Measuring for tonal characteristics

[Phelonious Ponk]

Sounds like they are the only speakers worthy of buying then

Are the Revel's what you have?

I think you could probably pick whatever Harman makes at whatever price you can afford and know you'd be doing well. Other speaker companies do well too, but often not as well, at much greater cost.

Tim

 

[fas42]

IOW you have absolutely no idea what the "recording sounds like", as long as transduction X "sounds good" to you. Ok.
We're in the same boat.

cheers,

AJ

Except, in my case, " transduction X sounds good" equals no audible defects - one can learn to listen for defects, and once attuned to such things it becomes painfully obvious that conventional playback is riddled with defects - they're so common that they are accepted as part of conventional hifi sound, people are blind, sorry, deaf to them because that's all they've ever heard, they believe it's intrinsic to reproduced sound.

 

[Mivera]

So Mike, Amir has confirmed that Harman is doing exactly what I said - measurement and very thorough blind listening evaluations, correlating measurements with sound, and developing products based on the resulting data. Yet you still haven't answered my simple question -- how are they doing that if, as you've said, measurements and sound cannot be correlated. Are you better at this than the engineering staff of a 6.5 billion dollar audio company, or could you perhaps be...what's the word...oh yeah, wrong?

Not that I expect anything more than another pithy, off-topic remark, or another diversionary question, but hope springs eternal.

Tim

How is that any different than I said most do it? The only difference I can see is they are manufacturing their own drivers, and perhaps doing more extensive subjective testing than some others. However before they have people do blind testing, I'm quite sure the design team listened to them first. And I'm quite sure that they may had gone through a few prototypes with different drivers, materials etc before they had prototypes ready to demo to the blind testers.

 

[fas42]

In regards to testing methodology, so far I'm not very impressed with what I can find out about Harman's setup - sorry amirm! - my understanding is that they're using a very ordinary, multichannel amplifier to do the driving, one that very few keen audio enthusiasts would give room space to ...

 

[AJ Soundfield]

Except, in my case, " transduction X sounds good" equals no audible defects - one can learn to listen for defects

If you said so. Compared to what defect free reference transduction of the construct?

 

[fas42]

If you said so. Compared to what defect free reference transduction of the construct?

An easy answer is the human voice. Direct experience of such makes all of us extremely sensitive to the nuances of the sound of that - and much audio reproduction generates a caricature of those qualities. When a system works correctly, a voice reproduced sounds, well, human - rather than a caricature.

Surprisingly, even on really "duff" recordings that humanness of the recorded voice can still trick the ear, if the cues in the recording come through cleanly enough. I have many recordings that I once thought were "hopeless", but upgrading the quality of the playback chain has allowed them, at least on occasion, to make the grade.

 

[AJ Soundfield]

An easy answer is the human voice.

Which defect free transduction of that recorded voice are you comparing yours to?
You're running in circles here.

 

[fas42]

Which defect free transduction of that recorded voice are you comparing yours to?
You're running in circles here.

It doesn't have to be any particular voice ... as I just implied, people in general can easily pick whether a voice coming from another room, say, is the real thing, or a recording, or a PA system - it's the overall qualities that people register in an instance, they don't have to think about it ...

Do you have trouble distinguishing the playback of a recorded voice from the genuine article?

 

[RayDunzl]

Do you have trouble distinguishing the playback of a recorded voice from the genuine article?

Not at all. If I am not speaking and I hear myself, I know it is fake.

On the other hand, my recorded voice sounds nothing like me!

 

[Phelonious Ponk]

How is that any different than I said most do it? The only difference I can see is they are manufacturing their own drivers, and perhaps doing more extensive subjective testing than some others. However before they have people do blind testing, I'm quite sure the design team listened to them first. And I'm quite sure that they may had gone through a few prototypes with different drivers, materials etc before they had prototypes ready to demo to the blind testers.

Really? Their design system is built on correlating measurements to sound.

That is not only different from what you said, it's exactly the opposite of what you said, Mike. You said measurements do not predict sound, and that's precisely what Harman is doing. They're using measurement data to develop products. They're designing to targets defined by the data, then they are getting statistically consistent results in blind preference tests. They are, statistically, correlating measurements to sound. So I'm still asking my first question: If, as you say, you can't do that, how are they getting the results? Luck? Alchemy?

Tim

 

[Mivera]

Really? Their design system is built on correlating measurements to sound.

That is not only different from what you said, it's exactly the opposite of what you said, Mike. You said measurements do not predict sound, and that's precisely what Harman is doing. They're using measurement data to develop products. They're designing to targets defined by the data, then they are getting statistically consistent results in blind preference tests. They are, statistically, correlating measurements to sound. So I'm still asking my first question: If, as you say, you can't do that, how are they getting the results? Luck? Alchemy?

Tim

Obviously you just don't understand. Different cone/dome materials have different tonal qualities. Even when they are used in Harman products. Even if Harman engineers pop the drivers in their cabinets they aren't immune to this reality. And at some point a designer made the decision to use that material.

This seems to be how you think the design process went:

1: They hired engineers who never heard a speaker before in their lives

2: Did up the driver designs in CAD based on theoretical data

3: Designed cabinets based on theoretical calculations

3: Built crossovers based on theoretical calculations.

4: Once together they went direct to measurement sweeps for evaluation.

5: Made changes to the drivers/cabinets/crossovers based 100% on measured data

6: Once they were satisfied with the measured data, they went direct to blind testing with thousands of listeners, comparing against all of the competition.

If this is how things went, I'm not sure why the listening tests were even required? They already knew based on the data that they were perfect speakers. What more could listening tests possibly bring to the table?

 

[NorthSky]

Different cone/dome materials have different tonal qualities. Even when they are used in Harman products. Even if Harman engineers pop the drivers in their cabinets they aren't immune to this reality. And at some point a designer made the decision to use that material.

Paper, cotton, textile, soft, hard, aluminum, gold, beryllium, plastic, ribbon, nylon, silk, silver, ceramic, kevlar, glue, stitches, geometry, dots, lines, tweeter's wave guide, rubber, felt, metal, foam, fiberglass, ...they all have various tonal characteristics.
The spider, the voice coil, the magnet material, the diameter, the length, the structure, the composition of the drivers, the crossover parts, their slopes, the frequencies where drivers cross, their united designs, the box enclosures, woods, density, thickness, the bracing, the material used inside the box, the glue, the aluminum rods, the rigidity, the non-parallel walls, the curvatures, the phase tuning, the positioning of the drivers, the coherence, their isolation and insulation from internal interference, the external shape, the 360° polar dispersion, the overall internal and external design with all the parts selected and computer tuned, the part's values, their selection, the research, the experiments, the wiring, the circuit boards (x-overs), ...all contribute to the speaker's final "identity/tonality".

Burn-in time? That too. Colors? No. ...Only psychologically.

* Reason for edit: Coherence.

 

[Mivera]

Paper, cotton, textile,soft, hard, aluminum, gold, beryllium, plastic, ribbon, nylon, silk, silver, ceramic, kevlar, glue, stitches, geometry, dots, lines, tweeter's wave guide, rubber, felt, metal, foam, fiberglass, ...they all have various tonal characteristics.
The spider, the voice coil, the magnet material, the diameter, the length, the structure, the composition of the drivers, the crossover parts, their slopes,the frequencies where drivers cross, their united designs, the box enclosures, woods, density, thickness, the bracing, the material used inside the box, the glue, the aluminum rods, the rigidity, the non-parallel walls, the curvatures, the phase tuning, the positioning of the drivers, their isolation and insulation from internal interference, the external shape, the 360° polar dispersion, the overall internal and external design with all the parts selected and computer tuned, the part's values, their selection, the research, the experiments, the wiring, the circuit boards (x-overs), ...all contribute to the speaker's final "identity/tonality".

Burn-in time? That too. Colors? No. ...Only psychologically.

Wow Bob are you sure you're not Sean Olive in disguise?

 

[NorthSky]

I was going to post couple scientific links on driver's material regarding their tonal properties, characteristics, but I've read enough audio articles in the last 45 years.

 

[Blumlein 88]

Obviously you just don't understand. Different cone/dome materials have different tonal qualities. Even when they are used in Harman products. Even if Harman engineers pop the drivers in their cabinets they aren't immune to this reality. And at some point a designer made the decision to use that material.

This seems to be how you think the design process went:

1: They hired engineers who never heard a speaker before in their lives

2: Did up the driver designs in CAD based on theoretical data

3: Designed cabinets based on theoretical calculations

3: Built crossovers based on theoretical calculations.

4: Once together they went direct to measurement sweeps for evaluation.

5: Made changes to the drivers/cabinets/crossovers based 100% on measured data

6: Once they were satisfied with the measured data, they went direct to blind testing with thousands of listeners, comparing against all of the competition.

If this is how things went, I'm not sure why the listening tests were even required? They already knew based on the data that they were perfect speakers. What more could listening tests possibly bring to the table?

This is supposed to be some sort of refutation of Tim's post? Sorry for laughing, but I am laughing at you not with you.

Are you so unfamiliar with how this started and progressed or just pretending? If not we can point you to descriptions of the basics.

Regardless of any of that, when a company has the policy of beating all the main competition at each price point and can do so then they have made the better speaker. Something you seem to disregard. Your post is so off it isn't even wrong.

 

[Blumlein 88]

An easy answer is the human voice. Direct experience of such makes all of us extremely sensitive to the nuances of the sound of that - and much audio reproduction generates a caricature of those qualities. When a system works correctly, a voice reproduced sounds, well, human - rather than a caricature.

Surprisingly, even on really "duff" recordings that humanness of the recorded voice can still trick the ear, if the cues in the recording come through cleanly enough. I have many recordings that I once thought were "hopeless", but upgrading the quality of the playback chain has allowed them, at least on occasion, to make the grade.

Harman has data on what types of recordings result in showing the differences most clearly and reliably to human listeners too. Hopefully I am not portraying it wrong to say mainly the idea is the simpler the better. Female voice, female pop, jazz trios are better at determining relative quality than say full orchestras playing the classics.

 

[Mivera]

This is supposed to be some sort of refutation of Tim's post? Sorry for laughing, but I am laughing at you not with you.

Are you so unfamiliar with how this started and progressed or just pretending? If not we can point you to descriptions of the basics.

Regardless of any of that, when a company has the policy of beating all the main competition at each price point and can do so then they have made the better speaker. Something you seem to disregard. Your post is so off it isn't even wrong.

You obviously don't understand what the topic of this thread is about. It's about measuring for tonal characteristics. There's a reason human listening tests are involved with the R&D process. If the measurement gear was enough, they could just put the speakers out on the market without ever even listening to them once.

 

[Blumlein 88]

You obviously don't understand what the topic of this thread is about. It's about measuring for tonal characteristics. There's a reason human listening tests are involved with the R&D process. If the measurement gear was enough, they could just put the speakers out on the market without ever even listening to them once.

Again, not even wrong.

If their premise is sound, and they can provide evidence it is largely right, they could indeed put out speakers based upon their model at this point. Whether they do or not I don't know. They do however design to that model which they confirm in listening tests. Do they ever have unexpected results and have to make alterations? I don't know. They seem to be at the stage of designing to the model and confirming with listening to expand and refine that model. A feedback loop which will improve the model as they go. Which will improve the designs as they go. Which will improve the sound quality at any give price point as they go.

 

[Mivera]

Again, not even wrong.

If their premise is sound, and they can provide evidence it is largely right, they could indeed put out speakers based upon their model at this point. Whether they do or not I don't know. They do however design to that model which they confirm in listening tests. Do they ever have unexpected results and have to make alterations? I don't know. They seem to be at the stage of designing to the model and confirming with listening to expand and refine that model. A feedback loop which will improve the model as they go. Which will improve the designs as they go. Which will improve the sound quality at any give price point as they go.

I can assure you that several listening tests are conducted in the design process in many stages before even the 1st blind test is conducted. Reality is not always what people who have no idea how the process works think based on the marketing material. They probably went through dozen's of drivers and cone materials in the process that underwent several measured and subjective tests. It's the very final stages when they have a product complete that they do the final blind testing. And if for some reason people think the tonal characteristics are off, they probably go back to the drawing board. And even the final product will still have different tonal characteristics than every other speaker. Put the Revel Salon2 beside the Vivid Giya 3 and compare. I can assure you they won't sound 100% the same. Does that mean 1 is better than the other one?

 

[NorthSky]

♦ http://www.stonessoundstudio.com.au/stone/diy_speaker_info/diy_driver_types_p3.htm

♦ http://www.techhive.com/article/2000385/what-makes-one-speaker-better-than-another.html

"Cone material is just one example of how materials can translate to a difference in sound quality. The materials used in everything from the wires to the glue in a speaker can impact the overall sound; every piece adds up to the entire sound output of a speaker, which is why speaker designers select each component carefully."
_________

• http://www.mother-of-tone.com/speaker.htm

• http://www.nutshellhifi.com/library/speaker-design2.html

• http://www.pocket-lint.com/news/131...d-it-what-to-look-for-when-choosing-a-speaker

• http://www.audioholics.com/loudspeaker-design/loudspeaker-drivers/diaphragm-material

"In the end, the overall sonic character of the speaker has as much—if not more—to do with the designer’s voicing decisions and other design choices as it does with any single arbitrary tweeter dome material. What one hears in the highest frequency ranges covered by the tweeter is a complex combination of materials, geometry of the parts and the tweeter frontplate, the baffle, the grille, and the crossover network, even if we eliminate the clear contribution of the amplifier and source material. This can vary significantly from one angle to another. Often moving 30 degrees off axis will radically affect the response you hear in the top octave. There is NO guarantee that a resonance will produce a peak more than a dip. This is a 50/50 split. There is NO guarantee of the Q or amplitude of that resonance. These things will often confound confuse and surprise even designers that have done it many times before. The amount of small issues that can go wrong and create real audible measurable artifacts is large enough to warrant a chapter on its own.

Bottom Line on Loudspeaker Diaphragm Material

Stiff cone drivers are typically better than flexible cone drivers (all things being equal). However, no magic cone material alone will determine the quality of the sound or performance of the driver. Cone geometry and proper dampening of the cone material to better manage its behavior at and above its break-up mode plays a vital role in how the cone will sound. It's critical to avoid operating the drivers in their break up mode so a clear understanding by the designer through a properly executed crossover network is key."