Sound Quality Vs. Measurements

[Wavebourn]

Originally Posted by bcarso

I'll have to find out who did that.

No powercube is needed to see the whole picture.

[ThorstenL]

Hi,

Originally Posted by wahab

I will do a test dedicated to even harmonics to have
some clues about it...

Given that there is a large body of work on the subject going back to the 60's and before, do we really need to re-do it, your call though?

Originally Posted by wahab

As for my speaker s HD content , i dont know since i ve no accurate enough microphone to perform a valuable measurement.

Looking at who makes them and what the drivers look like (look like french made or PRC OEM for that French harman shop), I suspect especially at 1KHz (where the woofer still handles the signal) there would be a lot of 3rd HD, even at 80dB.

If I guess the driver maker right you will have >0.6% 3rd HD at 85dB/1m (measured by K+T) and around 0.2% 2nd HD. At around 80dB due to the cubic law of 3rd HD may be down to around 0.1% as well as the 2nd Hd. Higher order stuff is mercifully low though.

So some of what you hear may in fact the result of combining a high distortion speaker with a high distortion amp. You may wish to crosscheck with inverted speaker polarity.

I think using a pair of headphones with known distortion performance AND with low distortion may be a better way.

Ciao T

[dvv]

If memory serves, "medium SPL levels" in a room would be in the range of 85...90 dB, ref. 1 m, both speakers active.

The actual value will vary due to possible room sound treatment, furnishings and ambient noise.

According to some not very good SPL meters, it seems I use my speakers at around 87...88 dB SPL re 1 m. That would be my medium volume, but then, my room is small by any measure, just 12 m2, or about 130 ft2.

[dvv]

As for HK, I told you - load tolerance is their very strong point.

My 680, nominally 85/130W into 8/4 Ohms, whacked me with a peak power burst of 512 Watts into 2 Ohms. Since I don't spend my time listening to sine waves at full rated power, I reckon that's quite enough for most people, with most speakers, in most rooms.

I guess their HCC (High intstantaneous Current Capability) is not just an ad departement catchphrase, rather a piece of Otala legacy.

[ThorstenL]

Hi,

Originally Posted by dvv

If memory serves, "medium SPL levels" in a room would be in the range of 85...90 dB, ref. 1 m, both speakers active.

Recommendations for listening levels vary somewhat. In germany for studio's we often see 94dB SPL for a -14dBFS (or 0dB VU). That means a "loud" but not excessively compressed CD will generate 94dB average SPL during the loudest passages, giving 111dB Peaks at 0dBFS.

Incidentally, the 94dB match well with SPL measurements I undertook at the Royal Festival Hall for an orchestra at ffff while playing "big" music (Mahler, Mussosky/Ravel etc.) at around 6th row centre.

I have seen similar practices of level (give or take a few dB) elsewhere, THX recommends 108dB peaks for 0dBFS (85dB plus 20dB headroom and 3dB correction for peaks).

Sadly many commercial music recording studio's do not adhere to such a convention, which usually makes their mixes less portable than ones from studio's that stick close to "best practice" for monitoring levels.

If listening levels are much lower than these references we would have to apply physiological frequency response correction, that is some bass boost. This boost is often build permanently into "audiophile" speakers under the misguided concept of "baffle step correction". Such speakers sound overly bass heavy (I usually call the bass produced "Stunt Bass") when played at realistic/reference levels (assuming they can reach them without gross distortion). But that is a whole other kettle of fish.

Ciao T

[GloBug]

Originally Posted by ThorstenL

If listening levels are much lower than these references we would have to apply physiological frequency response correction, that is some bass boost.

Shhh, don't tell a.wayne, he will say your amplifier is no good if it's not "flat".
He prefers a linear solution to reproduce non-linearity.

[wahab]

Originally Posted by ThorstenL

Hi,

Given that there is a large body of work on the subject going back to the 60's and before, do we really need to re-do it, your call though?

Looking at who makes them and what the drivers look like (look like french made or PRC OEM for that French harman shop), I suspect especially at 1KHz (where the woofer still handles the signal) there would be a lot of 3rd HD, even at 80dB.

If I guess the driver maker right you will have >0.6% 3rd HD at 85dB/1m (measured by K+T) and around 0.2% 2nd HD. At around 80dB due to the cubic law of 3rd HD may be down to around 0.1% as well as the 2nd Hd. Higher order stuff is mercifully low though.

So some of what you hear may in fact the result of combining a high distortion speaker with a high distortion amp. You may wish to crosscheck with inverted speaker polarity.

I think using a pair of headphones with known distortion performance AND with low distortion may be a better way.

Ciao T

That such test were already done doesnt make them useless ,
quite the contrary.

As for the speakers , right that the fundamental is rendered
by the woofer , but H2 fall on the crossover point while H3 and above
are in the range of the tweeter.

Dont know for the woofer , seems its marking was removed ,
i ll check later , but it s surely not a low cost OEM thingy,
while the tweeter is a Morel MDT30.

As you point it , increasing the amp output level
will increase its THD as well as the one of the speakers ,
but then , theses latters were driven at low level so the
figure of 0.6% at 85db SPL is rather unlikely.

[ThorstenL]

Hi,

Originally Posted by wahab

That such test were already done doesnt make them useless, quite the contrary.

Well, let's see. A number of test have been done with better instrumentation, methodology and controls and thus they do not make casual tests with worse instrumentation, methodology and without controls useless? I don't know, I'd likely argue that they in fact do.

Originally Posted by wahab

As for the speakers , right that the fundamental is rendered by the woofer , but H2 fall on the crossover point while H3 and above are in the range of the tweeter.

And yet the fundamental applied to the woofer will give rise to H2 and H3 in the woofer where it will mix with the amplifiers HD and will not generate higher order products.

Originally Posted by wahab

Dont know for the woofer , seems its marking was removed , i ll check later , but it s surely not a low cost OEM thingy, while the tweeter is a Morel MDT30.

The woofer looks to me like an OEM version of the common Audax Drivers, though I'd have to see the back. They are fairly common in british made monitors (see Soundcraft Zero series among others) as are Focal. Sadly Audax and Focal share the same magnet system design and that is one that is very high in distortion, compared for example to skandinavian drivers.

Audax did do some very advanced work on cones for drivers (their Carbon Fibre driver can for example be run completely open (no filter), but their magnet systems are very poor, leading to drivers with relatively speaking very high distortion.

Originally Posted by wahab

the speakers, but then , theses latters were driven at low level so the figure of 0.6% at 85db SPL is rather unlikely.

If the Bass Driver is indeed Audax, then the 0.6% H3 figure for 85dB/1m is correct, it is a feature of all their woofers, my figure comes from measurements by the German magazine Klang + Ton, where for example H2, H3 and H5 measurements are published for each and every driver tested (also waterfalls etc.

I did remark as you where at lower SPL's the 3rd and 2nd order HD of the speaker was likely reduced to around 0.1% each in your tests.

At any extent, I think you should seriously test your speakers first to see what distortion levels you are actually getting.

Linkwitz modified Panasonic (or Transound) Electret Capsules have low enough HD at these low levels to measure serious problems in Speakers, such as with the Focal/Audax woofers.

Ciao T

[dvv]

Well, you can argue all you like, but I'm happy with my Audax based speakers. The bass and mid have cones made of their take on the aerogel membranes, while the tweeter is their top of the (in its time) TWO25A16.

I'm not sure what you mean by poor magnetic assemblies, Thorsten, but my woofer has a magnet of 3.4 kilos, mounted on a die cast chassis. I wouldn't call that naive, especially after you hear it reproduce the strike of the 8 foot drum on the Blue Man group first CD - loud, clean and clear.

[john curl]

I'm with you, T. If people would read the harmonic distortion chapter in the 'Radiotron Designer's Handbook', from 1940 or so, they would have some clear info on distortion detection levels.
Not everything has to be tested for, over and over again.
For example, we know that loudspeakers generally have more distortion than the electronics that drives them. We debated this 40 years ago, but we still have found that amplifier distortion is audible, almost no matter how low it gets. This implies more subtle types of distortion, beyond conventional measurement equipment are being generated and passed through the loudspeakers to be detected by the ear. Two known measurements that fit this criterion are Hirata distortion, and PIM or FM distortion. This is where we should concentrate our efforts. I certainly have.