# Grimani SPL rule of thumb

#### robpbg

##### Member
I could, but I find using metric length units reduces the dynamics, staging, and rhythm. This is because of the simple integer ratios between the driver size and room dimensions.

You should be careful! Some will take you seriously.

OP
D

#### Dzhaughn

##### Member
The CTA-2034 has a section on specifying max SPL. With good stuff.

It is good, thanks for pointing it out. It is also interestingly different.

CTA-2034 Table 6 suggests only adding 1 dB per 3000 sq ft, where Grimani adds 3 dB. (Grimani's, I think he suggests, has an implicit model for how the distance from the listening position to the speakers varies as the room gets larger; maybe that is part of the difference.)

For all the doubters of the possibility of this kind of estimate: Table 6 also suggests all the wall materials and furniture only make +/- 1 dB difference in overall SPL. Do you believe that? (EDIT This is incorrect on my part, see below, ht NTK)

Last edited:

#### NTK

##### Major Contributor
Forum Donor
It is good, thanks for pointing it out. It is also interestingly different.

CTA-2034 Table 6 suggests only adding 1 dB per 3000 sq ft, where Grimani adds 3 dB. (Grimani's, I think he suggests, has an implicit model for how the distance from the listening position to the speakers varies as the room gets larger; maybe that is part of the difference.)

For all the doubters of the possibility of this kind of estimate: Table 6 also suggests all the wall materials and furniture only make +/- 1 dB difference in overall SPL. Do you believe that?
You misinterpret the instructions. The change is in "Listening Levels". "-1" is adjust one level up, e.g. from "Quiet" to "Moderate', etc.

#### nerdoldnerdith

##### Active Member
Even with using -7dB as a conservative estimate for my speakers placed 10 feet away in my ~3000 sq. ft. room, this tells me I need 105-(86-7)=26dB=400W into 8 ohms. In my case it needs to hold that down 3.7 ohms, so it can't start drooping as the impedence decreases. I thought maybe a Hypex NC1200 could pull that off, but it only manages about 600W into 4 ohms. I had to go with a Hypex NC2000 with a ridiculous 1600W into 8ohms and 2500W into 4 ohms to drive my speakers at reference levels without clipping the damn amps!

#### echopraxia

##### Major Contributor
Forum Donor
I don’t understand the point of these SPL “rule of thumb” measurements when we can’t even determine how loud we can safely physically drive most audiophile speakers.

Even Revel/Harman for example is particularly bad here. For all their marketing about how they are science-ing so much science into their speakers, none of their speaker specs seem to list a true power handling rating or max SPL chart.

Unlike e.g. Neumann and Genelec which will actually quote max usable SPL, from which you can then estimate your in-room results with rules of thumb. Neumann is best by far, where they cite max sustained and peak SPL for higher and low frequencies, and include full SPL vs frequency charts.

Last edited:

#### nerdoldnerdith

##### Active Member
I don’t understand the point of these SPL “rule of thumb” measurements when we can’t even determine how loud we can safely physically drive most audiophile speakers.

Even Revel/Harman for example is particularly bad here. For all their marketing about how they are science-ing so much science into their speakers, none of their speaker specs seem to list a true power handling rating or max SPL chart.

Unlike e.g. Neumann and Genelec which will actually quote max usable SPL, from which you can then estimate your in-room results with rules of thumb.
Their PerformaBe and Performa3 speakers have recommended amplifier power. Based on those recommendations the PerformaBe speakers should be able to hit 115dB, which is where they need to be for home theater.

#### echopraxia

##### Major Contributor
Forum Donor
Their PerformaBe and Performa3 speakers have recommended amplifier power. Based on those recommendations the PerformaBe speakers should be able to hit 115dB, which is where they need to be for home theater.
The problem is there is absolutely no clarification from Revel (perhaps intentionally) about what their “recommended amplifier power“ actually means, how it was derived, or how it relates (if at all) to the speaker’s abilities. Rather we are left to make our own dangerous assumptions when trying to estimate the speaker’s capabilities, in a domain where faulty assumptions can be catastrophic.

Sure, we can calculate the estimated in-room SPL at 300 watts, but they didn’t say they could sustain 300 watts safely. And 300 watts at what frequency, 30hz? 100hz? 1000hz? 10khz?

Given how vague the specs are on what “recommended amplifier power” means, I don’t know that any of us can really understand the safe SPL range of these without more data.

In contrast, Neumann and Genelec has data available for what SPL the speakers will safely and sustainably achieve at each frequency. Here is an example of the Genelec 8361A:

Here we can see it can achieves at 1 meter anechoic >120db at 300hz, ~114db at 80hz, and ~105db at 30hz, with <10% distortion.

Is this speaker therefore capable of 120db, or 105db? Well, both, obviously, depending on the frequency. There are many ways we can try to boil this down into one number, but any such simplification is bound to lose information versus a full plot like this.

Even so, it is completely fine to boil down a multidimensional measurement like this to just a single SPL “summary” rating value; but, for this to be meaningful at all, the “summary” rating’s definition/derivation must be unambiguously specified in a 100% reproducible, standardizable way.

In Revel’s case, it really can’t get any worse or more hand-wavy with terms like “recommended amplifier power” which really (probably intentionally for liability?) makes almost no promises about actual SPL capabilities. Not that Revel is worse than most other audiophile speaker brands in this respect, but they’re certainly not better.

And we know they can do better, if they wanted to, to compete with Genelec, Neumann, etc. They probably don’t bother because they don’t see Genelec/Neumann as competitors (pro audio vs home audio), but thanks in part to ASR, I think a lot more audiophiles are seeing the merits and advantages of some pro audio products over the vague and nebulously specified fashion products the audiophile speakers world is notorious for.

Last edited:
OP
D

#### Dzhaughn

##### Member
I don’t understand the point of these SPL “rule of thumb” measurements when we can’t even determine how loud we can safely physically drive most audiophile speakers.

Grimani actually offers some interesting ideas/claims about that in the video. In particular, he claims that, generally, as the power goes up, the frequency response curve starts to change. A 3dB change in that curve's shape is what he'd use as a max power rating for passive speakers. Not only does that reflect a change in performance, but he claims this level correlates with decreases in component life.

It would be not so hard to measure this. If you dare.

#### Sancus

##### Major Contributor
Forum Donor
Grimani actually offers some interesting ideas/claims about that in the video. In particular, he claims that, generally, as the power goes up, the frequency response curve starts to change. A 3dB change in that curve's shape is what he'd use as a max power rating for passive speakers. Not only does that reflect a change in performance, but he claims this level correlates with decreases in component life.

Properly measuring this sort of compression without a Klippel or taking your speaker on an outdoor adventure is pretty much impossible even if you are willing to risk damage, so that 3dB threshold is useless to the end user.

Erin used to do this test in his reviews and stopped because of how time consuming and annoying it was.

OP
D

#### Dzhaughn

##### Member
Well, tonight, in a 2400 ft^3 living room with lots of books and furniture, a single 85 db sensitivity speaker (measured by ASR) and a 50W Rotel amplifier, my Umik-1 measured 93 dB ) in the seating positions on several tracks, and 96 db on the first few bars of DM's "A pain that I'm used to."

Grimani's formula says 95 dB in a 20% larger room.

(Grimani would also say I need 106 dB, if I understand him. I used earplugs for this.)

#### aac

##### Active Member
And we know they can do better, if they wanted to, to compete with Genelec, Neumann, etc.
They compete with them in studios, but jbl pro isn't any better there. Not even a fr chart for their monitor line.

OP
D

#### Dzhaughn

##### Member
Properly measuring this sort of compression without a Klippel or taking your speaker on an outdoor adventure is pretty much impossible even if you are willing to risk damage, so that 3dB threshold is useless to the end user.

I don't get the difficulty. In particular, there's no need to go outdoors, no need to measure at angles, and no need to take a full minute. Do an REW sweep, and if the curve shape is the same +/- 3dB, increase the volume and repeat.

I expect to try it this week on an old infinity speaker.

#### Sancus

##### Major Contributor
Forum Donor
I don't get the difficulty. In particular, there's no need to go outdoors, no need to measure at angles, and no need to take a full minute. Do an REW sweep, and if the curve shape is the same +/- 3dB, increase the volume and repeat.

I mean if you don't care about accuracy at all, sure, whatever. That measurement will be full of reflections so it will be useless.

OP
D

#### Dzhaughn

##### Member
I mean if you don't care about accuracy at all, sure, whatever. That measurement will be full of reflections so it will be useless.

I think you're trying to measure too much at once! The Spinorama is one thing. The stress test a separate thing. We'd only look for the level that produces a change in the curve in the stress test, not at the curve itself.

#### Sancus

##### Major Contributor
Forum Donor
I think you're trying to measure too much at once! The Spinorama is one thing. The stress test a separate thing. We'd only look for the level that produces a change in the curve in the stress test, not curve itself.

I didn't say anything about spinoramas. Unless you gate the reflections out, they will be present, and will corrupt any measurement including an SPL measurement. And gating won't really help much because SPL capability only matters from 20hz-3khz or so, above that it's pretty much irrelevant.

So doing that will only be meaningful for one point and one speaker position -- it won't tell you anything general about the speaker.

If that's all you want, then OK I guess, but there's not really any point to measuring compression in that case IMO. You'd be better off just using your ears, pink noise, and an SPL meter and noting when the speaker starts to sound bad. Which will likely be long before the 3dB compression threshold anyways.

OP
D

#### Dzhaughn

##### Member
I didn't say anything about spinoramas.

Ah OK. I am assuming ASR has done a full review of the speaker in question. The only question in the present context is what is the upper bound for the power that a speaker will handle and sound as good (or bad) as usual. Surely there is a better answer than "trust the marketing material."

#### hardisj

##### Major Contributor
Reviewer
Properly measuring this sort of compression without a Klippel or taking your speaker on an outdoor adventure is pretty much impossible even if you are willing to risk damage, so that 3dB threshold is useless to the end user.

Erin used to do this test in his reviews and stopped because of how time consuming and annoying it was.

I perform compression/linearity testing. I just provide it in a easier to understand form now.

#### maverickronin

##### Major Contributor
Forum Donor
In contrast, Neumann and Genelec has data available for what SPL the speakers will safely and sustainably achieve at each frequency. Here is an example of the Genelec 8361A:

That max SPL vs frequency chart is actually from Sound and Recording's review.

#### Sancus

##### Major Contributor
Forum Donor
I perform compression/linearity testing. I just provide it in a easier to understand form now.

Oh, I know, I was referring specifically to the 3dB threshold test which you don't do anymore -- just sweeps at fixed levels. Not that I'm complaining, I never really liked the 3dB test, since it seemed almost guaranteed to trigger first on the tweeter for larger speakers. Real content is almost always -10dB (or much more) from 100hz by the time it hits tweeter frequencies, so it doesn't really make sense to me to put tweeters through >105dB tests.

Last edited:

Replies
0
Views
632
Replies
60
Views
14K
Replies
161
Views
28K
Replies
79
Views
5K
Replies
4
Views
797