GR Research (incomplete) speaker measurement methodology.

[ThatSoundsGood]

I get your drive with this, but I can't agree. As Einstein said ... ... don't over-simplificate. Speakers are tech/, music is something else. Let's keep them in their respective fields.

I only say: Burt Bacharach

I love some Bacharach! I just think there's more to someone's listening experience than low distortion and proper measurements. If people can make their own speakers and they think that the speakers sound great then that's all that matters. He's selling more of an experience than a scientifically perfect speaker. But Einstein was a lot smarter than me..........

 

[amirm]

The harmonic distortion measurements that Amir and Erin make are less accurate than they would be in anechoic chamber because of the room influences present.

I started by using the feature in Klippel NFS to generate anechoic results. The difference was small so I stopped it. My space has far more volume than Erin's so modal response is much more controlled than his. Regardless, since all speakers are measured the same, we can compare the results even if absolute values are not quite correct in low frequencies.

 

[ctrl]

Nah, the HD is not of so much importance. My bass guitar, which has to be considered a relatively 'clean' musical instrument, generates 200% of harmonics aka 'distortion' easily. If there is another 10% (!) in effect may depend on my picking style more than on the speakers.

The HD of a speaker reveals systematic design errors or bad compromises. There are so many ... .

We should distinguish between the sound of an instrument and playing a recording through speakers. The sound of my distorted electric guitar is largely HD3
Nevertheless, 10% HD3@1kHz should be easily audible in a loudspeaker.

If HD measurement reveals "systematic design errors or bad compromises" as you said, then it is important.

The paper "Graphing, Interpretation, and Comparison of Results of Loudspeaker Nonlinear Distortion Measurements - ALEXANDER VOISHVILLO et al" lists a few reasons for increased harmonic distortion:

"...the harmonic distortion curves, plotted separately as a function of frequency and level of input signal, provide useful information about loudspeaker under test.

For example, a strong level of high-order harmonics may be indicative of a rubbing voice coil or the presence of nonlinear breakups of a compression driver’s metallic diaphragm and suspension.

The relationship between harmonics of even and odd orders tells about the symmetry (or the lack of it) in loudspeaker displacement-dependent parameters.

The buildup of high order harmonics with an increase in input voltage may be indicative of approaching the limit of a spider’s deflection."

My personal finding is that literally no two way design must claim to be high fidelity. That is due to intermodulation and Doppler (in room) also. A quite subjective conclusion, but if it comes to GR research (research?!?) I still feel a little bit more informed.

On the Purifi blog you will find some interesting reflections on distortions. Also that Doppler distortion is barely audible, even with large cone excursion, compared to other types of distortion.

Doppler distortion vs IMD?

purifi-audio.com

 

[ROOSKIE]

On the Purifi blog you will find some interesting reflections on distortions. Also that Doppler distortion is barely audible, even with large cone excursion, compared to other types of distortion.

Doppler distortion vs IMD?

purifi-audio.com

Aren't these two signals both distorted @Purifi? I hear the modulation in both. I think the idea @Purifi is that IMD and Doppler go hand in hand. PLus they make high excursion drive units, so they have a vested interest in the ways this is discussed that some others may not.

Linkwitz's take is in section 'J'

Frontiers

His stuff is indeed older now and he did not have the full suite of equipment available today.

It is hard to test apples to apples to get an idea of how audible doppler distortion is. How does one find two drivers that have all traits in common except driver SD/required excursion for given SPL.

That said it is suggested by Linkwitz 'it may be perceived as a roughness of the tone'. Subjectively I think that holds true in my listening from time to time and of course I have no way to remove corresponding increases in IMD and HD to isolate the effect of doppler modulations.

 

[fineMen]

We should distinguish between the sound of an instrument and playing a recording through speakers. The sound of my distorted electric guitar is largely HD3
Nevertheless, 10% HD3@1kHz should be easily audible in a loudspeaker.

On the Purifi blog you will find some interesting reflections on distortions. Also that Doppler distortion is barely audible, even with large cone excursion, compared to other types of distortion.

Doppler distortion vs IMD?

purifi-audio.com

I'm happy to not agree on the two topics of your post.

If HD3 is just that HD3 then a difference between 210% and 200% shall not be audible. Or if so, it would be utterly obscured by external parameters like e/g the picking style of the player, or simply where the string is picked. A few millimeters make a measurable difference.

It is only so that elevated distortion nearly always points to some root cause which then also generates other artifacts that may be summarized as the typical loudspeaker sound. Which again some really appreciate, compared to a 'boring', clean reproduction. Don't we hear that once in a while, that with studio grade speakers the funky drama is missing?

When it comes to Doppler, I cannot disagree more. But anyway, thank's for reviving the link to the auditory test.

Would you (all!) mind to take the test? But do not follow the instructions fully: "Put on a pair of headphones and listen to both. I’ll leave you to draw your own conclusions."
Please try on your speakers in-room. There is a conjecture, supported by the famous S. Linkwitz with math and all, that Doppler, phase distortion namely, is transformed to amplitude modulation by room acoustics. If you think about it, it is all too obvious. In the steady ('diffuse') sound field phase correlation is lost. But that phase correlation is the only thing that distinguishes Doppler from AM. I could, anecdotally, confirm this reasoning practically with arbitray repeatability blindfolded at many sound pressure levels in different rooms. The effect is not subtle, but it is massive

ps, regarding the test of Doppler as suggested by the link: in the test signal the modulation frequency is not contained; hence if you hear intermodulation with Doppler, it is not your speaker generating it

 

[Rick Sykora]

Circling back to the OP’s complaint, regardless of a specific measurement or set of them, the primary issue with Danny’s crossover work is that it lacks completeness before he starts selling it. When I built his Encore speaker for review, I mentioned a few missing items from the instructions. One was the cabinet damping, this actually took years to get a change. Another was the baffle roundover and it still is not called out in the cabinet plans. This pattern is all over his upgrade kits. The measurement data is often incomplete and simply insufficient to support his mods.

His latest target is the Monolith speakers and he recently posted an upgrade kit for the tower. The tower has 4 woofers and is bass reflex. His main bass concern howewer is power handling in the crossover for the woofers. So he ignores the math and replaces one resistor with 4. The cabinet does not pass his knock test and so he adds his No-Res to the kit. He does not measure the alleged improvement or offer clear instructions on existing damping. His kit is over $1000 and so adds about 50% to the cost of the speakers and does not even include a (mandatory) new crossover board. While harping on clarity and claiming next level performance, his changes affect the bass design but he waves off any proof of real improvement as too room dependent. In the end, he offers no way to ensure you have improved your investment with his extensive mods either.

Save your money and enjoy what you have or buy better speakers. Danny does not deserve sales for unproven design mods.

p.s. Not surprisingly, the bigger issue is that the Danny’s whole process is incomplete not just the measurement aspect!

 

[ctrl]

Aren't these two signals both distorted @Purifi? I hear the modulation in both. I think the idea @Purifi is that IMD and Doppler go hand in hand. PLus they make high excursion drive units, so they have a vested interest in the ways this is discussed that some others may not.

Linkwitz's take is in section 'J'
Frontiers
His stuff is indeed older now and he did not have the full suite of equipment available today.

Linkwitz based his example on a 6.5'' woofer with +-5mm excursion. This achieves about 86dB SPL in the free field. But which speaker concept is this based on?

For this I have compared three concepts, BR, CB, and OB with a small baffle of 22.5cm x 40cm common for shelf speakers.
In all three concepts the SPL at 50Hz was chosen so that there was always 5mm excursion (for the BR concept the tuning frequency prevents large excursions at 50Hz, so the SPL was increased until 5mm excursion was reached near 50Hz).

This results in the following SPL at 50Hz for the loudspeakers in the free field:

For the OB and CB concept, this results in an SPL of 86-87dB at 50Hz - many CB or OB speakers with a single 6.5'' woofer are unlikely to exist. For the BR concept, which is the standard case, the SPL for +-5mm excursion around 50Hz is about 93-94dB@1m at free field.

In a listening room with two speaker in stereo and additionally some room load this corresponds to an SPL of close to 100dB@1m at 50Hz.
For frequencies above 50Hz, the SPL increases depending on the XO tuning and should be around 100-106 dB@1m.

What I mean is that Doppler distortion is not a problem at all for most listeners of 6.5'' bookshelf speakers, since the sound pressure levels are hardly reached until DD becomes dominant/audible.

If HD3 is just that HD3 then a difference between 210% and 200% shall not be audible.

What I meant is that it makes a difference whether a single instrument uses distortion as an effect or the speaker adds distortion to every recording (all clean vocals and instruments) all the time.

Would you (all!) mind to take the test? But do not follow the instructions fully: "Put on a pair of headphones and listen to both. I’ll leave you to draw your own conclusions."
Please try on your speakers in-room.

I took the test with headphone and speaker. Both signals showed audible distortion, but in both cases the FM signal was less disturbing. The more room influence is added (i.e. the worse my listening position is), the more the AM and FM signals converge.

However, one should also keep in mind here what the conditions are for the Purifi listening example. The carrier frequency is 1.4kHz, the modulation frequency 43Hz and the excursion +-12.5mm. Therefore delta f = 2*pi*fL*fH*Xp / c = 13.8 Hz. So the example from the Purifi Blog is much more worse then the example Linkwitz makes.

This corresponds to an SPL of around 92dB@43Hz@1m in the free field for a shelf speaker with PR. In stereo and with room load effect this should be close to 100dB@1m at 43Hz. This is already outside the range of use for a bookshelf speaker with 6.5'' woofer.

But of course, for a very good driver with extremely low distortion, Doppler distortion is a natural limit that cannot be undercut.

So, yes, theoretically the Doppler effect is a problem, but for 2-way bookshelf speakers, if operated in the intended volume range, rather less.


If one want to listen at high SPL with a 2-way BR concept, you need a 15'' woofer with XO to a horn around 1.2kHz.
With a baffle of 40cm x 70cm you need just 4mm excursion for 105dB@40Hz@1m.

According to Linkwitz delta f = 2*pi*40*1200*Xp / c = 3.5 Hz.
If the listening room has a normal T60 decay time of 0.3-0.4 s (Linkwitz used with 0.5s a really bad one). My room has T60 = 0.37 s (without any damping treatment), this results in a bandwidth of BW = 2.2 / T60 = 5.9Hz. And thus then 3dB / (5.9/2)Hz = 1 dB/Hz. Thus in the worst case 3.5Hz * 1dB/Hz = 3.5 dB amplitude change at 1.2kHz at 105dB@40Hz@1m - at these SPL no one cares.

The amplitude change of the Purifi listening example @1.4kHz would be in my room 13.8Hz * 1dB/Hz = 13.8dB.
At normal SPL, Doppler distortion is not a thing.

 

[fineMen]

Linkwitz based his example on a 6.5'' woofer with +-5mm excursion. ...
What I mean is that Doppler distortion is not a problem at all for most listeners of 6.5'' bookshelf speakers, since the sound pressure levels are hardly reached until DD becomes dominant/audible.

Direct link to S.Linkwitz' site:

Frontiers

Linkwitz states, supported by first order calculations, given the 6,5" woofer with excursion of 5mm: " ... or 20.7 dB below the 2000 Hz carrier AH, regardless of fL as long as Xp = 5 mm."

This is 10% of a Doppler distortion. Of course this has to be transformed to amplitude modulation first to become readily audible. I would argue that the projection that Linkwitz suggests, referring to room resonances and their damping doesn't hold fully. It is just and only the phase relation that discriminates the Doppler from amplitude modulation. In-room for the grand total of sound power any phase relation is lost entirely. Doppler becomes effectively amplitude by that.

My personal threshold for detecting sub-harmonics is as low as 0,3% with pure tones, but should be relaxed to 1% with complex sounds for good measure. Taking the Klippel test I voluntarily stopped at -53dB, with is roughly consistent with my very personal findings. Most critical bands are in the lower mids. This would restrict, for a very high standard, the excursion of a bass mid to, say 2mm?

Regarding H3 harmonic distortion and other

What I meant is that it makes a difference whether a single instrument uses distortion as an effect or the speaker adds distortion to every recording (all clean vocals and instruments) all the time.

Exactly the opposite of what I think. The H2, H3 ... are inherent to the instrument's sound, and not an effect. Only synthesizers can, but never will generate pure tones. Because it was so boring. Everything else is composed of the base tone and the overtones. Harmonics are a decisive part in the musical representation and vastly underestimated in amplitude by hifi enthusiasts (two-hundred percent easily, no 'effects' applied!).
Reiterated, it is not the harmonics, or THD, but the generating machanism in the replay device(s) which generate other nasty stuff also. That all combined 'makes' the typical speaker sound. All that drama

I took the test with headphone and speaker.

Thank you!

The more room influence is added (i.e. the worse my listening position is), the more the AM and FM signals converge.

Yep, I find this interesting.

However, one should also keep in mind here what the conditions are for the Purifi listening example.

I can only hold anecdotal experience against that. My switch to three or even four way, when I was still in DIY, made a huge difference. It may be, though, that part of the enhancement came from avoiding the 'cone edge' resonance, namely the surround's resonance. The latter is a cruel , unforgiving source of AM with many drivers, but obviously not with the Purifies. Because they look so ugly

 

[kemmler3D]

The H2, H3 ... are inherent to the instrument's sound, and not an effect. Only synthesizers can, but never will generate pure tones. Because it was so boring. Everything else is composed of the base tone and the overtones. Harmonics are a decisive part in the musical representation and vastly underestimated in amplitude by hifi enthusiasts (two-hundred percent easily, no 'effects' applied!).

Yes, and I wanted to comment on this also. I don't think it makes sense to compare the harmonic content of a bass guitar tone and THD of a speaker. A bass guitar CAN'T produce a pure sine tone under normal circumstances, except later in the decay. It's not how a plucked or strummed string moves.

An ideal speaker should be able to produce a pure sine wave, if we ask it to.

There's also the fundamental point that harmonic content from an instrument is desired, harmonic distortion from playback equipment is (usually) undesired.

I agree that most hifi enthusiasts don't really understand timbre or harmonics in music, and that is harmful to their approach to sound reproduction. For example, the top note of a piccolo (~5khz) is sometimes cited as the maximum frequency you really need to care about when you reproduce music. Well, not really, even the piccolo has overtones.

 

[ctrl]

This would restrict, for a very high standard, the excursion of a bass mid to, say 2mm?

Okay, if that is your personal limit for Doppler distortion, then we simulate for a 6.5'' woofer a loudspeaker with 22.5x40cm baffle in free field and allow maximum +-2mm excursion down to 40Hz.

Tuned to high sensitivity, we get f3=70Hz and f6=56Hz. At 2mm cone excursion this tuning would produce 94dB@1m.
When tuned to a f3=56Hz and f6=45Hz, 2mm cone excursion would produce 90dB@1m.

This should be more than enough SPL for most listeners of small bookshelf speakers with about 2m listening distance.

Exactly the opposite of what I think. The H2, H3 ... are inherent to the instrument's sound, and not an effect.

I think we mean the same thing. It is clear that any (non pure electronic) instrument produces harmonics. But if your bass has 200% distortion (as you stated), then this is an effect (caused by an pickups, effect device, amplifier, bass speaker,...) not the "natural" overtones of your bass guitar string.

This cannot be compared to the reproduction of distortion by a loudspeaker, which gives every recording an unnatural addition of overtones.

My switch to three or even four way, when I was still in DIY, made a huge difference.

If the quality of the drivers remains high, one is certainly always better off with three or four ways in terms of distortion. However, one should not exaggerate the influence of Doppler distortion in "normal listening".

For open baffle speakers, high excursion is much more of a problem and Doppler distortion might cause problems with two-way concepts. However, it is not uncommon to use two 15'' woofers, which should not be a real concern in terms of Doppler distortion with a crossover frequency around 1.0-1.5 kHz.

 

[kemmler3D]

But if your bass has 200% distortion (as you stated), then this is an effect (caused by an pickups, effect device, amplifier, bass speaker,...) not the "natural" overtones of your bass guitar string.

This cannot be compared to the reproduction of distortion by a loudspeaker, which gives every recording an unnatural addition of overtones.

+1 for emphasis here. Harmonic distortion and an instrument's own natural harmonic series shouldn't be compared. Not sure if that was the intention, but someone might read it that way.

 

[fineMen]

This should be more than enough SPL for most listeners of small bookshelf speakers with about 2m listening distance.

To be clear, that is the limit then. Just 2mm of excursion even with a Purify driver offering 12mm.

But if your bass has 200% distortion (as you stated), then this is an effect (caused by an pickups, effect device, amplifier, bass speaker,...) not the "natural" overtones of your bass guitar string.

It is not 'distortion'. It comes from picking the string, basically. Picking forces the string into a triangular shape. That shape brings the overtones into being. That simple actually. With an upright bass you easily have 2000% ( 4 digits ) of harmonics due to limited effectiveness of radiating the base tone.

If the quality of the drivers remains high, one is certainly always better off with three or four ways in terms of distortion. However, one should not exaggerate the influence of Doppler distortion in "normal listening".

That is where my anecdotal experience chimes in. One can clearly under-estimate the phase aka Doppler distortion. But it is of lesser concern to me, if you really get the clue. It is just a hint, whatever you do with it. See above ('look up!').

 

[restorer-john]

Circling back to the OP’s complaint, regardless of a specific measurement or set of them, the primary issue with Danny’s crossover work is that it lacks completeness before he starts selling it. When I built his Encore speaker for review, I mentioned a few missing items from the instructions. One was the cabinet damping, this actually took years to get a change. Another was the baffle roundover and it still is not called out in the cabinet plans. This pattern is all over his upgrade kits. The measurement data is often incomplete and simply insufficient to support his mods.

His latest target is the Monolith speakers and he recently posted an upgrade kit for the tower. The tower has 4 woofers and is bass reflex. His main bass concern howewer is power handling in the crossover for the woofers. So he ignores the math and replaces one resistor with 4. The cabinet does not pass his knock test and so he adds his No-Res to the kit. He does not measure the alleged improvement or offer clear instructions on existing damping. His kit is over $1000 and so adds about 50% to the cost of the speakers and does not even include a (mandatory) new crossover board. While harping on clarity and claiming next level performance, his changes affect the bass design but he waves off any proof of real improvement as too room dependent. In the end, he offers no way to ensure you have improved your investment with his extensive mods either.

Save your money and enjoy what you have or buy better speakers. Danny does not deserve sales for unproven design mods.

p.s. Not surprisingly, the bigger issue is that the Danny’s whole process is incomplete not just the measurement aspect!

You pretty much sum it all up there.