Sound Quality Vs. Measurements

Status
Not open for further replies.
It seems reasonable to me to assume that a driver, sold as capable of accepting 150 WRMS, or (since it's a nominally 8 Ohm driver) 34.6 Vrms. Let us, for the moment, assume that this is indeed so.

Does it then not stand to reason that this driver, if properly implemented, should feel little or no compressions at 20 Vrms, i.e. the point where those small speakers started to go nuts?

My point is this - if we generally think of a nominally 100 WRMS amp as simply a "safe" 50 WRNS amp, since we have to allow for transients and do not want any clipping thank you, why should be think of speakers completely out of that idea?

I agree that giving a loudspeaker in general any truly meaningful power rating is one hell of a fiddly job, but by and large, it should be safe to assume that it can handle about half the rated power without undue distress.

Ultimately, in a home enivornment, we all rarely use large power levels, we are typically well below 5 watts conitnuous, which could become much more in transients, and people erroniously think that a transient requiring say 40 watts to be properly (i.e. as the producer intended it to be) reproduced will sound much louder. This is simply not so, but if done properly, your tympani will maintain its sonic balance as it should, simply using much more power to be able to do so.

No doubt the voice coils will heat up after an hour of higly dynamic music, but let's give some credit to the designers of that driver, these people are neither complete fools, nor professional thieves, and let's assume they do know what they are on about.

In the end, I think much more than is usually assumed falls squarely on the field of the customer; I find that far too many times, they don't really know what they want.
 
Hi,

From Sound On Sound:
"The final input level of 20V RMS is pretty loud, but it was important to drive the speakers hard in order to reveal any differences without any ambiguity."

Thorsten, do your 300Bs reach 20V RMS (without clipping) ?

One must beware of single numbers in comparison when comparing apples and oranges.

The speakers I used with my 300B Amplifiers where 97dB/1W/1m efficient (and 16 Ohm). So 20V was not required to reach high SPL's hence 20V output capability was not a design requirement. It may be if I was using lower efficiency speakers.

As the Speakers in the Sound On Sound test varied between 86 and 88.8dB/2.83V/1m, we can take a mean of 87dB vs. 97dB or 10dB difference in SPL with level for ease of calculation.

In other words, my 300B Amplifier only needed to produce 6.3V (2.5WW) for the same SPL that 20V in the Sound On Sound test would produce, except the 2.5W in this case are much less likely to cause significant compression compared to 50...100W

Also, you may be surprised by the actual levels such a simple 300B SE Amplifier is capable of. You may wish to peruse the article "Measurements of Amplifier Peak Output Voltages Under Dynamic Conditions and Into a Real Loudspeaker Load" - by Peter van Willenswaard.

Part 1

Part 2

BTW, my current tube amp manage around 18V without gross distortion, my current speakers are around 89...90dB/2.83V/1m and a near resistive 6 Ohm Load.

Ciao T
 
Hi,

Perhaps the word "domestic" is unfamiliar? So, any contrary data to share?

It is not unfamiliar, it is merely a weaselword that allows the user to claim anything they like, without the ability to falsify any such claim as there is no definition.

Instead I would propose we apply the only known standard for Home Audio (domestic audio) THX.

This specifies 85dB SPL PER CHANNEL at the listening position for a -18dBFS signal. It implies a maximum SPL at the listening position of 103dB SPL for "domestic levels".

If we consider a common listening distance of 3m it means we require around 110db/1m from each of the Speakers. For a 87dB/2.83V/1m sensitive Speaker (which according to John Atkinson's work represents an "average" sensitivity for HiFi Speakers) this means +23dBW Power input, or 40V or 200W per speaker.

If we use instead a 97dB/2.83V sensitive speaker the same condition is fulfilled using only +13dBW or 12.7V or 20 Watt.

We may now debate how much compression and distortion we wish to allow under these conditions as "low impairment" and this can be a mega meta thread in itself.

However no matter how we look at it, the numbers are not going look good next to any semi-competently designed Amplifier. But what are these numbers?

Soundstage measures at 2m distance, usually at 90, 95 and in extreme cases 100dB, single speaker for compression and distortion, note their distortion plots cut off -45dB, so this is only around 0.5% HD... So my previously noted 110dB SPL would be 104dB SPL in the measurements that Soundstage does

Let's see what we have, shall we?

Dynaudio Focus 110

Nearly no (< 0.5dB) compression at 90dB from 70dB, worsens a bit at 95dB to -2dB but only below around 300Hz, mid/hi unaffected. Distortion is a mixed bad, fairly low from 500Hz-4Khz, rises a lot at either end though. Considering that is is basically a "toy" speaker with a 5.5" Woofer not a bad results.

SoundStage! Measurements - Dynaudio Focus 110 Loudspeakers (12/2006)

Magico V2

A -1dB broadband response variation at bass and treble variation between 70dB and 90dB... Distortion is very low.

SoundStage! Measurements - Magico V2 Loudspeakers (5/2009)

NHT Classic Three

A -1dB broadband response variation at bass and treble variation between 70dB and 90dB, but gets notably worse at 95dB with > -2dB... Distortion is comparably low.

SoundStage! Measurements - NHT Classic Three Loudspeakers (3/2007)

Paradigm Reference Signature S1 v.2

A -1dB broadband response variation between 70dB and 90dB, but gets much worse at 95dB with +1/-3dB... Distortion is comparably low but rises at high and low frequencies.

SoundStage! Measurements - Paradigm Reference Signature S1 v.2 Loudspeakers (12/2007)

Paradigm Studio 60

A -1dB broadband -2dB narrowband frequency response variation between 70dB and 90dB, but only gets a little worse at 95dB... Distortion is bad.

SoundStage! Measurements - Paradigm Reference Studio 60 v.5 Loudspeakers (5/2009)

Revel Concerta F12

This is a low compression Speaker! A <<-1dB frequency response variation between 70dB and 90dB, ~-1dB frequency response variation between 70dB and 95dB, even at 100dB frequency variation to 70dB is only around 2dB... Distortion is reasonably low, excepting a nasty spike at around 3khz, could be acoustically bothersome.

SoundStage! Measurements - Revel Concerta F12 Loudspeakers (3/2006)

Usher Audio Technology Be-718

A <-1dB frequency response variation between 70dB and 90dB, quite broadband with bass and upper mid/treble being turned down, no measurement at 95dB... Distortion is reasonably low.

SoundStage! Measurements - Usher Audio Technology Be-718 "Tiny Dancer" Loudspeakers (10/2007)

Wilson Audio Watt/Puppy 8

A <-1dB frequency response variation between 70dB and 90dB, but looses it to ~-2dB frequency response variation between 70dB and 95dB, saving grace is that this is quite narrowband around 5KHz... Distortion is reasonably low.

SoundStage! Measurements - Wilson Audio Specialties WATT/Puppy 8 Loudspeakers (3/2007)

YG-Acoustics Anat

A +/-1dB frequency response variation between 70dB and 90dB, +1/-2dB frequency response variation between 70dB and 95dB... The less said about distortion the better.

SoundStage! Measurements - YG Acoustics Anat Reference Main Module Loudspeakers (4/2007)

QED

Ciao T
 
"maximum SPL at the listening position of 103dB SPL for "domestic levels".

If we consider a common listening distance of 3m it means we require around 110db/1m from each of the Speakers. "

How did you calculate that ? Would this not depend on the room size ? Or are you listenining in free space ?
 
Hi,

"97dB/1W/1m efficient" quite different from a LS-3/5 you proposed as an "example" no ?.
As always you just use whatever seems to support your position, this makes a discussion somewhat tiresome.

Anyone who would use an LS-3/5 with a 300B SE Amp would be in a very questionable position indeed.

What I am giving are data-points, not what supports any "position".

Lets look:

System 1 - 300B-SE 6W & Tannoy Monitor Red - 10W/108dB/1m sans Compressison

System 2 - 50W (20V) & 87dB "Monitor" - 50W/104dB/1m sans Compressison

System 3 - 100W (30V) & LS-3/5 - 100W/103dB/1m sans Compression

If compression is approximately linear to applied power we can see which system will be subject to more compression...

We know that the "Monitors" had between around 2 - 5dB Compression, as the SE Amp High Efficiency Speaker system operates at 1/10 of the power of the Monitor (and is actually blessed with a much larger voice coil and most of the Monitors) it is safe to consider that thermal compression will not be measureable at SPL's below 100dB.

The same cannot be said about the smaller systems.

Incidentally, I also happen to have distortion measurements for the TMR 15", this driver that went out of production in 1967. It is below -46dB (0.5%) in the midrange at 100dB/1m (that is 1 Watt power input) and stays below 1% all across the band.

In fact, as the distortion is H2 dominant it can be lowered considerably by using an amplifier with appx. 1% HD at 1 Watt H2 dominant with the correct phasing of distortion components. Usable bandwidth in my system was < 30Hz-18KHz.

So, using 1930's Amplifier Technology and 1940's Speaker technology we can easily make a system that is low compression, low distortion and wide bandwidth, something that seems to present considerable problem to 21st Century technology.

Amusing, innit?

Ciao T
 
Thermal compression is real ...

No-one argues that it isn't real. But is it a significant factor in domestic audio (assuming one is aware of whether one is in a living room or a studio)? Howard's data, on a sample that would tend to fall on the worse side, indicate no. If there's contrary data, as opposed to reflexive dismissal, I'd be interested to know.
 
Hi,

"maximum SPL at the listening position of 103dB SPL for "domestic levels".

If we consider a common listening distance of 3m it means we require around 110db/1m from each of the Speakers. "

How did you calculate that ? Would this not depend on the room size ? Or are you listenining in free space ?

The 103dB are obvious, 85dB at -18dBfs = 103dB.

The 7dB increase in SPL for a 3m listening distance is an estimate and will vary considerably with Speaker directivity (or lack thereoff) speaker placement, room size and room furnishing.

It is a ballpark number, nothing more. If you insist it should be 107dB or 113dB I will not quibble.

Ciao T
 
Hi,

Yes really ? From your post #2998581: "Three to four dB thermal compression is nothing unusual ..."

This is in the standard way of specifying thermal compression in Pro Audio, that is at rated continuous maximum power.

That said, I probably would not want to run many HiFi Speakers at rated power for any amount of time, in most cases it appars that the makers are very optimistic.

FWIW, you can take the AMR LS-77 and sweep it with 150W RMS (that is it's rated input power) from 26Hz to 40KHz (if your amp plays along) with no ill effects on the drivers or other parts.

Things like this I simply expect. When I see what most "HiFi" Speakers do I'm bemused.

I know quite a few 3-Way speakers where doing the same thing would leave you with three blown drivers and a damaged crossover.

Ciao T

PS, FWIW, I can clip a 180W RMS Amp into the LS-77 (87dB/W/m our rating - 88dB according to Paul Messengers measurements) in domestic settings without it being EXTREMELY LOUD (but loud, yes), but it does require quite dynamic Music (not much compression). That should be 109dB/1m but despite quite extreme drivers in terms of thermal design we also have some thermal compression at these levels, but quite mild.
 
Hi,

No-one argues that it isn't real. But is it a significant factor in domestic audio (assuming one is aware of whether one is in a living room or a studio)?

Well, okay, Thermal compression is real, can be measured at quite low SPL's.

Lets add to this, for those who want to understand the issues better, may I advise "Dynamic Linearity and PowerCompression in Moving-Coil Loudspeakers" by Mark R. Gander - JAES, Vol.34,N o. 9, September 1986.

As to "Is one aware of whether one is in a living room or a studio", precisely why would this matter and why should it make it a difference? Do the laws of Physics change because we step outside the studio? Hardly.

As to does it matter for music listening or not, let each draw their own conclusions. One thing that can help immensely is to listen to both an extreme example of compression (LS-3/5 & 100W Amp) and one of extremely low compression (say Avant Garde Uno) with the same music, something large scale classical, not compressed and at realistic levels.

It is quite instructive.

Howard's data, on a sample that would tend to fall on the worse side, indicate no. If there's contrary data, as opposed to reflexive dismissal, I'd be interested to know.

Let us be clear.

First, KH's test is a single Data Point. If it falls on the worse side or not would have to be established from a much larger sample.

Second, looking at the glitches his test-setup produces makes me wonder just how well it was working at measuring what it was intended to measure.

Third, the test lacks positive and negative controls being documented and hence we cannot presume that it was working as intended or with sufficient accuracy.

As such even the single datapoint that was produced in the experiment is not relevant as data.

My dismissal is not reflexive, but evidence (or rather lack of relevant evidence on the part of the writer) based.

Now, having dismissed the experiment as flawed, and it's results as irelevant, I would add that to declare a nearly 0.7dB change in Efficiency in one driver (LF) that is not matched by second driver (HF) can certainly lead to an audible change in balance of the speaker.

Moreover, the LF tuning will be altered as well.

Even though I consider that KH's experiment UNDERSTATED the issues considerably, even the results it produced do not justify his closing remarks, but rather the opposite.

Ciao T
 
Hi,

So the answer is still no, you have no data contrary to Howard's.

We should note that Soundstages ACOUSTIC measurements show much larger changes than his 0.67dB for lower SPL's than he claims to have operated in his test.

So I would say there is better data than his, more relevant data. Is it contrary to his? No, it merely suggests that his experiment understand the effects present, for whatever reasons.

With a single Data point against well over a dozen (I only showed a few) I think the conclusions are easy to draw for anyone who has a working sense of logic and no agenda, so I'll leave it everyone to draw their conclusion from the ACTUAL EVIDENCE.

Ciao T

Pour me another Purple Haze lassie, I need my sedatives...
 
Sy,

Use and application are different.

Please present actual evidence that listening to a cut in a mastering studio or in a living are different uses and applications.

The reverse in fact should be true, the Mastering Engineer in the Studio should be hearing what you will be hearing in your living room, in order for the Mix to translate well from his studio to your living room.

Ciao T
 
Thermal compression is real and my worry is more that it wacks passive xovers out of alignment. Good reason to filter actively.

Thermal compression is real , but i have never been able to measure such in any one of our products.

Typically the speaker system is ran at an input of 10 watts for 15 mins , then sweeps are done at 1 watt,5 ,watt, 10 watt, 25 watt and compared, when overlapped you get one line .

Speaker is 89db/2.83v/m / 8 ohm Nom...Zmin 7 ohm


There is no whacking of xover out of line, now that is not to say a poorly designed xover nor speaker system would not exhibit such ..
 
diyAudio Member RIP
Joined 2005
There have been proponents of thermal compression issues at Harman for years, in addition to Gander et al. in JBL Pro. But I suspect many of their concerns have been anecdotal.

Greg Timbers likes horns for all but bass, and I heard his home system with the mid-high powered by Quicksilver tube amps. Sure played loud! I should have brought some material I was familiar with, as virtually everything he had was female vocals. I finally dug out Bela Fleck's Flight of the Cosmic Hippo, and the title track was indeed most impressive --- no difficulty with bass extension. I think I found one sort-of-pops-classical recording. Different strokes.
 
Status
Not open for further replies.