Dr Berlin on key Hz
"The portion of speech energy that carries much of the information that makes speech understandable is concentrated in the frequencies between 300 and about 4000 Hz. "
from a Professor of Hearing Science, re hearing loss and its impact on communication www.capitolent.net/berlin.htm
in that range ~ avoid XOs or lumpy drivers .
"The portion of speech energy that carries much of the information that makes speech understandable is concentrated in the frequencies between 300 and about 4000 Hz. "
from a Professor of Hearing Science, re hearing loss and its impact on communication www.capitolent.net/berlin.htm
in that range ~ avoid XOs or lumpy drivers .
Some more on the voice 'target'
Excerpts:
The band of the telephone was for economical reasons confined to 300-3400 Hz, and this, of course, was based on the observation that it suffices for high-intelligibility speech communication.
AM radio, extending from about 100 - 4000 Hz, was for decades regarded as acceptable even for transmitting music.
And some IMO somewhat conflicting statements:
- consonants are about as important as the first two formants of vowels, as the former essentially are represented above about 1800 Hz, the latter, below
- with respect to intelligibility of speech, 700 Hz appears to be the "most important" frequency
- from about 400 to 2000 Hz . . the auditory system's dominant region for speech intelligibility
From another technical paper:
www.mmk.e-technik.tu-muenchen.de/persons/ter/top/dominant.html
Cheers
Excerpts:
The band of the telephone was for economical reasons confined to 300-3400 Hz, and this, of course, was based on the observation that it suffices for high-intelligibility speech communication.
AM radio, extending from about 100 - 4000 Hz, was for decades regarded as acceptable even for transmitting music.
And some IMO somewhat conflicting statements:
- consonants are about as important as the first two formants of vowels, as the former essentially are represented above about 1800 Hz, the latter, below
- with respect to intelligibility of speech, 700 Hz appears to be the "most important" frequency
- from about 400 to 2000 Hz . . the auditory system's dominant region for speech intelligibility
From another technical paper:
www.mmk.e-technik.tu-muenchen.de/persons/ter/top/dominant.html
Cheers
Yup!
A E U I O (and all European AE Å Ä Ö U versions) are in low range and fundamental. Even frequency response is needed and freedom from gross resonances especially those from the room acoustics.
Without consonants the language would still be just some mumble and the L M N relatively low with T S P F K in the highest range is needed and here we need low harmonic and IM distortion and stored energy. Good impulse/time behaviour since these sounds are transient in nature. Wall reflectins in the room destroy the resolution of these sounds due to multiple impulses where there should be only one.
Just venting..
/Peter
A E U I O (and all European AE Å Ä Ö U versions) are in low range and fundamental. Even frequency response is needed and freedom from gross resonances especially those from the room acoustics.
Without consonants the language would still be just some mumble and the L M N relatively low with T S P F K in the highest range is needed and here we need low harmonic and IM distortion and stored energy. Good impulse/time behaviour since these sounds are transient in nature. Wall reflectins in the room destroy the resolution of these sounds due to multiple impulses where there should be only one.
Just venting..
/Peter
Rick,
With all due respect you are starting to over-analyze the situation. I can only tell you from first hand experience and 30yrs of mucking around with all this stuff that the Fostex 206E is a very special driver (and cheap too so you've got nothing to lose). Its perfect for AV use, you get electrostatic-like sound quality without the size and cost penalties of the panels.
My reference point for pure detail, nuance and true tone in a speaker is electrostats and ribbons (yes I use the Raven R-1 too). In particular the ultra thin membraned (3.6 microns) electrostat I use, namely the ESL-3 made by www.eraudio.com.au , is extra special to me as it gives an electron microscope's view on a recording.
The Fostex though not without its faults is a no miss product for the AV purposes you propose, so long as bass augmentation (a subwoofer) is added.
Regards,
Steve M.
With all due respect you are starting to over-analyze the situation. I can only tell you from first hand experience and 30yrs of mucking around with all this stuff that the Fostex 206E is a very special driver (and cheap too so you've got nothing to lose). Its perfect for AV use, you get electrostatic-like sound quality without the size and cost penalties of the panels.
My reference point for pure detail, nuance and true tone in a speaker is electrostats and ribbons (yes I use the Raven R-1 too). In particular the ultra thin membraned (3.6 microns) electrostat I use, namely the ESL-3 made by www.eraudio.com.au , is extra special to me as it gives an electron microscope's view on a recording.
The Fostex though not without its faults is a no miss product for the AV purposes you propose, so long as bass augmentation (a subwoofer) is added.
Regards,
Steve M.
Pan said:
I agree with this completely. I've heard enough metal cone systems that had perfect response curves, perfectly notched breakup nodes, and perfect LR rolloff shapes, but still sounded harsh, edgy, bright or fatiguing because they simply crossed over too high.
I've got some related data on the L18 design at my web site, if anyone wants to see this visually. The 3rd harmonic is nasty, and happens way below the breakup node.
Just for grins, I did an elliptical crossover with the L18 and a Vifa XT19 at 2.7kHz and it sounded terrible. I can't be sure if it might have been that I just don't like the sound of 8th order slopes, but I suspect it was the 3rd HD coming through the woofer that gave it an edgy and bright sound.
I prefer metal cones over paper, provided they are crossed over right. The problem is so many designs are not crossed over well, thus the general consensus that paper sounds better.
Pan posted:
Would you rather have a 50W amp that had 0.05% THD for <5W, and THD rose to 10% at 50W, or an amp that had 0.1% THD at all power levels below 50W?
Sometimes the wise choice is to accept a tenth or two percent in THD (especially if it's low and even order) for much lower THD over a wide range of operation. Having a driver with THD below the limits of audibility but limited dynamics (stroke, either from design or use of the driver) imposes many other problems related to the package as a whole.
Don't choose your car simply on the basis of the torque of the engine - you should look at the entire package.
IMHO, if the errors are low enough to be at or below the limits of audibility (and THD down in the 0.5% to 1% range is, IMHO) then go for the widest frequency and dynamic range you can. The entire system will gain from that choice. Going back to the car analogy, giving up some HP and torque to gain better traction and extra gears in the transmission. Better able to harness what you have.
Dan Wiggins
Adire Audio
To eliminate nonlinearities from exciting the resonances of the driver, you'd want to HIGH PASS the driver - eliminate the need for stroke. Low pass filters don't help here.
Within it's linear range of motion, yes. But there are also thresholds below which it no longer matters...
Would you rather have a 50W amp that had 0.05% THD for <5W, and THD rose to 10% at 50W, or an amp that had 0.1% THD at all power levels below 50W?
Sometimes the wise choice is to accept a tenth or two percent in THD (especially if it's low and even order) for much lower THD over a wide range of operation. Having a driver with THD below the limits of audibility but limited dynamics (stroke, either from design or use of the driver) imposes many other problems related to the package as a whole.
Don't choose your car simply on the basis of the torque of the engine - you should look at the entire package.
Then you also necessarily force yourself into the use or multiple drivers, comb filtering, off-axis dispersion problems, complexity in the crossover, higher susceptibility to crossover and driver parameter errors, etc. Especially when considering center channels, as has been discussed in this thread - wide horizontal dispersion is very important, and going to multiple drivers will not help there at all.
IMHO, if the errors are low enough to be at or below the limits of audibility (and THD down in the 0.5% to 1% range is, IMHO) then go for the widest frequency and dynamic range you can. The entire system will gain from that choice. Going back to the car analogy, giving up some HP and torque to gain better traction and extra gears in the transmission. Better able to harness what you have.
Dan Wiggins
Adire Audio
Re: Some more on the voice 'target'
Mild nit to pick, AM's transmission bandwidth was reduced to 10 kHz in North America with the last modification the NRSC curve under twenty years ago. Prior to that the station I worked at cut off at 15 kHz. The 4 kHz seen in better AM radios (sigh) is due to the economics of manufacturing a wide band, low noise AM receiver vs. the market potential. It's not a broadcast standard by any means and we monitor off-air to the full 10 kHz.
rick57 said:
Mild nit to pick, AM's transmission bandwidth was reduced to 10 kHz in North America with the last modification the NRSC curve under twenty years ago. Prior to that the station I worked at cut off at 15 kHz. The 4 kHz seen in better AM radios (sigh) is due to the economics of manufacturing a wide band, low noise AM receiver vs. the market potential. It's not a broadcast standard by any means and we monitor off-air to the full 10 kHz.
What we need is a driver with good bandwidth over the critical (300HZ-3.5K), moderate to good efficiency, with controlled break-up at the top-end and the ability to play at least an octave or two below 300HZ & above 3.5K. No nasty break-up nodes that need notch-filters & the ability to play with low distortion (< 1% THD) at reasonable levels & only moderate levels(<5%) as we approach let's say 105db. It should get to 105db/1W/1M without significant power compression (let's say less than 10%) and should be small enough(how about a 5.25") to have good off-axis performance up to the desired 3.5K-4K crossover.
Could you design a driver like that for us Dan?
Could you design a driver like that for us Dan?
Dan,
I don´t know if I´m bad at express myself and write but I believe I have said most of what you say allready and are perfectly aware of what you say.
"To eliminate nonlinearities from exciting the resonances of the driver, you'd want to HIGH PASS the driver - eliminate the need for stroke. Low pass filters don't help here."
Wrong, as I allready state you need both. At normal levels the harmonics most likely to excite a break up is the 2nd and 3rd at the fr. of concern. That would be (as I already explain and I assume you are aware of) 2k and 3k for a driver with a 6k resonance. Now if you push things of course the higher order products of much lower frequencies will be of magnitude high enough to excite the breakup therefore, as you suggest, a highpass.
"A driver with inherent low distortion will always excite the breakup less than a driver with higher dist.
Within it's linear range of motion, yes. But there are also thresholds below which it no longer matters..."
And of course we take this into consideration when designing a high performance system.
"Would you rather have a 50W amp that had 0.05% THD for <5W, and THD rose to 10% at 50W, or an amp that had 0.1% THD at all power levels below 50W?"
Absolutely. Used with drivers with sensitivity high enough. 0.1% is absolutely not acceptable for an amp with todays low distortion drivers. There are drivers with about and below 0.1%HD and I would never use an amp that does not have distortion at least 1/10 of the driver. Using a 90dB driver I get by with a couple of watts for most listening and it is ther I want my rig to shine. I would never sacrifice the performance on acoustic instruments and voices for partylevels with pop music.
"Sometimes the wise choice is to accept a tenth or two percent in THD (especially if it's low and even order)for much lower THD over a wide range of operation."
I don´t buy into that hype anymore. If a system/component has relatively high even low order HD and low higher odd order this will change as soon as a complex signal is put into the system. IM dist is something that should be avoided in any situation IMO.
But the last part, wasn´t that exatcly what I said? Or am I missing someting..? I´ll be glad to take overall low distortion with a small peak here and ther insted of a driver with a flat dist. curve but overall higher.
"Having a driver with THD below the limits of audibility but limited dynamics (stroke, either from design or use of the driver) imposes many other problems related to the package as a whole."
Yes, but one does not have to exclude the other. Also few if any drivers have been made yet with distortion below the thresholds of audibility.
"Don't choose your car simply on the basis of the torque of the engine - you should look at the entire package."
Absolutely, that´s the only way to end up with high perfrmance.
"Then you also necessarily force yourself into the use or multiple drivers, comb filtering, off-axis dispersion problems, complexity in the crossover, higher susceptibility to crossover and driver parameter errors, etc."
A multiway speaker is a must for low distortion, good dispersion and wide dynamic range. I´m sure you would opt for a two way minimum and typically the x-ing will be somewhere (but not necessarily) in the upper mids/lower highs so you will get the potential problems there anyway. A three way will can have the lower crossing where wavelengths are long enough not to present any problems regarding loobing and combing.
I don´t really know what you mean with the last part though.
"Especially when considering center channels, as has been discussed in this thread - wide horizontal dispersion is very important, and going to multiple drivers will not help there at all."
What? horizontal dispersin is typically better with several drivers since you can operate them where they are not beaming or changing dispersin greatly.. So it´s the other way around really. Not that I use centers myself though.
"IMHO, if the errors are low enough to be at or below the limits of audibility (and THD down in the 0.5% to 1% range is, IMHO)"
I would never let a driver with that high distortion into my home I can say. I used such drivers many years ago but there´s no reason to when there are so much better performing examples now. Some of the better performers aren´t even very expensive... but many are unfortunately.
"then go for the widest frequency and dynamic range you can. The entire system will gain from that choice."
Here I agree but the nice thing is that this can easily go hand in hand with low overall distortion. You don´t need to sacrifise anything and it´s even more likely that a low distortion driver will excel even in dynamics. It´s really very simple, distortion in a driver is dynamic compression more or less.
Now, talking about low distortion and dynamics.. I´m thinking of using four tumults in two stereo push-push closed box subs. Ever tried that? Flat down to 5-10Hz and low distortion would be interesting to experience.
Happy listening!
/Peter
I don´t know if I´m bad at express myself and write but I believe I have said most of what you say allready and are perfectly aware of what you say.
"To eliminate nonlinearities from exciting the resonances of the driver, you'd want to HIGH PASS the driver - eliminate the need for stroke. Low pass filters don't help here."
Wrong, as I allready state you need both. At normal levels the harmonics most likely to excite a break up is the 2nd and 3rd at the fr. of concern. That would be (as I already explain and I assume you are aware of) 2k and 3k for a driver with a 6k resonance. Now if you push things of course the higher order products of much lower frequencies will be of magnitude high enough to excite the breakup therefore, as you suggest, a highpass.
"A driver with inherent low distortion will always excite the breakup less than a driver with higher dist.
Within it's linear range of motion, yes. But there are also thresholds below which it no longer matters..."
And of course we take this into consideration when designing a high performance system.
"Would you rather have a 50W amp that had 0.05% THD for <5W, and THD rose to 10% at 50W, or an amp that had 0.1% THD at all power levels below 50W?"
Absolutely. Used with drivers with sensitivity high enough. 0.1% is absolutely not acceptable for an amp with todays low distortion drivers. There are drivers with about and below 0.1%HD and I would never use an amp that does not have distortion at least 1/10 of the driver. Using a 90dB driver I get by with a couple of watts for most listening and it is ther I want my rig to shine. I would never sacrifice the performance on acoustic instruments and voices for partylevels with pop music.
"Sometimes the wise choice is to accept a tenth or two percent in THD (especially if it's low and even order)for much lower THD over a wide range of operation."
I don´t buy into that hype anymore. If a system/component has relatively high even low order HD and low higher odd order this will change as soon as a complex signal is put into the system. IM dist is something that should be avoided in any situation IMO.
But the last part, wasn´t that exatcly what I said? Or am I missing someting..? I´ll be glad to take overall low distortion with a small peak here and ther insted of a driver with a flat dist. curve but overall higher.
"Having a driver with THD below the limits of audibility but limited dynamics (stroke, either from design or use of the driver) imposes many other problems related to the package as a whole."
Yes, but one does not have to exclude the other. Also few if any drivers have been made yet with distortion below the thresholds of audibility.
"Don't choose your car simply on the basis of the torque of the engine - you should look at the entire package."
Absolutely, that´s the only way to end up with high perfrmance.
"Then you also necessarily force yourself into the use or multiple drivers, comb filtering, off-axis dispersion problems, complexity in the crossover, higher susceptibility to crossover and driver parameter errors, etc."
A multiway speaker is a must for low distortion, good dispersion and wide dynamic range. I´m sure you would opt for a two way minimum and typically the x-ing will be somewhere (but not necessarily) in the upper mids/lower highs so you will get the potential problems there anyway. A three way will can have the lower crossing where wavelengths are long enough not to present any problems regarding loobing and combing.
I don´t really know what you mean with the last part though.
"Especially when considering center channels, as has been discussed in this thread - wide horizontal dispersion is very important, and going to multiple drivers will not help there at all."
What? horizontal dispersin is typically better with several drivers since you can operate them where they are not beaming or changing dispersin greatly.. So it´s the other way around really. Not that I use centers myself though.
"IMHO, if the errors are low enough to be at or below the limits of audibility (and THD down in the 0.5% to 1% range is, IMHO)"
I would never let a driver with that high distortion into my home I can say. I used such drivers many years ago but there´s no reason to when there are so much better performing examples now. Some of the better performers aren´t even very expensive... but many are unfortunately.
"then go for the widest frequency and dynamic range you can. The entire system will gain from that choice."
Here I agree but the nice thing is that this can easily go hand in hand with low overall distortion.
You don´t need to sacrifise anything and it´s even more likely that a low distortion driver will excel even in dynamics. It´s really very simple, distortion in a driver is dynamic compression more or less.Now, talking about low distortion and dynamics.. I´m thinking of using four tumults in two stereo push-push closed box subs. Ever tried that? Flat down to 5-10Hz and low distortion would be interesting to experience.
Happy listening!
/Peter
Pan said:
None of those absolutely lovely singled ended no-feedback amps for you then... a single number distortion number is next to meaningless... i'd much rather have an amp with 1% and predominantly 2nd order vrs one with 0.01% and high order because the 1st amp will sound better.
Distortion measurements as usually bandied about are red herrings.
dave
planet10 said:
True, true. I enjoyed Lynn Olson's article on the subject.
http://www.nutshellhifi.com/library/FindingCG.html
Joules said:
I'm using two per side 15" TC Sounds Aluminum subs in a 2" thick mdf sealed box per side , Lightly EQ'ed, 3dB down at 16 hz
More woofers were less expensive than more "quality amplifier"
Sounds nice. I hope TC go back to offer UH drivers again since that would be my first choice. Alu/Ti 15" UH four pieces as you use in closed boxes.
Quality don´t come cheap but it sure is nice.
/Peter
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.