Iplaynaked asked:
"Er, what specifically do you feel are "hard and fast rules" that are counter-intuitive, as you say? I believe everyone here would like to know, actually. I would."
Well, that's not really what I said. I said that much of audio is counter-intuitive.
Okay, I don't know if these facts are so widely accepted as to be considered "hard and fast rules", but I think they are counter-intuitive:
1. Total harmonic distortion figures fail to reliably predict subjective preference.
2. The perceived tonal balance of a loudspeaker can change with loudness level even if the measured frequency response curve stays the same (Fletcher-Munsen curve; this one might qualify as a "hard-and-fast rule").
3. The on-axis anechoic frequency response curve is not a reliable predictor of perceived sound quality or tonal balance.
4. The ear largely ignores reflections arriving later than .68 milliseconds after the first-arrival sound as far as directional cues goes, but takes into account these same reflections as far as tonal balance and loudness goes.
5. Semi-random distribution of multiple low frequency sources in a room results in smoother bass, both measured and perceived, than careful placement of a single low frequency source.
6. The ear perceives sound quality very differently from the way the eye judges a frequency response curve; narrow-band peaks and dips that leap out to the eye are often inaudible to the ear.
7. The ear has a characteristic called "masking" that tends to ignore a low-level signal in the presence of a louder signal if they are close in frequency. This principle is applied in audio data-compression algorithms.
8. This principle of masking works in the frequency domain but not in the time domain, so that if the low-level signal is a distortion that arrives later in time, perhaps because of a path lenth difference (as with diffraction), then it is much more likely to be audible.
9. Speaking of diffraction, this is a type of distortion that our ears have a level-dependent sensitivity to; that is, we don't hear it at low volume levels but we do at high volume levels (as a harshness) - and this is one of those distortions that is not revealed by a frequency response curve.
10. The ear is relatively insentitive to the preservation, or lack thereof, of the phase relationships in a music signal. I'm not saying it's undetectable, but certainly not as readily detectible by the ear as by a microphone.
I could go on, but ten seems like a reasonable number to stop at.
By the way, I am not advocating 20 degrees of rotation of the speaker/listener triangle as a "hard and fast rule". It is an application of a principle to a specific situation. I trust that a re-reading of my posts in this thread will reveal what that priciple is. Being able to apply principles is more useful than memorizing rules, because the "rule" may not be practical to apply in the next person's room, but perhaps the principle can be applied in another way.
Regarding what size drivers and what type of box a speaker should have in a small room, I think you are arguing based on generalities that apply to most speakers, and I'm saying that if done right the specifics of a more nearly ideal solution in this case are different from what an overview of generalities would predict.
Duke
dealer/manufacturer
"Er, what specifically do you feel are "hard and fast rules" that are counter-intuitive, as you say? I believe everyone here would like to know, actually. I would."
Well, that's not really what I said. I said that much of audio is counter-intuitive.
Okay, I don't know if these facts are so widely accepted as to be considered "hard and fast rules", but I think they are counter-intuitive:
1. Total harmonic distortion figures fail to reliably predict subjective preference.
2. The perceived tonal balance of a loudspeaker can change with loudness level even if the measured frequency response curve stays the same (Fletcher-Munsen curve; this one might qualify as a "hard-and-fast rule").
3. The on-axis anechoic frequency response curve is not a reliable predictor of perceived sound quality or tonal balance.
4. The ear largely ignores reflections arriving later than .68 milliseconds after the first-arrival sound as far as directional cues goes, but takes into account these same reflections as far as tonal balance and loudness goes.
5. Semi-random distribution of multiple low frequency sources in a room results in smoother bass, both measured and perceived, than careful placement of a single low frequency source.
6. The ear perceives sound quality very differently from the way the eye judges a frequency response curve; narrow-band peaks and dips that leap out to the eye are often inaudible to the ear.
7. The ear has a characteristic called "masking" that tends to ignore a low-level signal in the presence of a louder signal if they are close in frequency. This principle is applied in audio data-compression algorithms.
8. This principle of masking works in the frequency domain but not in the time domain, so that if the low-level signal is a distortion that arrives later in time, perhaps because of a path lenth difference (as with diffraction), then it is much more likely to be audible.
9. Speaking of diffraction, this is a type of distortion that our ears have a level-dependent sensitivity to; that is, we don't hear it at low volume levels but we do at high volume levels (as a harshness) - and this is one of those distortions that is not revealed by a frequency response curve.
10. The ear is relatively insentitive to the preservation, or lack thereof, of the phase relationships in a music signal. I'm not saying it's undetectable, but certainly not as readily detectible by the ear as by a microphone.
I could go on, but ten seems like a reasonable number to stop at.
By the way, I am not advocating 20 degrees of rotation of the speaker/listener triangle as a "hard and fast rule". It is an application of a principle to a specific situation. I trust that a re-reading of my posts in this thread will reveal what that priciple is. Being able to apply principles is more useful than memorizing rules, because the "rule" may not be practical to apply in the next person's room, but perhaps the principle can be applied in another way.
Regarding what size drivers and what type of box a speaker should have in a small room, I think you are arguing based on generalities that apply to most speakers, and I'm saying that if done right the specifics of a more nearly ideal solution in this case are different from what an overview of generalities would predict.
Duke
dealer/manufacturer