How does speaker efficiency/sensitivity effect sound quality?

MrJolly

Member
2014-04-23 9:54 am
I've previously followed the Boominator thread and had a lot of fun making a portable box powered by a 12v battery, a 6w amp and some efficient drivers I managed to get cheap on eBay:

qgFbulY.jpg


(2 x vintage 12" + 2x P-Audio tweeters)


As expected, it gives out plenty of volume considering the underpowered amp and sounds pretty good too :D. What surprised me was how easily it handled dynamics in the recordings - live jazz and blues actually sounded live, everything jumped out and had a presence, all while feeling completely at 'ease' and plenty of headroom to spare. The speakers beside it sound great and probably have a much flatter response but I feel something is lacking when switching over to them.

I see a lot of speaker projects designed with high efficiency in mind to compliment low powered SET amps - is there a benefit to high efficiency regardless of power requirements?
 
...The speakers beside it sound great and probably have a much flatter response but I feel something is lacking when switching over to them...I see a lot of speaker projects designed with high efficiency in mind to compliment low powered SET amps - is there a benefit to high efficiency regardless of power requirements?

Yes. The problem is that there are many people that don't like the answers.

The papers on this subject say that direct radiating loudspeakers produce many times more modulation and intermodulation distortion than if they were simply placed in a well-designed horn, i.e. higher efficiency loudspeakers equate to lower distortion, of the type that gives the "flat" or even "loud" sound that you might describe in loudspeakers.

Most of these papers were written by Klipsch, but more recently, papers by Klippel and articles by Keith Howard re-examine the arguments that have raged since Klipsch and others before him (see pg. 9ff) have been saying since the inception of cone driver development at Bell Labs. In the 1980s, there was an attempt by Allison and Villchur (JAES) to pooh-pooh the audibility of modulation distortion, but as the Howard article above discusses, these arguments against the audibility of MD just don't hold up. If you say that modulation distortion isn't audible, then you're also saying that excessive amounts of jitter also aren't audible--take your pick.

Driver modulation distortion is proportional to diaphragm motion, i.e., at higher instantaneous and sustained SPL. The less motion that the diaphragms travel for a given output SPL, the less modulation distortion (AMD and FMD) there exists.
 

Ron E

Member
2002-06-27 10:41 pm
USA, MN
I see a lot of speaker projects designed with high efficiency in mind to compliment low powered SET amps - is there a benefit to high efficiency regardless of power requirements?


Pros:
Better, more realistic dynamics.
Less power compression
Horns can have good pattern control

Cons
Usually lighter cones which are less well damped and more prone to breakup and coloration
Generally larger boxes
Sometimes limited in the bass by low excursion motors
Horns have high order modes, honk and reflections

Most of the single driver solutions out there from the cheap (pioneer whizzer cone) to the expensive (fostex or especially lowther) tend to have an untidy sound due to all of the resonances and breakup that they are doing over most of their range.
 

Sk8Ter

Member
2011-08-09 12:12 am
Yes. The problem is that there are many people that don't like the answers.

The papers on this subject say that direct radiating loudspeakers produce many times more modulation and intermodulation distortion than if they were simply placed in a well-designed horn, i.e. higher efficiency loudspeakers equate to lower distortion, of the type that gives the "flat" or even "loud" sound that you might describe in loudspeakers.

Most of these papers were written by Klipsch, but more recently, papers by Klippel and articles by Keith Howard re-examine the arguments that have raged since Klipsch and others before him (see pg. 9ff) have been saying since the inception of cone driver development at Bell Labs. In the 1980s, there was an attempt by Allison and Villchur (JAES) to pooh-pooh the audibility of modulation distortion, but as the Howard article above discusses, these arguments against the audibility of MD just don't hold up. If you say that modulation distortion isn't audible, then you're also saying that excessive amounts of jitter also aren't audible--take your pick.

Driver modulation distortion is proportional to diaphragm motion, i.e., at higher instantaneous and sustained SPL. The less motion that the diaphragms travel for a given output SPL, the less modulation distortion (AMD and FMD) there exists.


do you own horn speakers?....I am in lewisville lets compare notes ya?

Lawrence
 
Pros:
Better, more realistic dynamics.
Less power compression
Horns can have good pattern control

Cons
Usually lighter cones which are less well damped and more prone to breakup and coloration
Generally larger boxes
Sometimes limited in the bass by low excursion motors
Horns have high order modes, honk and reflections

This is pretty typical response (which wasn't in the original question, I might add). Just like direct radiating loudspeakers, there are both good and bad implementations in high efficiency systems. The following "cons" listed above are generally from an age when analysis tools weren't available to design better loudspeakers:

1) Usually lighter cones which are less well damped and more prone to breakup and coloration, and
2) Horns have high order modes, honk and reflections

The second item above is something that Danley, Geddes, Delgado, Woods and many others (starting about 20-30 years ago) can help you with. Well-designed horns don't have those issues.

Even well-designed high efficiency direct radiator drivers (full range) typically have SPL limitations due to their DR design. In essence, they provide a subset of the benefits of high efficiency without what I'd consider the most important characteristic: the ability to reproduce recordings with their full dynamics--without introduction of audible amounts of modulation distortion.

Chris
 
Such a perk of living in the states. And what a great audio buddy in Cask! Whenever he talks, I'm impressed. Here in Winnipeg it's been 15 years since I've had a conversation on Hi-Fi. And when I did it was with a snake oil shop trying to sell me... anything.

Off topic but I am really floored by how few young people are into the hobby. Whenever I mention I am into Hi-Fi or speaker building people's eyes just kinda glaze over. I still hold out hope though..

Anyways I'm interested about the topic of high efficiency since my first build that I'm working on is a high efficiency design.

What's the difference between modulation distortion and THD? I'm confused about distortion. When I talk to Dr. Geddes he tells me that THD is not correlated significantly with perceived sound quality. He wrote a paper a while back where he identified the Gedlee metric which he said did have a correlation, but nobody else seems to give a hoot about it. But then there are members of this forum who are really concerned with THD plots, basically designing builds around it by using cone drivers instead of CD in horns. Sacrificing directivity/polar response for reduced THD. Curious to hear your view on this Cask. And anyone else's.

-Shawn
 
What's the difference between modulation distortion and THD? I'm confused about distortion. When I talk to Dr. Geddes he tells me that THD is not correlated significantly with perceived sound quality. He wrote a paper a while back where he identified the Gedlee metric which he said did have a correlation, but nobody else seems to give a hoot about it. But then there are members of this forum who are really concerned with THD plots, basically designing builds around it by using cone drivers instead of CD in horns. Sacrificing directivity/polar response for reduced THD. Curious to hear your view on this Cask. And anyone else's.

Hi Shawn!

I think that the Klippel reference above will give you a graphic on what the differences are between harmonic and modulation distortion. Nelson Pass has also written on the subject of modulation distortion--which he calls "the elephant on the dance floor".

I'd recommend talking to Geddes on what Geddes has written (...he can be fairly prickly on the subject, and he frequents this forum...). I think a lot of people care about the subject. I do. I believe that it's the higher order sidebands that are objectionable in listening.

Generally, it is the higher-order harmonics that convolve (as in modulate with) the original recorded signal to produce sidebands seen in the Klippel graphic that are audible. The human hearing system generally ignores the lower order side bands, but the upper or lower sidebands that stand further away from the center higher frequency (both higher and lower in frequency) can be heard as a thickening of the sound, and is responsible for loudspeakers that sound "loud" and "flat" when turned up to concert levels (see pg. 382 of Toole's book on this subject).

THD is easy to measure. It usually doesn't correlate well with listening trials, I've found. Geddes is right on the audibility of the harmonic distortion part, I believe. THD is typically used by manufacturers to cover what really matters: what harmonic, and by frequency. The real proof in the pudding would be dual tone testing using the Klippel/international standard testing procedures. I've seen zero results by manufacturers using these test procedures.

As far as discussions on various types of distortion types that are audible (HD, AMD, FMD, compression distortion, IMD, etc.), and at what level, there was a long thread on this forum that discussed that, recently, but I don't believe that the "killer study" has been performed to finally put the discussion to rest. Some university needs to do the study with real humans, and at over 80 dBC...without headphones, IMO.

Chris
 

MrJolly

Member
2014-04-23 9:54 am
Thanks for the info and links guys, plenty for me to read!


The human hearing system generally ignores the lower order side bands, but the upper or lower sidebands that stand further away from the center higher frequency (both higher and lower in frequency) can be heard as a thickening of the sound, and is responsible for loudspeakers that sound "loud" and "flat" when turned up to concert levels

Yeah that pretty much sums up what i was trying to describe, at the extreme end of the scale it's the same effect you get from small bluetooth speakers which 'sound' loud, yet what is being heard is lots of distortion.

Like most people i've come from a love of HiFi as a consumer, high power amplification is cheap and sensitivity is irrelevant. What has surprised me on the DIY front is the amount of people using horns and using PA/Pro audio style drivers, units i thought were only relevant in a field with the sound of a generator whirring in the background. Happy to find a bit of insight into why these drivers are used and what they can bring to the table for enthusiasts at home :)
 
Yes. The problem is that there are many people that don't like the answers.


Driver modulation distortion is proportional to diaphragm motion, i.e., at higher instantaneous and sustained SPL. The less motion that the diaphragms travel for a given output SPL, the less modulation distortion (AMD and FMD) there exists.

If you have two drivers of the same size, and one is much lower sensitivity than the other, for a given SPL output at a specified frequency, I believe that the cones will be moving exactly the same amount. The cone excursion will be the same. The difference is that the lower sensitivity driver will require a lot more power to put out the same SPL as the higher sensitivity driver.

But, since the stiffer, lighter cone of the higher sensitivity driver means that Fs is higher than the lower sensitivity driver, then we typically use much larger drivers when designing with high sensitivity drivers. In that case, if the cone is much larger in order to get a lower Fs, then for a given frequency and SPL, the cone of the higher sensitivity driver will move much less than an equivalent driver with the same Fs but much lower sensitivity because the cone of the lower sensitivity driver will be smaller.

Thus, less cone motion can be achieved, but the sacrifice is size, both of the driver and the box. In order to achieve lower FR with high sensitivity drivers, you need to go big.... sometimes really big.
 
I believe that the statement that you quoted above was talking about direct radiating drivers vs. horn-loaded drivers.

For instance, if you put a woofer at the mouth of a well-designed and well-implemented horn, you will need to move the horn-loaded driver diaphragm about 1/5th or less the distance to produce the same on-axis SPL on its output as the direct radiating case. That's what the real advantage of horns happens to be on this subject area, i.e., much lower modulation and compression distortion than typical direct radiating loudspeakers, due to significantly reduced driver diaphragm motion.

Chris
 
Hi, when MrJolly is mentioning jumping at thee and being present, he is prolly also pointing to high directivity, the same as low dispersion and beamwidth, and good impulse response. Short history of western hifi:

In grammophone and electron valve times, it was all hily efficient and beaming rather narrowly. Efficiency and directivity both rise with diaphragm size. It became hifi around, when Hiroshima and Nagasaki were destroyed. Sounding right but still not live, because one could easily hear it coming from a loudspeaker, from off. So stereo came up. Now people found out, that stereo sounds most engulfing and confusing --confusion can be a virtue--, when loudspeakers of wide dispersion are used. Hence small woofers and tiny tweeters were introduced. Frequency range was extended into kontrabass in order to sound more grounded, yet bass extension kills efficiency, too. Impulse response was sacrificed, too. For half a century it got more well-behaved and ethereal but less plain and cutting.

Use more diaphagm area, not all area located in one driver but rather evenly distributed among several ones, and cluster all drivers on a baffle not too small. Use flat filtering and sealed enclosures. The result shall be o.k. efficiency, moderate directivity and flawless impulse response. So you may either be satisfied with mono or find out, that stereo does not work with two loudspeakers, but one rather needs three or one special (quasi-coincident stereo) one.
 
I believe that the statement that you quoted above was talking about direct radiating drivers vs. horn-loaded drivers.

For instance, if you put a woofer at the mouth of a well-designed and well-implemented horn, you will need to move the horn-loaded driver diaphragm about 1/5th or less the distance to produce the same on-axis SPL on its output as the direct radiating case. That's what the real advantage of horns happens to be on this subject area, i.e., much lower modulation and compression distortion than typical direct radiating loudspeakers, due to significantly reduced driver diaphragm motion.

Chris

OK, that was not clear to me from his post, obviously. But yeah, of course what you say is correct.