Why Do Most Speakers Suck at Dynamics?

Discussion in 'Meadowlark Audio Forum' started by Prime Minister, Apr 30, 2018.

  1. Prime Minister

    Prime Minister Site Owner Staff Member

    Wow. Another interesting idea there. I guess thats where the ultra high end guys have it easier, in a sense. They know the stores their gear will be sold in, know they amps it will likely be matched to, and they can design accordingly.

    One would think this would lead to a number of really beautiful, musical speakers. However, my experience has been that speakers in the silly high dollar brackets usually sound more impressive, then they do beautiful. A "wow, doesn't that sound fantastic" reaction, rather then a "man, that's really lovely music" one.

    So is that all about building to an expectation? Needing a product to WOW! people, rather then seduce them?
    MikeT. likes this.
  2. 4-2-7

    4-2-7 Smart Ass Member

    Martin Logan
  3. 4-2-7

    4-2-7 Smart Ass Member

    BTW sometimes it's not the speakers sucking, I'm a firm believer in head room and having amps that will over power the speakers. If you generally have 100 watt amps and 200-400 watt speakers your going to have a lot of sucking to deal with. The dynamics of a speaker changes with power behind it.
    Celt and JimPA like this.
  4. Amplifier power is cheap. USE LOTS!

    While developing our tiny 5.5" Little Wing, I apprehended that 2x125W per speaker would suffice. Nope. Kept banging the stops on the amps. Went to 2x250W. So....tee, hee, hee....a pair of skinny little bookshelves sporting 1KW. And still we can watch the meters peak into the red.

    For Nightingale, I found that stepping up from 4x400W per side to 5x400W and doubling up on power supply made all the difference; a sense of strain at sustained high energy went away. Pleasant, indeed. Especially if you like concert level Jeff Beck.

    Ideally, you want enough power to bang the stops on the speaker before you hit the limit on the amps. At the extreme, you want listener to reach the limit before the stereo ;-)

    As soon as I get the chance I'll start on the promised Dynamics and Punch exposition. Much fun.
    rogerfederer and 4-2-7 like this.
  5. dogscanskate

    dogscanskate Junior Member

    Are there limits to how much dynamics a speaker can put out though? Are they not also limited by the recording instruments such as microphones, compressors and the hardware such as how large must a record groove have to be to actually be capable of really large dynamics to feed the signal through to the speakers?

    If vinyl were not compressed, how many minutes could we possibly be able to fit on a 12''. Doesn't that help builders choose parameters that account for what we actually get instead of what the instruments can give out in a venue? Just asking...
  6. 4-2-7

    4-2-7 Smart Ass Member

    The old saying here is crap in crap out.

    If we talk about records, cartridge, arm, turntable as the source, there is a lot there that needs to be good before going down the line. But you want to give it a unfettered short signal path with the power to handle all the transient peaks in the music. If we clip the signal we are not getting the complete dynamics that's present in the recording or I should say the mastered pressing of a specific record. If a amp uses it's stored energy as well as it's continuous production to move a driver for a burst of transient peaks, it will need to recover to do it again. Other wise depending on the music, the needs of the driver and rapid succession of peaks in the music you'll start losing dynamics from clipping. Now I'm not talking about diving the amp really hard and "hearing the bad clipping we know we can get. But subtle clipping happens without one really hearing it, but we might sense it in low dynamics because we're not getting all of the music at that point.

    Something like this Awesome recording and on a really good mastered record will demand a lot from a system to sound it's best. You can just hear the burst of 20db throughout this record.

    dogscanskate and JimPA like this.
  7. dogscanskate

    dogscanskate Junior Member

    My point is that before you feed a speaker with information, the signal has already been (heavily many times) modified. To store it in a record groove or tape or cd, compromises must be made concerning these exact dynamics. There are physical limitations that apply such as a 2'' 30 ips tape will secure more information than a 1/4'' 7.5 ips tape can in the same span of time.

    When designing a speaker, the recipe changes for every design. I see a lot of validity in the statements posted above but we are still designing for a flawed source. Since I have not tried all speakers, my interest is piqued. I don't believe that we are talking about low frequency dynamics because of the compromises made in the recording studio, I may be wrong however.
    MikeT. likes this.
  8. MikeT.

    MikeT. Senior Member

    This is such a great thread. I’ve always thought about dynamic response in terms of full spectrum content but am thinking a bit differently now. The attack of single or multiple violins, from a silent rest, could also be considered, yes?
    Where my speakers don’t do huge bass swings, I love what they do (head-shaking realism) when reproducing attack and detail from mid-bass up.
    Thanks to @Pat McGinty and all for sharing.
    dogscanskate likes this.
  9. Dynamic Range is a great and very fun subject that serves as a nice foundation for a discussion of the larger subject of Dynamic Fidelity. And you are completely correct that, starting with the recording and ending at the speakers, net DR will be limited by the element in the pipeline with the least range. That's a key point.

    Since audio recording, media distribution and playback began with Edison's wax cylinder, DR has been a dominant limitation; it has been the single biggest reason why audio does not sound the same as reality. The issues of bandwidth and (most) distortion on the recording and media sides came well to heel with the LP. And our better stereos have done a pretty decent job of keeping up with the DR of LP and to a some extent CD.

    The benchmark, here, is that the DR of human hearing is about 140dB. LP can do roughly 75dB and CD is limited by it's 16 bit word size to 96dB. Now our 24 bit datastream finally hits the 140dB ideal that matches our senses.

    The DR limitation imposed by LP and CD reduced the DR target for both recording guys and reproduction guys. Who needs either a recording or a stereo that can swing 140dB when the intervening media 'hits the wall' so much sooner? So the whole chain has pretty much settled in under those limits.

    Today that barrier has fallen. Now it's up to the source guys, and the playback guys to catch up. It's a chicken-and-egg thing: it will be much easier for the source guys to swing the entire gamut because they already have the gear. But what's the point if the we on the playback end can't handle it? Fantastic recordings will go entirely unappreciated.

    So the sharp guys in playback are getting very busy knowing that we simply cannot hit the new DR target the old way. The bar is being raised in front of us. Happily the technologies we need to leap over are coming along nicely.
  10. I think you hit on what I was about to ask with sources..

    I get reasonable dynamics with vinyl and CD.. But when I stick on a BluRay over the same system many times there’s an abundance of dynamics and I’m hunting through the menu to enact dynamic compression as it’s way too much. Sometimes those explosions are way too loud after relatively average levels of speech.

    To me I’d rather have that dynamism with music rather than film but the music industry I doubt will ever embrace it. In fact CD quality is now the gold standard for audio quality and audio dynamics as for most people digital is streaming. The strides in this area of source dynamics don’t have seemed to have made it into music reproduction for the mass market because few people outside of people like us care that much. Convenience over sound quality I guess. I feel like source matters more than equipment and we already have the tech to do it. There’s just no market for it.
    Last edited: May 13, 2018
    Fran604g likes this.
  11. When I did my first few attempts at HT I thought there was a tendency by mix engineers to exaggerate the dynamic contrasts. But now, having done a few serious builds that actually meet the (pretty darn) comprehensive industry standards I see what's going on. The systems are surprisingly robust. For instance we're working on a Dolby 7.4.4 Atmos system that comports with merely mid level specs that uses 35 transducers and 14KW of amplification in a 20x30'. The engineer who's mixing a film knows what that machine will do and takes full advantage. Naturally, when you try to push that soundtrack thru a lesser system you could can get an annoying result.

    Both the HT guys and the Pro guys are way ahead of the 2 channel guys because they have some say over what's downstream and because they're constantly aiming higher. All along, the 2 channel guys have had to mix for environments like radio and auto sound. In both cases compression is the only way to get good results especially considering that most playback systems stink. As a kid I listened to 60's Rock on AM radio. Since the service diameter of the transmitter was limited by signal minimum amplitude, the best results required gobs of compression. For auto sound, since the ambient noise floor is so high, naturally dynamic music doesn't work; compressed music actually sounds much better. So compression sells.

    That's not to say there are not plenty of recordings with fantastic dynamics.

    It's a tantalizingly funny thing - you don't really get to find out just how dynamic any recording actually is until your system exceeds it. I'm presently finding that recordings I've used as references for decades have had their dynamics concealed from me by the limitations of my earlier systems. And I don't know much more is "hiding".
  12. McGinty's Laws of Speaker Engineering

    Briefly, as a predicate to addressing the subject of dynamics in speaker design, I'd like to put these ideas on the table as a cornerstone.

    Murphy's Law and its many correlates reverberate through every branch of engineering. Their humor informs us that our problems are our gifts.

    To wit, here are my humbly minor tangential derivatives.

    1) In order to get an attribute you want you must give up another attribute that you want.

    2) Many relationships involve adverse square functions. Each time you ask for twice of something it hurts you times four. Each time you ask for twice of two things it hurts you times 16. And so on.

    3) Most of the trade offs include money. And #2 accelerates that.

    Next up: Dynamic Linearity
    Prime Minister and dogscanskate like this.
  13. Dynamic Linearity

    Ideal Dynamic Linearity means that across the band of interest, within the amplitude envelope of interest, each time you double the input you double the output.

    For us that means from, say, 20Hz to 20KHz from 0dBSPL to, say, 120dB room there is a 1:1 relationship between input amplitude and output amplitude. This makes a three dimensional transfer function, frequency/input/output that, when plotted, forms a flat plane splitting the I/O axes at a ratio of 1:1.

    When we examine real results we see a curvy shape that falls below the ideal plane, very badly in some places, which is lucky for us; we have plenty of work to do.

    So, what gets in the way of ideal dynamic linearity? Numerous things. The two 'biggies' are thermal compression and motor force nonlinearities. For simplicity's sake I'll restrict the story to woofers. But tweeters are no less interesting.

    Next up, Thermal Compression.

    perfect dynamic linearity.jpg
  14. Thermal Compression

    Surprise, surprise...we're facing an adverse square function per Law #2: each time you double the current flowing thru the voice coil you quadruple the heat generated. Impedance rises linearly with heat so as current increases output falls off and fast. Complicating things, the time constant is long; think of the wires in your toaster, it takes a while to heat up and a while to cool.

    Plus the crossover's transfer function is modulated by the changing impedance. For a low pass filter, corner frequency rises with impedance. So heat = glare. Plus both filter Q and total system Q wobble with impedance. Ugly stuff.

    The trick is not to heat your coils, a tall order considering the typical woofer turns 2% of the power into sound; the rest goes to heat. Since the idea is to not push in power faster than the heat can dissipate you can just refrain from juicing it.....but phooey on that, what we really want is superior performance. To that end we have four main avenues for expanding the envelope: convection, conduction, diffusion and multiplication.

    The first means better venting; a common method is to vent thru the back plate via a hollow pole piece thereby turning the moving system into an air pump. There are several additional ways to move more air past the coil such as a perforated former and elevated spider. The pic shows a driver with all of these features (plus, look closely, an extra tricky strategy for mechanically isolating the motor from the basket!).

    Conduction involves using the right materials, often lightweight metals like aluminum or titanium for the bobbin and sometimes for the cone. An aluminum cone can serve as dandy radiator. Some of the newest woofers take full advantage of that idea.

    Delightful and, ha ha ha, costly: heat can be diffused or 'spread out' by increasing the diameter of the coil. Then the magnetic circuit needs to grow even more to deliver adequate flux density. (We use large coils wrapped on metal and heavily vented because that's what it takes to withstand the LF boost we're shoving into them via digital equalization.)

    Last, and my favorite, multiplying the number of drivers divides the heat.

    Obviously each of these schemes comes at a price and obviously we want them all. Right away this turns into a 'run-away' problem, money wise. See Law #3.

    Sad but true: typical speakers are pretty weak on all of these things. The limiting factor is often cost, a $4K/pr. retail figures backwards into about a $100 woofer. Most designers are going to look for low Fs, a clean midrange and a smooth roll off to minimize crossover complexity before they begin to worry about how it scales with power. Anyway, good dynamics aren't worth a damn without proper performance in those other areas. So most of the time it's fourth or fifth on the list.

    Attached Files:

    MikeyFresh likes this.
  15. Prime Minister

    Prime Minister Site Owner Staff Member

    Hmmmmm. I never, ever considered the factor of heat. Very, interesting.

    Multiple drivers. Is that why they have become so popular? Multiple voice coils to dissipate heat through?

    Always so much to think about in your posts, Pat!
  16. Prime Minister

    Prime Minister Site Owner Staff Member

    And further, wouldn't cone weight be an issue? In cars, unsprung weight in suspensions is a huge issue. The lighter you make the wheel and tire combo, for instance, the easier it is for the suspension to do it's thing.

    How about aerodynamics? Yes, we want to move air, but is air resistance working against the driver a good thing?
  17. Yep, heat is a problem. You've never cooked a woofer coil?

    Digression: daughter Meagan who currently works with me recently admitted that, during her wilder teen years, the only Meadowlark speaker she was unable to intentionally fry were Nightingales. I do remember coming home to find seized woofers quite often but accepted it as karmic payback for what I did to my father's stereo. And cars.

    Anyway, yes, heat is a problem wherever it occurs in the signal path. Early on I noticed that shelving resistors in tweeter circuits were heating up. Under load, the rising resistance of the resistor sums with the simultaneously rising resistance of the voice coil thereby diminishing output and lowering the filter's corner frequency. Heat + more low frequency = blown tweeters. So we multiplied the number of resistors to spread out the heat plus heat sinked them.

    We ended up going with Caddock resistors because they are carbon on a ceramic block in a TO220 package designed for quick heat transfer to a radiator. It was really surprising how much that improved performance.

    Yep, multiple drivers solves many problems. But that idea does make for problems in two way speakers.

    In my next post I'll describe another nasty source of compression; one that also leads us to the multiple driver solution.
  18. My best guess - in a car, unsprung mass works against keeping the tire in contact with the road by tending it to go ballistic? The tire is elastic so you're dealing with a mass/spring resonant system - so more mass takes stiffer springs and stiffer damping which interferes with nimble suspension articulation?

    Maybe too, at speed, there's also a gyroscopic effect made worse by more mass?

    You're right about mass in a driver's moving system. F=MA so you want max F and min M to get max A. The car analogy works here; light car + plenty of power = acceleration.

    The trade off in drivers is that low M also means high resonant frequency - just as it does in a bell - which works against you for getting the low system F3 you want for a nice bottom end. So in conventional speakers you need to strike a balance. Happily, with modern digital PEQ we can sidestep this restraint :-)

    For cars you're trying to decouple from the air so that less force is needed to move thru it. For transducers the air IS the load so coupling with it is the whole idea.
  19. Mechanical Compression

    This is a subject I've never seen discussed on the retail side of things, but it's fundamental to understanding how woofers behave.

    Because the motor works by moving a coil against a static magnetic field, and since the amount of that coil inside the sweet spot of that field diminishes with excursion, motor force falls off both fore and aft of center. And as the moving system approaches its mechanical limits the elastic components are also stretched tight. This graph shows the force/position performance of one of the finest 7" woofers.


    You can see that motor force is kinda-sorta linear for the first few millimeters then it falls off quickly. So, for small excursions you get pretty much double the output when you double the input. But ask the woofer to really stroke and you'll see a falling relationship. This means that the entire band's dynamic linearity is modulated by coil position. When you ask for either low bass or big output you are asking for increased stroke which decreases the percentage of time the coil spends in the sweet spot therefor compressing everything including the mid band.

    Now you can see why a designer who is greedy for a low F3 will be trading away dynamic performance. This is waaaaay too common. Frankly, I think many designers don't really understand this. Too, some customers are just 'asking for it' by comparing F3 specs when making their purchase decision; they have no idea what they're trading away.

    Thermal Compression sums with Mechanical Compression. So all of the shortfalls add up. That's why so many speakers suck at dynamics. There are other reasons, but those are the big two.

    Doing it right takes big linear stroke which is expensive because it requires powerful magnets and tricky to execute magnetic circuits. Fortunately, the best driver guys know that both Thermal and Mechanical Compression need be addressed together because their effects sum. So they naturally go hand-in-hand as quality and price ascend.

    Multiplying the number drivers divides their excursion duties, ideally keeping all drivers in the linear zones. And cool.

Share This Page