Sound Quality Vs. Measurements

Status
Not open for further replies.
diyAudio Member RIP
Joined 2005
I recently tried Spice . I got so bogged down . I still use a spreadsheet . A calculator is useful , often my phone as an idea comes to me sometimes miles from home . I an terrible at arithmetic and have to recheck . Like when pico or nano farads . It matters a lot and sometimes .........! I often get people to check my stuff after , usually nothing is changed as by then I have built it . I use Burr Brown Filter Pro freeware ( I bought it originally ) . I lost my fear of Chebishev filters with that . Sometimes the best solution a Chebishev . .

http://www.ti.com/lit/an/slyt113/slyt113.pdf

I dodged computers in general for many years, after I realized I'd quickly wind up a programmer. When I got an HP-25 I spent the next two weeks doing nothing but writing programs for its ludicrously limited capacity. Some years later I started to play with one of the early Compaq luggables and once again was spending huge amounts of time.

One prof I knew well at UCLA bought each new HP offering, and spent almost all of his time for the next two years (!) writing programs for the 65 and 67, all the while maintaining how sensible this was, as the resulting tools would be so powerful and save him so much time (he did a lot of optical design). After a while the code he developed with a grad student, far beyond the limits of the HP calculators, took over, and he spent a lot of time on the main machine at the university, an IBM 360 mod 91KK, iirc. At one point the astonishing amount of fast memory with that machine was about 2 million words!

But I managed to design and build a lot of equipment before embracing simulators. It's fun to go back now, and dig out some of the schematics and plug them into Spice. Generally the design decisions are vindicated, though I'd do a lot of things differently today. One of the things that made my UCLA position so educational: there was a budget of <50 dollars a day for buying things without going through the UCLA highly-bureaucratic purchasing dept. As a result, I could buy transistors but not modules, unless the latter went through purchasing and of course as well were approved by the Chair.

By the time the particular instrument was completed, it had exactly two off-the-shelf modules in it: an ADC and a sample-hold. Of course time could have been saved with more modular opamps, power supplies, etc., but much of the required circuitry really did need to be custom.
 
Administrator
Joined 2004
Paid Member
Well, Hi-Eff drivers do have their quirks, that's for sure. It takes a good bit of crossover skill to tame those quirks. But once you have, it's well worth it.

Medium efficiency drivers often seem much easier to build crossovers for. I mean the crossovers are closer to textbook.
 
Loudspeakers are interesting, to be sure. One thing that DOES count with speaker efficiency, is improved magnetic path and magnet, which, like compression ratio in an automobile, gives you more efficiency, no matter what else you do. Unfortunately, really good magnetic assemblies, this includes the magnet and the steel cost REAL money, and are not the first choice of 'practical' designers.

Also, horns, which will make ANY loudspeaker much more efficient, do give improved 'internal clarity' because the speaker cone has to do so little to get so much sound output. One weakness of horns, unfortunately, is that it is almost impossible to make a full range one, therefore, a 2 or 3 way horn speaker having a stepped arrival time, normally, the K-horn being an excellent example of this.

I was once told by a professional loudspeaker designer from UK, staying nameless as he still works in the LS company, that their design goal was to spend 10-12% of the FACTORY sales price on drivers. Top models were stretched to 15%.

When you calculate backwards, the remaning sum for the drivers is generally pitiful. It makes you want to cry.

However, do bear in mind that many of highly touted European driver manufacturers have VERY different prices for wholesale buyers, like loudspeaker companies. Thus, driver you may be asked to pay €220 (app. $300) for a driver, but an LS manufacturing company can have it for €86 (app. $112). If they do well, everybody will know they use X's drivers, and sales will be good.

My own 10" bass driver cost some €350 on the street, but I got it for €210, which was the manufacturer price from Audax at the time.

Just one remark on the automoile engine analogy, John - my dad, a mechanical engineer who worked in development of engines for around 12 years just after WW2 used to say that NOTHING eats up the engine from the inside as high compression ratios after wild revving up to 7,000+ rpm. Combining the two was u asure fire way to eat up your engine quickly, unless you resorted to superior materials, in which case it became very costly.

The entire US muscle car era is based on 11.5:1 and greater compression ratios, but this was compensated by extremely large capacities, which allowed them to work at below 4.000 rpm. Kind of a petrol fuel diesel approach. At the time, Europeans preferred faster revving engines with much smaller compression ratios and always outsold their cars in the US as compared to US cars in Europe by something like 10,000:1, or some such.

Just a remark, if you like we can expand on it in the car thread.
 
@Pano ..... Does your pass the headphones test .... ?

Would you be alluding to a direct comparison with a pair of cans?

That's a thankless road to go on, Wayne. The industry creates headphones outputs by using series 220 Ohm or so resitor, with another one to the ground, to act as a voltage divider. Cheap and cheerful, also tend to kill the sound dead.

If on the other hand you do use a specialized headphone amplifier, and have half-decent headphones, you live on another planet. My own amps operate in pure class A and have an output impedance of 0.1 Ohms 0...20,000 Hz, etc. THAT really shows what headphones ara capable of, and it digs up the wealth of information in the signal.

I have never seen or heard any loudspeaker which could stand up to that, but on the other hand, some people plain hate headphones, and having the sound all around you IS a bit more natural than just on your ears. Yet, it's only then that you discover what kind of ultra high quality bass some cans are capable of. Cases in point - Sennheiser 598, Koss Pro4AA Titanium, etc.

Yessir, you'd be surprised what a fully complementary, pure class A headphone amp can do for your musical pleasure, with two 50W devices in a SEPP confuguration. :D :D :D

Oh yeah - it doubles down to 5 Ohms. :cool:
 
I was once told by a professional loudspeaker designer from UK, staying nameless as he still works in the LS company, that their design goal was to spend 10-12% of the FACTORY sales price on drivers. Top models were stretched to 15%.

It was 10% the retail price for everything as I knew it . Still they did not make money as the sales teams were often not so very red hot . Naim I know made considerably less as R&D gobbled it up . I think Naim sited that recently as the reason for the merger . Simply not having enough R&D money , forgive me Paul if I am wrong ?
 
I was once told by a professional loudspeaker designer from UK, staying nameless as he still works in the LS company, that their design goal was to spend 10-12% of the FACTORY sales price on drivers. Top models were stretched to 15%.

When you calculate backwards, the remaning sum for the drivers is generally pitiful. It makes you want to cry.

Rule of thumb for any successful manufacturing operation is retail price is 5x cost of BOM. You need drivers, crossover, cabinet, packing material, and box, so that's about the right allocation for drivers.
 
A good set of cans are good for voicing , very few speakers will pass this test, coloration abound...


I was once told by a professional loudspeaker designer from UK, staying nameless as he still works in the LS company, that their design goal was to spend 10-12% of the FACTORY sales price on drivers. Top models were stretched to 15%.

5x1 is about right , cabinet cost Kills you , shipping kills you and frankly , real expensive drive units are a waste of money ( diminishing returns) then just when you have it lock , some kook writes the only way to go is with diamonds and Bam ...:rolleyes:

Hi end prices ..... :crazy:


Just one remark on the automoile engine analogy, John - my dad, a mechanical engineer who worked in development of engines for around 12 years just after WW2 used to say that NOTHING eats up the engine from the inside as high compression ratios after wild revving up to 7,000+ rpm. Combining the two was u asure fire way to eat up your engine quickly, unless you resorted to superior materials, in which case it became very costly.

The entire US muscle car era is based on 11.5:1 and greater compression ratios, but this was compensated by extremely large capacities, which allowed them to work at below 4.000 rpm. Kind of a petrol fuel diesel approach. At the time, Europeans preferred faster revving engines with much smaller compression ratios and always outsold their cars in the US as compared to US cars in Europe by something like 10,000:1, or some such.

Just a remark, if you like we can expand on it in the car thread.

This was because they never had the electronics we have today to alter timing and fuel based on load and we spend those big blocks to 9000 RPM ..

:)
 
People often forget that the designers, themselves, have to make something, and the person who sells it, let's say, retail, has to make something, and then there are the company costs like the manufacturing building, storage, etc., telephone, building alarm service, electricity, etc. etc. This drives the cost of a lot of audio components up, and/or profits down.
This is where 'promotion' even selling a story, much like Mark Levenson often does, helps to get the extra sales that can make a company a success. I KNOW that just making the best product I can make and trying to sell it at a reasonable price does not 'cut it' and that is why I never returned to manufacturing products after I closed down Vendetta Research.
I KNOW that making a world class audio component can become outrageously expensive, just because we want it to be as 'perfect' as possible, both sonically and visually. We fight it every project.
Now many of you might think that all you need to do is a circuit sim to design it, and a little time to build a prototype, well that will get you a 'hoopty' but NOT a race car. '-)
 
Administrator
Joined 2004
Paid Member
Would you be alluding to a direct comparison with a pair of cans?
Yeah, that must be it. It's a strange comparison because cans are so limited. That in-head sound, the lack of space and feeling. But yeah, my speakers are better in most ways than my Sennheiser phones. They better be, considering their size, cost and effort! Very similar in tonal balance, more revealing of space. Not the hyper detail of good headphones, but they aren't scant millimeters from my ears, either. Certainly more life-like.

It's sort of a kumquat vs grapefruit comparison.
 
Yeah, that must be it. It's a strange comparison because cans are so limited. That in-head sound, the lack of space and feeling. But yeah, my speakers are better in most ways than my Sennheiser phones. They better be, considering their size, cost and effort! Very similar in tonal balance, more revealing of space. Not the hyper detail of good headphones, but they aren't scant millimeters from my ears, either. Certainly more life-like.

It's sort of a kumquat vs grapefruit comparison.

No it's not, cans make a good reference tool and I'm not talking about that in your head sound , check for tonal balance and information retrieval .
 
It is unfortunate that generalisations are made without any real supporting evidence. The standard +/- 3dB tolerance for what makes a fairly good speaker amounts to bad practice. What many people don't realise, some designers included, is that measurement methodologies are poor when it comes to speakers. In my experience +/- 0.3dB change in frequency balance is audible on music material. Not only that, +/- 0.3dB change in frequency balance between two frequencies 1/10th octave apart is also audible on music material. If this is true then Stereophile's measurements don't mean a whole lot, but people, myself included, like to read them.

The task is really to identify methods that do give rise correlation with what you hear, rather than take the position that measurements don't giive rise to correlation. My favoured way of lining up a set of studio speakers is to use an AKG C414 microphone. It does not have a ruler flat response but it does have a hyper cardioid setting which allows the measurement to be made at the point of interest. When correction is made for its frequency balance by calibration against a flat microphone, very good correlation results.

It is also necessary to understand that all frequencies are not made equal. This goes beyond phon curve measurements which indicate the frequencies that the ear finds most sensitive. Certain key frequencies outside the most sensitive range have significant bearing on overall voicing of a speaker. It amazes me how often this is ignored.

Horns can be demonstrated to have moving mass benefits compared to direct radiating cone speakers. For those of who cannot support a flat panel envorinment, they are a good option especially where controlled directivity is also a beneficial factor.
 
Status
Not open for further replies.