Tweeters suffer from power compression at lower levels, due to their smaller coils lack of heat transfer etc at the low end. Wouldn't doubt that this causes issue starting with a only a couple of watts.
Yup, also why I choose an AMT over any metal dome. Why I also dislike most metal cone drivers, especially midrange/midbasses/woofers with their peaky sharp breakup modes.
Attachments
Is this a fact, or an opinion? i.e. is there any documented evidence for it? You would think a tweeter would be designed with a coil commensurate with its mass and expected displacement, and that market forces would demand that it actually work.
As probably the only person here who has done extensive measurements of thermal power compression, I have a problem with it constantly being linked with listener fatigue.
Thermal compression will simply increase the DC resistance of the various voice coils. It will create a frequency response change much like putting several (equivalent) ohms in series with voice coils. Net effect is a response curve variation with most impact where driver impedance is low (away from resonance and inductance rise). Tweeter coils are lighter (quicker time constant) and likely will get hotter (a conductivity issue) and see more compression.
So when you crank up the music someone will magically turn down the treble a bit. (midrange will go down some, but the mid treble will go down more). Will this lead to more listener fatigue, or possibly less?
David
Thermal compression will simply increase the DC resistance of the various voice coils. It will create a frequency response change much like putting several (equivalent) ohms in series with voice coils. Net effect is a response curve variation with most impact where driver impedance is low (away from resonance and inductance rise). Tweeter coils are lighter (quicker time constant) and likely will get hotter (a conductivity issue) and see more compression.
So when you crank up the music someone will magically turn down the treble a bit. (midrange will go down some, but the mid treble will go down more). Will this lead to more listener fatigue, or possibly less?
David
If it varies over a short-ish timescale I would think it would be fatiguing. But didn't Stereophile find that power compression at normal listening levels was a myth?
Hot Stuff: Loudspeaker Voice-Coil Temperatures | Stereophile.com
Last edited:
You think so? Someone turning the tweeter level down a fraction as the volume advances?
2dB less treble. Very fatiguing.
http://www.jblpro.com/pub/obsolete/4411.pdf
JBL 4411, one of my older designs with power compression measured at 85, 95, and 105 dB nominal levels. The actual compression realized would be tied to the sweep rate at which the curve was taken, but these were at a fairly normal pace as evidenced by the curve resolution.
http://www.jblpro.com/pub/obsolete/443035.pdf
Also shown here at 1, 10 and 100 Watt nominal levels for 4430 and 4435.
Regards,
David S.
Attachments
From John Atkinson: "Of course, it would be possible to worsen these figures by driving the speaker still harder, but there would soon be objection from both the speaker and my ears. More realistically, on the wider-dynamic-range material I more usually listen to, and at my habitual replay levels, the rise in voice-coil temperature and the concomitant thermal compression will be lower still. So I strongly suspect that, for most hi-fi users—those who don't habitually wind the volume control to its highest position and indulge in PA listening levels—thermal compression is a paper tiger. "
By the way, his measurement methodology is similar to what I described a few pages back.
David S.
By the way, his measurement methodology is similar to what I described a few pages back.
David S.
Indeed- I was lumping them together. I'd say the lack of <100hz makes for the boredom fatigue, as opposed to the "sore ears" fatigue which is related to gross misbehavior. I don't get very much listening in, only a minute or two at most, with "sore ears" systems, so for me, the longer term boredom fatigue is the more insidious one, which can come in from a lack of resolution/intelligibility (bad drivers or crossover/enclosure design), poorly performing frequency extremes or lumpy frequency response, anything that takes me further away from the live reference.
Sheer SPL, or too-low SPL, will do it too.
Maybe not "thermal" compression, but rather other non linearities resulting from pushing tweeters and midbasses hard. When i say it is unpleasant i don't mean it is fatiguing but rather results in an apparent collapse in the presenation. It's a cringe-inducing effect i have heard on smaller bookshelf speakers, especially with lower order crossovers.
I was talking about metal cone breakup in midrange drivers being mistaken for a tweeter issue. Of course soft domes are in breakup as low as 5khz by design.
None of my thermal work has been made public, I don't do that anymore (knowledge it a competitive advantage). But it is true that typical 1" tweeters can suffer very large differences in frequency response with SPL levels making them highly variable with the signal changes. Is this the source of fatigue? It could well be one of them. As I said, I do not believe that it is one thing, it is many.
Then let's see your measurements or simulations.
I have shown you several that I have taken. I have described the various tests that we used back at KEF. All support my contention that thermal power compression, while measureable, is a fairly minor effect even with lower efficiency speakers.
That it would not be audible and certainly not a large contributor to fatigue is a personal opinion, but I think the measurable magnitude and the nature of the change makes that a pretty safe opinion.
If you want to promote this as a new and significant distortion form, we need to see some data. Otherwise it is all supposition (on everyone's part).
David S.
I said "If it varies over a shortish timescale" it would be fatiguing i.e. a dynamic effect that loosely tracks the signal level. What was the timescale you were suggesting for (the mythical?) tweeter power compression to kick in and out?
People seem to think that a couple of dB is neither here nor there - I've heard that before - but if it was a couple of dB added onto, or taken away over the entire audio range above 2.5 kHz, then yes it would be quite significant. With my system I find that 1dB increments in baffle step correction are too coarse.
Earl, you have been part of several threads commenting on the aspect of power compression and thermal modulation and have gone to lengths to make comment on such issues, even recently. Anything I may have said is and was not taking your comments in vain or out of context. I have far too much respect for you on that front 🙂
Mike
As explained before, the the timescale is exactly related to the thermal time constant of the driver in question. From my measurements about 2 to 5 seconds for tweeters and maybe 5 to 10 for most woofers. That means, precisely, that the power input rectified and sent through an RC time constant equal to the thermal time constant would represent the time variation of temperature. As described earlier by Earl, this would be an "envelope" change rather than a "waveform" change.
Nobody said the response effects were inaudible. I would assume that, pushed hard enough, you would hear a subtle frequency response shift. At the same time I have shown at least 2 systems measured at high steady state levels to show the typical magnitude of the shift. My objection is the assumption that this is a typical cause of listener fatigue. For the most part all our knowledge about typical fatigue causes point to elevated treble, rough HF response, elevated higher order distortion, etc. A gentle hand on the treble control would not be one of them (in my opinion).
I have pointed out that a number of speakers have used light bulbs as protection devices. This is the same phenomonon conscripted into driver protection. As current to the tweeter increases a light bulb heats up and the rising filiment resistance limits the level increase. Proponents of this protection type like to quote how it seamlessly it does its job. A bulb filament can get pretty hot (hot enough to glow), so the net effect is much greater than typical thermal compression.
Interesting to note that John Atkinson in the Stereophile article was coming to similar conclusions about this being a Straw Man.
David S.
I am convinced by the Stereophile article, but I think that dynamic compression is fatiguing however it occurs. As is dynamic expansion. An otherwise hard core audiophile friend of mine is insensitive to high levels of dynamic range compression, whereas it drives me mad.
What if a speaker is overly-sensitive to one frequency range? (A speaker that can't do bass below 100Hz might fit this description.)
A transient containing the sensitive spectrum will stand out against all sounds that don't have strong content in that range. If the transient comprises a form of 'spectral sweep' (as is common) then the transient's volume envelope will be modified, possibly exactly like dynamic range expansion. Instead of an ear-caressing soft leading edge it may take on a much harder edge, causing sore ear fatigue..? In other words, if I hit a drum and reproduce it on a speaker that cuts off below 100Hz, it may sound awfully similar to the same recording played through a dynamic range expander - which sounds like a recipe for sore ears.
Last edited:
Modern recording using plenty of compressors and the net effect is well more than 2 dB compression. It may squash dynamic range and be generally undesirable, but it isn't always heard as distortion.
The biggest parameter in the audibility of any compressor circuit is the speed of attack and release. The time constants need to be slow with respect to the waveform, so that compression isn't coming and going every cycle. This would clearly lead to distortion and could be potentially fatiguing. If the attack speed is relatively slow relative to the signal frequency, then it becomes a compression of the waveform (a squashing of volume). Some amplifiers (NAD and McIntosh) have used compressors or soft clipping to squeeze waveforms and prevent the greater sin of hard clipping. They always trade off speed of reaction against waveform distortion. At NAD, I believe they left a few cycles of clipping in, so as not to rob the signal of its sense of loudness.
The thermal compression we are talking about would be very similar to any of these slow acting soft compressors. They would have a more gradual slope change (the JBL curves show at most 2dB change for 10dB input increase, a very soft knee). They are not limiting per se, the signal would always continue to grow, at least until thermally destructive levels were reached.
They are partially reducing dynamic range in a frequency dependent way.
David S.
I have done with some listening yesterday by modify may amp to make more fatiguing sound (I am playing on its phase related to feedback and damping).
As for cone breakup, it may not true as suspect, because it still exist in very low level and only low frequency. And by returning my amp to normal configuration with same speaker it has no fatiguing. I also trying TV (matsushita) speakers that used in sony with TDA2009 chip.
As for <100Hz, it also not true. It may happened on high damping amp that has annoying mid-high and good bass. Other external noise (like water, mechanical sound of my winding machine) also help my ears not fatigued by this annoying speaker sound.
Using 1K ohm resistor in series with loudspeaker always slightly reducing the fatiguing effect, but it need slightly increased amp volume too. As for heated driver, I don't think so, because some public address horn (toa, etc) are very hot in operation, and listening fatigue happened also in low level sound.
As for cone breakup, it may not true as suspect, because it still exist in very low level and only low frequency. And by returning my amp to normal configuration with same speaker it has no fatiguing. I also trying TV (matsushita) speakers that used in sony with TDA2009 chip.
As for <100Hz, it also not true. It may happened on high damping amp that has annoying mid-high and good bass. Other external noise (like water, mechanical sound of my winding machine) also help my ears not fatigued by this annoying speaker sound.
Using 1K ohm resistor in series with loudspeaker always slightly reducing the fatiguing effect, but it need slightly increased amp volume too. As for heated driver, I don't think so, because some public address horn (toa, etc) are very hot in operation, and listening fatigue happened also in low level sound.
