I tried an experiment. I created a 1000Hz tone in one channel and a sweeping tone in the other (400 - 3000Hz). The signal source and play back have very low measured distortion. When I play the sound back (at comfortable level) I can clearly hear effects of faint rising tones and even a tone decreasing in frequency during the test which are heterodyne effects.
With headphones on I do not hear this. This tells me that when the sound mixes in the air I can hear these beat frequencies because:
1) Non-linearity caused by speakers and/or...
2) Non-linearity of air itself and/or...
3) Non-linearity of my own ears.
What causes this is most intriguing to me and the fact that it is clearly audible. Any thoughts?
With headphones on I do not hear this. This tells me that when the sound mixes in the air I can hear these beat frequencies because:
1) Non-linearity caused by speakers and/or...
2) Non-linearity of air itself and/or...
3) Non-linearity of my own ears.
What causes this is most intriguing to me and the fact that it is clearly audible. Any thoughts?
Probably caused by signal frequency phasing due to reflections in the room from walls, ceiling etc.
Welcome to the real world of speaker and speaker amplifier combinations. Speakers are always the weak link in the chain.
-Edit: What Mr. Snell said. Move some furniture around, and the speakers too, and see if it changes
What kind of speakers are you using?
-Edit: What Mr. Snell said. Move some furniture around, and the speakers too, and see if it changes
What kind of speakers are you using?
An extra tone must be caused by nonlinearity, not reflections. Speaker nonlinearity is the most likely cause, as both tests involve air and your ears.
I brought the speakers near, lowered the volume and put my head by them for a near-field test. Same effect. Not sure that JonSnell would be right because some non linearity would be needed to generate the "outside" tones. In fact, playing the rising tone alone with sound bouncing off the wall yields nothing unusual besides the peaks and troughs of volume as the rising frequency cancels and adds with the reflection. EDIT: DF96 says pretty much the same thing (we replied at about the same time).
I tried various speakers from DIY, modded and Boston Acoustic A-100s. Same thing.
I agree, it is probably a speaker thing, but still wonder about the hearing and psychoacoustic parts of it.
I tried various speakers from DIY, modded and Boston Acoustic A-100s. Same thing.
I agree, it is probably a speaker thing, but still wonder about the hearing and psychoacoustic parts of it.
Last edited:
Might be happening here or not, don't know that cabinet speaker components and crossover, BUT speaker output is depends not only on applied voltage but also on magnetic flux at the voice coil.
If you modulate useful magnetic flux you AM modulate any tone being reproduced ... the essence of intermodulation.
You modulate magnetic flux when the lower frequency tone moves voice coil back and forth, because gap flux density is not uniform, and drops sharply (eventually down to 0) a short distance away (think a couple mm) unless you have a very long voice coil (2X or 3X gap length) , so in loudspeakers intermodulation is sort of built in, at most you minimize it.
1000Hz is "nothing" for a woofer, but midrange speakers are not only much smaller but also tend to be built with "short" coils: gap length or just above.
If you modulate useful magnetic flux you AM modulate any tone being reproduced ... the essence of intermodulation.
You modulate magnetic flux when the lower frequency tone moves voice coil back and forth, because gap flux density is not uniform, and drops sharply (eventually down to 0) a short distance away (think a couple mm) unless you have a very long voice coil (2X or 3X gap length) , so in loudspeakers intermodulation is sort of built in, at most you minimize it.
1000Hz is "nothing" for a woofer, but midrange speakers are not only much smaller but also tend to be built with "short" coils: gap length or just above.
Most likely the speakers themselves, many sources of intermodulation distortion abound both in the drivers and in the cross-over depending on the type of chokes used. (Ferrite cored inductors can be quite nasty even at ridiculously low levels due to core saturation.)
Such problems you're experiencing may well not be attributable to any single factor, but I'd be inclined to weigh more heavily on room acoustics. Without meticulous experimentation of placement, just plain blind luck, and some of either combination with room treatment, I think many more systems will have some degree of acoustic issues than is commonly admitted.
For a range of speakers with presumably some degree variation of driver design, crossover frequencies, etc to exhibit the same problem in the same room, which is not experienced with headphones certainly suggests the room is in play
For a range of speakers with presumably some degree variation of driver design, crossover frequencies, etc to exhibit the same problem in the same room, which is not experienced with headphones certainly suggests the room is in play
Interesting article. Yes I can hear that fluttering effect of the difference in frequencies when the two tones pass each other in frequency. It explains some of what I hear. As the fundamental tones spread apart to nearly a 2:1 ratio, I can still here what I believe are intermodulation effects.
You are hearing additional frequencies that were not in the original signals. It has already been said that these frequencies can only be generated by nonlinearities. Just like a mixer in a superhet creates the IF.
To generate these (f1+f2) or (f2-f1) frequencies two signals of different frequencies have to be applied to a nonlinear element.
In your original post you explained that you have one signal on the left and the other on the right channel of your stereo setup. Did I understand that correctly?
So, in the playback chain there is no adding or mixing of the two signals.
I hope that sound transmission in the air takes place linearly.
The first element where the two signals meet is the ear. So I reckon: an ear must be non linear.
This also explains why the effect does not happen with headphones. Each ear hears only one tone. Do they add up in the brain? If so, maybe that is a linear process.....
To generate these (f1+f2) or (f2-f1) frequencies two signals of different frequencies have to be applied to a nonlinear element.
In your original post you explained that you have one signal on the left and the other on the right channel of your stereo setup. Did I understand that correctly?
So, in the playback chain there is no adding or mixing of the two signals.
I hope that sound transmission in the air takes place linearly.
The first element where the two signals meet is the ear. So I reckon: an ear must be non linear.
This also explains why the effect does not happen with headphones. Each ear hears only one tone. Do they add up in the brain? If so, maybe that is a linear process.....
- Status
- Not open for further replies.