Preference vs. audibility - please keep them separate.
- Thread starter j_j
- Start date
- Status
Yes Orb; I was curious to learn from Ethan's perspective.
...And actually more on the amp's slew rate than the damping factor; and how both can affect our audibility.
I realize that with everything in life a good match between amp, preamp and speakers is key to 'preference-versus-audibility's mind processing.
...And add wiring and source, of course.
Damping factor of the rest of your system must match what the speakers expect. Otherwise you see things like I saw in one manual for a speaker, which is (not quoted word by word, but quoted accurately as to meaning) "you must have less than .06 ohm resistance in your amplifier output impedance and your speaker wiring.
I've built speakers, from inexpensive drivers, that are not quite state of the art, but that do have reasonable power response as well as direct response, that have an impedence that never goes above 15 ohms or below 7.8 ohms. Anything will drive it, pretty much, that will drive 8 ohms. It's not really hard, but it's also not terribly easy. You must pay a lot of attention to a variety of annoying physics.
Such requirements as .06 ohms indicate to me that some redesign might be in order.
I've built speakers, from inexpensive drivers, that are not quite state of the art, but that do have reasonable power response as well as direct response, that have an impedence that never goes above 15 ohms or below 7.8 ohms. Anything will drive it, pretty much, that will drive 8 ohms. It's not really hard, but it's also not terribly easy. You must pay a lot of attention to a variety of annoying physics.
Such requirements as .06 ohms indicate to me that some redesign might be in order.
I can't imagine why an amplifier (or any audio device) whose response extends beyond the audible range should sound any different from an amplifier that's limited to the audible range. Your comparison above isn't meaningful because one limit is +/- 0.5 dB and the other is +/- 3 dB. As long as both are flat within half a dB or so in the audible range, that should be flat enough to not matter. What happens beyond that range shouldn't affect the sound. If you play a CD for the source, its own response is hard-limited at around 20 KHz, so an amplifier that can pass frequencies higher than that won't have anything to pass! And good luck finding a speaker than can reproduce 100 KHz!
I recently wrote this short article about slew rate to complement my Audio Expert book:
Slew Rate And Slew Rate Limiting
As for damping factor, it's important, but it's not an issue with modern solid state power amps. In practice, damping factor is limited to around 50 simply because the speaker driver's own voice coil is in series with everything else. I'm addressing both of these issues in more detail in my workshop Lies, Damn Lies, and Audio Gear Specs at the AES show in a few weeks.
--Ethan
There is a reason: while a mono source that extends beyond 20 kHz (or let's face it for most WBF members beyond 14 kHz) cannot be heard or perceived, in stereo two sources can mix acoustically producing sum and difference frequency components where the difference components can be heard. How important this is for typical recordings I don't know, but it is a conceivable reason why extended frequency ranges are not completely ridiculous.
I can't imagine why an amplifier (or any audio device) whose response extends beyond the audible range should sound any different from an amplifier that's limited to the audible range. Your comparison above isn't meaningful because one limit is +/- 0.5 dB and the other is +/- 3 dB. As long as both are flat within half a dB or so in the audible range, that should be flat enough to not matter. What happens beyond that range shouldn't affect the sound. If you play a CD for the source, its own response is hard-limited at around 20 KHz, so an amplifier that can pass frequencies higher than that won't have anything to pass! And good luck finding a speaker than can reproduce 100 KHz!
I recently wrote this short article about slew rate to complement my Audio Expert book:
Slew Rate And Slew Rate Limiting
As for damping factor, it's important, but it's not an issue with modern solid state power amps. In practice, damping factor is limited to around 50 simply because the speaker driver's own voice coil is in series with everything else. I'm addressing both of these issues in more detail in my workshop Lies, Damn Lies, and Audio Gear Specs at the AES show in a few weeks.
--Ethan
Are you sure, Ethan? Haven't we been told, on this forum, that if a high res file or SACD, with supersonic content is played into an amp that can reproduce that content that, while we are not capable of hearing the supersonic content itself, it can create IM within the audible range? So an amp that's range extends beyond audibility could sound different, when given supersonic content.
Tim
Dont forget the speakers , most have significant roll after +- 20 khz anyway
What comes in doesnt come out
What comes in doesnt come out
Last edited:
Dont forget the speakers most have significant roll after +- 20 khz anyway
What comes in doesnt come out
For normal speakers, yes, but there are these supertweeters that can go out beyond 20 kHz. As I say, I don't know how significant it is with real music, but you could certainly make a synthetic track where the only audible sound was from acoustic mixing in the air. I don't know whether it's something that could happen with a 'genuine' stereo recording, or only artificially-created tracks.
I ve compared with diamond cones extension over 60 khz ,i did notice some bats flying outside.
In my opinion 25 khz is more than enough
In my opinion 25 khz is more than enough
I will probably try a berylium set later on , which sits in the middle +- 40 khz and is much cheaper than diamond
But for the moment i ve had it with ultra hard membranes domes/cones
But for the moment i ve had it with ultra hard membranes domes/cones
No designer of an extended bandwidth amplifier will tell you that his amplifier sounds better because it is able to reproduce 100 kHz. Usually the extended bandwidth is a consequence of another objective of the design - the designers of Spectral, Burmester and Soulution, just to nominate a few known examples, have explained in the past the whys of their options.
Sum and difference IM frequencies are generated only in the presence of nonlinearity. Simply combining two frequencies, whether electrically or acoustically, does not generate IM products. Now, you could argue that there's always some distortion, so if the amp or speaker driver distorts, that will create IM products. And that's true! And it happens at all audio frequencies, not just ultrasonic. But ... and these are big ones:
1) The IM distortion of a competent amplifier should be low enough that the artifacts are too soft to hear anyway.
2) Even if the tweeter adds an audible amount of distortion, which is likely, you don't want to hear those IM products! So I consider filtering out ultrasonic junk a feature, not a failing.
3) Even if the amp and speaker are perfectly linear, our ears distort at high volume levels to create IM products. And as with #2, I much prefer that stuff be filtered out which avoids the problem. Many audio circuits routinely include such filtering for just this reason, to avoid IMD caused by content at frequencies we can't hear anyway. This won't happen when playing CDs, but it can happen in a live sound setting from, for example, microphones used for cymbals or other percussion.
BTW, #2 above was the downfall of the infamous Oohashi experiment we often see posted as "proof" that people hear or perceive ultrasonic frequencies. That experiment played a number of ultrasonic tones through a single tweeter, and the tweeter distorted creating difference tones in the audible range. When the experiment was repeated by someone else using separate tweeters for each ultrasonic tone, nobody could tell when the tones were turned on and off.
--Ethan
It was in that line of thinking that I made my earlier comment. And I was waiting for a scientific explanation by an expert amp designer.
I'm always under the strong belief (past, present & future) that a more linear and extended bandwidth in an amplifier is a desirable attribute.
...No feedback, or very little, good slew rate (thanks Ethan by the way for the earlier link), and a sufficient damping factor across the entire frequency range and at various and lowest impedance to match several loudspeakers out there (high-end ones including stats).
Can you hear the difference(s) between two amps of various bandwidth? ...I believe so, and the one with the largest should reproduce accuracy the closest, even if we can't hear over the human frequency range, and that the source and the speakers are limited anyway, to say from 5Hz to 50kHz.
--> It has to do with rejection of harmonic distortion very high outside the human audio spectrum as it can spill right back into it if not extended high and low enough. ...Say from DSD for example.
- Ralph (Atmasphere) talked about it a bit earlier here in this thread (posts #119, 126, and 131), and 2Hz to 200kHz (+/- 1dB) should be a good start.
...That is if you're serious enough about Sound & Science (Preference vs. Audibility).
Last edited:
Well there's a product called the Acoustic Spotlight that relies on nonlinearity of air:
And something else along the same lines here:
http://www.acoustics.org/press/133rd/2pea.html
If there's anything to it, the effect from our speakers at ultrasonic frequencies would be very small - but we are operating in a field where people worry about noise and distortions at -100dB!
Indeed.
Crossing two beams of ultrasonics at reasonably high levels (unfortunately the exact level for best operation depends on both temperature and pressure, which makes these devices somewhat unpredictable) creates a classic distortion mixer (like the mixers in radios that do hetrodyning to some extent, but a lot harder to control).
Air is not even remotely a linear or non-lossy medium.
I don't have the link handy, but google for "atmospheric loss as a function of frequency". Ahh, wait, I'll do it. http://sengpielaudio.com/calculator-air.htm there you go for the loss with distance.
While I don't have a site that expresses it handily, when you get to circa 100dB SPL in a free field you can start to see nonlinearity in air. At 120, it's not so hard to see. At 140dB SPL the idea of linearity is gone, you can create mass flow with sound waves. Some guys next door in acoustics research used to have a really cool demonstration of this at Bell Labs.
Now, below 90dB SPL or so, the distortion is under atmospheric noise level at the ear. (which is, by the way, some 6dB SPL to 8.5dB SPL itself, white noise, 20-20K). You can't hear that, quite, because in any critical band of the ear it's not above threshold. Barely, in fact, just barely, for the most sensitive frequency for a person with unimpaired hearing. That noise you can not get rid of without removing the air around the eardrum, which is obviously NOT going to happen to any living subject for what I hope are obvious reasons.
The people who talk about hearing 100dB below 0dB SPL need to explain how they avoid having the noise of the atmosphere (due to its molecular nature, this is not something that can be avoided) interfering with this...
In general, agreed. I am being somewhat cautious with preferences, because there are some preferences (not audio ones, mind you) that have the potential to cross the gulf between 'a's nose and 'b's by existing.
I've known people like that, even. (No insinuation intended there.)
- Status
Similar threads
