Pan said:
Thanks for the clarification. How should we compare the 10ms burst test with either pink noise tests or with worst-case music?
I cannot find a description of the burst test protocol (e.g. as used by Morel), but would be interested to see results for repetitive bursts over at least several seconds (to better represent music).
A single burst is 10J, so 1 per second is 10W with a crest factor of 100. For a typical pair of domes average of 10W might correspond to slightly over 90dB at ~4 m in a reasonably large room. Not astoundingly loud. I suspect that this sequence of bursts would show some, compression as the voice-coil temperature steps up at each pulse.
I think too that this is still an easier test than real music (where the crest factor is lower). A pink noise test as used in pro audio may be going too far though (certainly on the safe side).
I've never used the domes you name, but have used Morel (also give 10ms burst tests, and use Dynaudio technology), some Peerless and Seas units all of which seem about the same in power handling. Recently I've been using a compression driver (DE250 on XT1086) specifically for the reason that I was concerned about compression in domes - I was uncomfortable with a change in sound that I thought I was hearing in loud music. I did some testing that suggested that the effect I was hearing was not harmonic distortion.
I do not have fair, objective test results to present (too much work to do properly). The effect I thought I was hearing with (very) loud music is no longer present with the DE250. I use active crossovers, and because of noise at the moment I have a 20dB attenuator before the DE250. The power going to the coil is of order 10x lower and the effective thermal mass is larger by a similar factor. Thermal compression could be of order 30~100 times less.
I cannot prove that this was definitely the problem with dome tweeters, but I'm confident that it is no problem with the DE250. It is actually a bit scary, as it is very hard to tell how loud the music is now, without an SPL meter.
It seems to me that compression drivers of a given size are all much the same with regard to thermal performance, and dome tweeters similarly (provided they have ferrofluid), lets guess within factors of order 2 provided we stick to state of the art in each case.
In that case your criticism (which, BTW, reads as harsh and unacceptably agressive - whether you mean that or not), may be misguided. I suspect confusion arises from different test protocols, making the comparison of your favourite domes with compression drivers, which are tested for use in a tougher environment, very difficult.
Ken
Pan said:
there is more important argument - who really needs this? Who wants lifelike SPL in an apartment, in a room of <<100 square meters?
I remember someone has written quite recently on this forum (perhaps it was Dr Geddes): "I do object to live music being the standard as that is a narrow deffinition of the audio experince"
indeed!
and National Research Council of Canada (commenting the way they measure distortion in speakers) states:
"an SPL level of 90dB measured anechoically is very loud and considered far beyond normal listening levels"
http://www.soundstagenetwork.com/measurements/test_loudspeakers.htm
best,
graaf
Both sine and pink noise may be usable for comparison if the voltage sensitivity is taken into account between different drivers.
As you suggest it doesn't help us understand exactly how a driver will perform with music though.
I'm working on a method for evaluating thermal compression as real music is playing. Stereophile did somehitng like this and that indicated no significant thermal compression from a tweeter in a two way. The midbass suffered some compresssion though. However I don't remember what the time constant was in that Stereophile test.
I want to use fine resolution to see even eventual fast modulation of the voicecoil temperature. Still logging the voicecoil temp only tell us what's going on in the voice coil while the effective acoustic compression also depends on drive impedance.
I'm fully confident though that for most musiclovers a good quality dome tweeter is no problem regarding dynamics.
/Peter
As you suggest it doesn't help us understand exactly how a driver will perform with music though.
I'm working on a method for evaluating thermal compression as real music is playing. Stereophile did somehitng like this and that indicated no significant thermal compression from a tweeter in a two way. The midbass suffered some compresssion though. However I don't remember what the time constant was in that Stereophile test.
I want to use fine resolution to see even eventual fast modulation of the voicecoil temperature. Still logging the voicecoil temp only tell us what's going on in the voice coil while the effective acoustic compression also depends on drive impedance.
I'm fully confident though that for most musiclovers a good quality dome tweeter is no problem regarding dynamics.
/Peter
Graaf,
yes, and even though short peaks may be high in the top decade it is in the mid registers (100Hz to 1000Hz) the most power is with most music.
It's very enlighting to play a sine at 2-5kHz into a tweeter at 90-100dB spl and experience that.
Continuous high levels in the top octaves is something few people can stand... sans some headbangers possibly.
/Peter
yes, and even though short peaks may be high in the top decade it is in the mid registers (100Hz to 1000Hz) the most power is with most music.
It's very enlighting to play a sine at 2-5kHz into a tweeter at 90-100dB spl and experience that.
Continuous high levels in the top octaves is something few people can stand... sans some headbangers possibly.
/Peter
experiment
I took an old Peerless 810921 (ferrofluid) tweeter, and did some testing with 2kHz-20kHz band limited tone bursts.
1000W/10ms (10J) seemed a bit scary so I tried 100W 30ms (3J) but repeated at 3 Hz.
Unfortunately the 10W average was too much for the tweeter and it failed before I had a good set of current data from which to work out the resistance rise. (The lead out wire probably over-heated where it was not well cooled as voicecoil looked otherwise OK apart from the broken wire.)
The highest temperature I recorded (from the resistance change) was around 50C (about 2dB compression). I think that was probably louder than I'd ever have listened to full bandwidth music. (Tweeter was in a small box, and was not very loud outside the box.) I think the time constant for temperature rise was in the order of 10 seconds, but again a better measurement would have been needed to tell for sure.
This leads me to wonder if the problem I heard with these tweeters was perhaps them running out of excursion after all (perhaps due to too-low crossover).
I don't recommend experiments like this if you want to use the tweeter afterwards.
Ken
I took an old Peerless 810921 (ferrofluid) tweeter, and did some testing with 2kHz-20kHz band limited tone bursts.
1000W/10ms (10J) seemed a bit scary so I tried 100W 30ms (3J) but repeated at 3 Hz.
Unfortunately the 10W average was too much for the tweeter and it failed before I had a good set of current data from which to work out the resistance rise. (The lead out wire probably over-heated where it was not well cooled as voicecoil looked otherwise OK apart from the broken wire.)
The highest temperature I recorded (from the resistance change) was around 50C (about 2dB compression). I think that was probably louder than I'd ever have listened to full bandwidth music. (Tweeter was in a small box, and was not very loud outside the box.) I think the time constant for temperature rise was in the order of 10 seconds, but again a better measurement would have been needed to tell for sure.
This leads me to wonder if the problem I heard with these tweeters was perhaps them running out of excursion after all (perhaps due to too-low crossover).
I don't recommend experiments like this if you want to use the tweeter afterwards.
Ken
Pan said:Graaf,
It's very enlighting to play a sine at 2-5kHz into a tweeter at 90-100dB spl and experience that.
Continuous high levels in the top octaves is something few people can stand... sans some headbangers possibly.
/Peter
possibly headbangers!
few seconds and the head just ... BANGS!
Pan said:
Certainly many live experiences aren't all that loud, but that's really immaterial. I don't want a system that works ok 95% of the time... those moments when I do need huge peaks, they better be there. As far as the drum set goes, I'm definately on the other end of the spectrum... that impact is one of my favorite parts of seeing a live band. Since someone will probably reply with something about hearing loss, I better through in the obligatory "one has to be careful though... I do monitor exposure time when really cranking it up." As for the answer to the question, when you account for compression knocking down the extreme peaks in almost all recorded music, I've found that a system capable of hitting 112-115 dB peaks in the 80-10k Hz range per channel and 125 dB peaks 30-80 Hz in the mono sub is just about right for being able to reproduce at a realistic level anything I (want to) throw at it. I added (want to) because there are amplified concert experiences that are much louder than I want to recreate.
Pan said:
From what I remember, with the handful of drivers I tested the compression was about 1 dB from 1.5-8 kHz... but remember that's 120 dB RMS at 1m... given my goals as stated above, the compression driver never sees that much continuous power. As far as distortion, with a sine wave at those levels THD was on the order of 1-5% and almost all 2nd order. Everything I know indicates this is negligble.
For fun I've done a few (very nonrigorous) experiments with cranking a system up outside and listening to it from varying distances to keep the listening volume about constant while varying the playback level and the compression/distortion at high volumes seems negligible so far as I can tell. This type of experiment doesn't really mean much of course.
Pan said:
The results you have here are different from what I have personally seen, but I'm not discounting them. I just have a hard time believing the .5 dB compression at 105 dB for the dome tweeters... aren't those about 90 dB 1w/1m? That would mean you'd be throwing about 30 watts RMS into it to get 105 dB... hell, there are 3" coil midwoofers that compress .5 dB at 30 watts RMS. I have burned up tweeters putting that much continuous power into them before.
Pan said:
You have a point there. I would not feel comfortable putting that much peak power into the tweeter but it appears one can get around the mechanical limits. On a side note, I was thinking about what a system that used such peak capabilities would look like... you mentioned some testing with 10 kw peaks before. I'm picturing a setup with something like a QSC RMX-5050 bridged to each tweeter... hope your electronics don't have turn-on pop
What makes you think "accuracy and detail" (however YOU define them) can not be reproduced with well designed compression driver waveguide combination at either high or low SPL levels?
What leads you to think this?
Of course they can. This is likely because the dome tweeters are producing more distortion at high SPLs than YOU can handle.
If the domes were distortion free, you would not notice the high SPLs.
You've got it backwards.
Really clean systems will wreck your hearing because they don't 'sound loud' at very high SPLs.
Rybaudio said:
You have mentioned "recorded music". What is peak/average ratio of the CDs You "throw at it"? When You hit 115 in peak it means that average is...?
And at the same time for National Research Council of Canada "an SPL level of 90dB measured anechoically is very loud and considered far beyond normal listening levels". This I can confirm from my own experience. 100 dB in peaks is an absolute max I have experienced. And I didn't like it
It was rather frightening experience in a room of 28 square meters.It seems that people have very different SPL needs.
I think that it's OK.
And can be explained.
Human hearing is sensitive to relative SPL so the max SPL requirement is a function of noise floor. Also differences in hearing sensitivity can be in the range of 15 dB within the norm of "normal hearing".
Not taking into account those differences produces misunderstandings and pointless and also endless discussions.
We have different needs and it's perfectly OK
nobody is "better" from this perspective or nobody has "better equipment" because he can achieve "realistic level" with whatever he "throw at it" (="all other have worse")
realistic SPL level in music reproduction at home is something subjective
it is not the same as measured SPL level of a given live performance
and even if a living room can be with extensive acoustic adaptation and with help of high power equipment turned into a stadium not everyone of us needs it
best,
graaf
graaf said:
At the listening position the A-weighted RMS is at most 105 dB and usually in the 95-100 dB range for full volume listening. Up to this point I've just been monitoring this to keep exposure time safe so I have been using A-weighted (side question- is this the right thing to do?). This really doesn't sound that loud if the system isn't distorting.
graaf said:We have different needs and it's perfectly OK
nobody is "better" from this perspective or nobody has "better equipment" because he can achieve "realistic level" with whatever he "throw at it" (="all other have worse")
Maybe I don't understand what you are getting at here or maybe you're just trying to make people feel good, but it sure seems to me that if one introduces a value system in which achieving life-like SPL is good, then all other things being equal, a system that can do that is better than one that can't. If that is not an aspect of system performance that is highly valued... well then it doesn't matter. Isn't that obvious/tautological?
graaf said:
If absolute realism isn't your goal, that's fine, but to relativise a concept whose standard definition incorperates the notion of absolute just so everyone can feel good about it being applied to themselves seems rediculous. The bottom line is that a live performance happens at a certain SPL and in order to mimic the experience of that event one has to play it back at a similar SPL. Of course, processing done to the recording can screw this up along the way, but I am talking ideal case here.
Rybaudio said:
Thats is something a good dome can do.
Well depending on the time here that sounds like good performace. The compression driver in the Summa showed a compression of aprox 0.5dB after 5seconds at 105dB 1 meter.
Sounds good as well. I think those are tolerable distortion levels at that spl, have not done any testing at those levels myself though.
The following is from one of my posts in the "the problem with hifidelity thread";
I have measured good 5" midwoofers that have worse thermal performance than good tweeters. One should also remember that the demands rms powerwise always is higher in the registers below 2kHz.
You have a point there. I would not feel comfortable putting that much peak power into the tweeter but it appears one can get around the mechanical limits. On a side note, I was thinking about what a system that used such peak capabilities would look like... you mentioned some testing with 10 kw peaks before. I'm picturing a setup with something like a QSC RMX-5050 bridged to each tweeter... hope your electronics don't have turn-on pop
I corrected myself after that.. 3kW was the correct transient testing for those tweeters. 10kW was for midbasses. Dynaudio test all their drivers (as I understand it) with 1kW.
/Peter
Rybaudio said:
I believe that it is not only matter of distortions
It is also a function of room size and room acoustics.
You have to have a big room or an acoustically dead room that appears to be larger than it is.
Otherwise the experience is frightening not because of distortions but because of mismatch of SPL and room size. You can try to put a stadium SPL into a living room but unless the room is really big and/or its own acoustics is suppressed that experience won't be anywhere near "realistic"
It is just like heavy metal band playing at a jazz cafe
Rybaudio said:
as I said above achieving life-like SPL of big auditorium in a small auditorium (even big listening rooms are small as auditoria) has hardly anything to do with realism
just like heavy metal band playing at a jazz cafe
with exceptions maybe of really big, acoustically dead room
that doesn't mean that realism is not achievable in a typical listening room
what You can realistically have in a typical listening room as a small auditorium is a realistic perspective on a big auditorium - the walls behind speakers disappear and You can "look into" a big auditorium eg. a stadium, without trying to put it into the room
Rybaudio said:
realism is my goal but it is realism achievable in a realistic circumstances
and also I insist that realism of sound reproduction is always subjective - it is always "realistic for" someone and not so for someone else
for me the objective goal is accuracy, acccurate is accurate, not "for someone", it is or it isn't
it is the standard definition of "accuracy" that is absolute - not of the "realism"
and accuracy is about reproducing accurately what is in the recording, that is it
and what is in the recording? compression, 60 dB of dynamic range, 5<25 dB between average and peak
also for other people (including me) there are more important things in realistic sound reproduction than life-like SPL and their achievement often conflicts with achievement of life-like SPL
I mean preserving the time structure of life sound, I am talking about time domain
Rybaudio said:
ideal cases! exactly
why not talk about real cases?
why not?
what about some realism?
the truth is that what we need to mimic dynamic range of live music is a recording with dynamic range of live music, the suitable listening room and high power equipment
the first element is missing, the second is rather rare case
all we have is high power equipment
I ask - what is it for when the first two elements are missing?
the above is MY point of view
we differ and as I said it is perfectly OK
and when Dr Geddes writes "I do object to live music being the standard as that is a narrow deffinition of the audio experince" than who am I to argue?
On the contrary - I strongly concur with regard to the question of "life-like SPL"
best,
graaf
FrankWW said:
I haven't heard a compression driver waveguide combination that could match the SQ of a good dome tweeter yet (at what is normal hi-fi listening levels to me).
FrankWW said:
I can listen to certain music at very high SPL and the only limit is the excursion on my woofers, not distortion, however I prefer acoustic instruments and like to listen to them at more realistic levels.
I agree that clean systems don't sound loud at high SPLs, ironically the tweeter that hurt my ears the most was a compression driver but then you get bad domes as well.
André
Pan said:
At what distance? When I refer to playback levels I am always referring to the listening position (where it counts). Most of the time I sit about 3m away from the speakers... that translates to peaks of 122-125 dB at 1m. Personally I think the 120 dB claim is pushing it... 125 I highly doubt.
Pan said:
This is so different from what I have seen out of dome tweeters I have to ask how you measured the resistance?
Pan said:
That may be the case, but I am talking 15" pro midbasses/subs here. Take a look around Beyma's site for example... they have fairly basic power compression charts in a lot of their data sheets. Most of the numbers there are similar to what I've measured. I just find it hard to believe a 1" thin voice coil can dissipate heat like that.
" Well depending on the time here that sounds like good performace. The compression driver in the Summa showed a compression of aprox 0.5dB after 5seconds at 105dB 1 meter."
Hello Peter
That doesn't sound right. If you look at small format drivers many are in the 110db @ 1 meter with a suitable horn/waveguide. Lets say for arguments sake that it works out to be 111db 1 watt 1 meter. If you drop 6db in output to 105db then your are running about 250mw into the compression driver from say 1K-4K where their efficiency peak lies.
Depending on the compensation you will be running more power in the upper octaves with a noise source. If you example you have 13db of passive compensation then using the 250mv reference you have 5 watts worst case in upper octaves where there is no attenuation from the passive network.
That's not much power at all. I would expect almost no measurable power compression at levels like that. Even in a dome for that matter.
Rob
Hello Peter
That doesn't sound right. If you look at small format drivers many are in the 110db @ 1 meter with a suitable horn/waveguide. Lets say for arguments sake that it works out to be 111db 1 watt 1 meter. If you drop 6db in output to 105db then your are running about 250mw into the compression driver from say 1K-4K where their efficiency peak lies.
Depending on the compensation you will be running more power in the upper octaves with a noise source. If you example you have 13db of passive compensation then using the 250mv reference you have 5 watts worst case in upper octaves where there is no attenuation from the passive network.
That's not much power at all. I would expect almost no measurable power compression at levels like that. Even in a dome for that matter.
Rob
The industrial limit for sound is around 85 Db. Beyond that you need protection. (ear condum?)
I live in an apartment and i am older , i have no need for 100 plus Db. Every time i have gone to a resturant or other lately and the sound levels get so great that i cant hear someone talking at my table it gets irritating. Please save your hearing, one day you will realize why.
Insofar as "live" performance , most performers turn the volume up past the distortion limits of the amplifiers they are using. I guess the phrase of " if you cant perform, turn it loud" applies. Note! this does not apply in all cases.
ron
I live in an apartment and i am older , i have no need for 100 plus Db. Every time i have gone to a resturant or other lately and the sound levels get so great that i cant hear someone talking at my table it gets irritating. Please save your hearing, one day you will realize why.
Insofar as "live" performance , most performers turn the volume up past the distortion limits of the amplifiers they are using. I guess the phrase of " if you cant perform, turn it loud" applies. Note! this does not apply in all cases.
ron
ronc said:
If you are referring to OSHA's standards, then that is precisely what I use for my exposure time guidelines. Their maximum exposure time for 85 dB is 8 hours!... I never listen to a stereo that long in one sitting. My rough guideline is, for A-weighted RMS SPL:
105+ dB: 2 songs max
100 dB: half hour
95 dB: hour
less: rarely listen long enough to make it a concern.
Unfortunately I can't crank it up hardly at all living an apartment now so these don't come into play much these days.
The real problem is stored energy. If you calculate the sum of stored energy over the first 0.3ms or so and add that to the original response, it will look really ugly.ronc said:
I have found that if the stored energy situation is improved, I can crank the volume much higher and not feel irritated. This effect of stored energy probably effects sound reproduction much more than other forms of distortion in most hifi drivers.I've sat on top of tanks without ear muffs as they were target practicing, and can say for sure that no audio system is going to reach levels like that, let alone how realistic it can sound.
Unfortunately I can't crank it up hardly at all living an apartment now so these don't come into play much these days.
In your later years you will appreciate this. No facts here but IMHO listening to distortion( at high levels) over a long peroid of time will have more effect of wear and tear on the eardrums.
(good masters thesis or even a PHD there).
ron
In your later years you will appreciate this. No facts here but IMHO listening to distortion( at high levels) over a long peroid of time will have more effect of wear and tear on the eardrums.
(good masters thesis or even a PHD there).
ron