disappointed"

from that perspective I think that distinction You have made between "comfortable peak" "loud peak" and "too loud towards uncomfortable peak" misses the point of the JBL brochure statement
"comfortable peak" means there "a peak loud enough but not too loud"
louder than "comfortable" in this sense means just "uncomfortable"

Does this sound logical?

best,
graaf

 

[graaf]

First, you can bet that those numbers are dBA, add 10 dB or so for dBC, which is what all of my data is in. Then add 10-20 for "louder" levels and another 10-20 for peaks, so 30-50 dB onto those numbers seems reasonable and we are right back to what I said in the begining.

why dbA?
this statement in JBL L200 brochure is a note to the information concerning efficiency (as an explanation to the customer, as an answer to the question "is it loud enough"?)

is loudspeaker efficiency given in dBA?

why louder?
louder than "comfortable" in the sense used in the brochure ("loud enough") is just "uncomfortable", "too loud"
why should anybody listen uncomfortably loud?

I agree that we can add up to 20 dB for peaks. This is also taken into account in the brochure (under "power/amplifier requirements")
80 dB average, 100 dB peaks
this is in agreement with my own measurements and also with NRC statement that 90 dB average is "far beyond normal listening levels"
assuming that normal=typically comfortable

best,
graaf

 

[gedlee]

80 dB average, 100 dB peaks
this is in agreement with my own measurements and also with NRC statement that 90 dB average is "far beyond normal listening levels"
assuming that normal=typically comfortable

best,
graaf

But its not in agreement with my measurements in my room. Sorry.

 

[graaf]

from "Acoustics and Psychoacoustics" by Howard and Angus, 4th ed., Focal Press 2009, Chapter 6: "Hearing Music in Different Environments"

(an average absorption coefficient of 0.2 is typical for a typical domestic listening room and it corresponds to a RT60 of around 0.4 s)

it follows that the SPL produced by a loudspeaker in a typical domestic listening room at any normal listening distance is 4 dB higher than it's measured SPL output at 1 meter distance

 

[bear]

The addition of the "hanging around" reverberant field?

I do not know but do these coefficients correspond to a rather "live" room?? Seems likely... like a tiled bathroom, very reverberant?

_-_-bear

 

[graaf]

The addition of the "hanging around" reverberant field?

I do not know but do these coefficients correspond to a rather "live" room?? Seems likely... like a tiled bathroom, very reverberant?

_-_-bear

definitely a typical living room according to Professors Howard and Angus

actually their calculations are much more conservative than those of the authors of the handbook from the Swedish Royal Institute of Technology linked before in this thread here

 

[Hentai]

This thread grew pretty rapidly from last time i looked at it. Not sure if there is room still at the discussions table but i would like to add some thoughts to this as i have been enjoying high spl speakers for a few years now.

For me a SPL value on it's own means close to nothing. Sensitivity value tells a little more if its related to a driver and not a system. For a system you can reach a pretty high sensitivity if you use many low sensitivity units but that kind of system never gave the feeling of a real high sensitivity system.
By this i mean high efficiency and it is my belief that this is what actually matters. High efficiency of the drivers and (always preferred but not as important) high efficiency system.
So high efficiency drivers in a horn is from my point of view the best way to achieve the best impact. I have met however designs with high efficiency drivers but the system had a lower overall sensitivity (mostly by smoothing things passively) but still keeping close enough characteristics of great transients, great dynamics and impact.