[sander99]

Surely, one driver can be used.
Yet, there are limitations of resonances, cross-couplings, etc.

One membrane can hardly practically vibrate perfectly as needed to reproduce music in 3-decade range of 20 Hz to 20 kHz.

Then a lot of efforts is involved in DDs to make them work better, as mentioned - very strong field, multiple coils, compensation...
To me an array of small drivers each working in a well-defined limited range of 2-3 octaves is a more flexible solution.

Yes I understand. I just wanted to give some more context to my other post in which I state that an additional set of drivers by itself will not be perceived as a different channel, I mean will not be perceived as sound coming from a different direction. So using different drivers to work on different frequency bands can be usefull, but it does not mean it would be usefull to also make another "division of work" by adding extra sets of drivers for other channels (for 5.1 etc.).

What I really mean I gues is: If you think 4 or 5 drivers per ear make a good optimal choice for covering the audio spectrum with good resolution and sufficient sound pressure levels to listen to stereo recordings, then I say that same design is equally good for multichannel (using binaural signals, and multichannel over iems or headphones without binaural signals is meaningless).

 

[PhonoPhi]

Yes I understand. I just wanted to give some more context to my other post in which I state that an additional set of drivers by itself will not be perceived as a different channel, I mean will not be perceived as sound coming from a different direction. So using different drivers to work on different frequency bands can be usefull, but it does not mean it would be usefull to also make another "division of work" by adding extra sets of drivers for other channels (for 5.1 etc.).

What I really mean I gues is: If you think 4 or 5 drivers per ear make a good optimal choice for covering the audio spectrum with good resolution and sufficient sound pressure levels to listen to stereo recordings, then I say that same design is equally good for multichannel (using binaural signals, and multichannel over iems or headphones without binaural signals is meaningless).

More drivers (both in capabilities and quantaties) offer better flexibility and more opportunities.

With BAs, my limited experience that 4-5 frequency bands work really well for the resolution (the price to pay is definitely their coherence).
Another (more subtle and of minor relevance to this discussion) point is that multiple BAs working in the same frequency range can offer better smoothness with the partial canceling of their individual resonances (similar to a single violin vs. symphonic violin section).

Then spatially, it would be just simpler to generate the desired wavefront with all the required details at the ear channel entrance using several spatially separated groups of drivers.

 

[sander99]

Then spatially, it would be just simpler to generate the desired wavefront with all the required details at the ear channel entrance using several spatially separated groups of drivers.

I am still doubting if you understand my main point: Using more drivers is fine but the spatial placement of the drivers will not influence the perceived direction of the sound.

 

[bigshot]

I don’t think he has experienced special sound aside from the secondary depth cues in recordings. I have a lot of trouble explaining what soundstage is to people who have only used headphones. It just isn’t part of their frame of reference.

 

[castleofargh]

Then spatially, it would be just simpler to generate the desired wavefront with all the required details at the ear channel entrance using several spatially separated groups of drivers.

I'm also not sure about that. You're already at the entrance of the ear canal(or in it) so you do not need to create a "sound field". What you do need it a signal processed for that playback method(including your own HRTF applied to it).
Using multidriver+crossovers aren't only bringing benefits, so the pros and cons have to be weighted carefully. I've heard several IEM manufacturers talking about 3ways being a good compromise for BAs. I do not know if it's accurate, but they know more than me for sure. That's specifically with BA drivers in mind. When free to chose, I'm still not sure that multidriver is the optimal way. The easiest, sure, but best for fidelity?

 

[bigshot]

When you get into multiple drivers, you run into issues with overlapping frequencies. Not sure how that plays out in IEMs. In an ideal world, a single full range driver should be easier to balance, and it seems there are high quality single driver models on the market, so it shouldn’t be impossible.

 

[PhonoPhi]

I am still doubting if you understand my main point: Using more drivers is fine but the spatial placement of the drivers will not influence the perceived direction of the sound.

I think you (and me) are both concentrated on our main points.
Let us (me) stop here.

 

[PhonoPhi]

When you get into multiple drivers, you run into issues with overlapping frequencies. Not sure how that plays out in IEMs. In an ideal world, a single full range driver should be easier to balance, and it seems there are high quality single driver models on the market, so it shouldn’t be impossible.

Single driver should have this "ideal" transducer independently simultaneously vibrating from 20 Hz to 20 kHz.
I do agree that historically single drivers were most developed solution. Can it be practically sufficient, hard to say.

Overlapping frequencies are addressed by different crossovers designs, starting from simple frequency filters. Still, definitely the most difficult aspect to work out with multiple transducers.

 

[PhonoPhi]

I'm also not sure about that. You're already at the entrance of the ear canal(or in it) so you do not need to create a "sound field". What you do need it a signal processed for that playback method(including your own HRTF applied to it).
Using multidriver+crossovers aren't only bringing benefits, so the pros and cons have to be weighted carefully. I've heard several IEM manufacturers talking about 3ways being a good compromise for BAs. I do not know if it's accurate, but they know more than me for sure. That's specifically with BA drivers in mind. When free to chose, I'm still not sure that multidriver is the optimal way. The easiest, sure, but best for fidelity?

A good question. Thinking of the simplest implementation of phase delays, I would think it is just easier with physically distinct drivers (?)

 

[bigshot]

How would you have timing error inside an ear?

 

[PhonoPhi]

How would you have timing error inside an ear?

(I am not sure if it is addressed to me, also I realized that I exceeded "my reasonable usefulness" in this thread. I am not an engineer. I do work with science, but not directly relevant...)

Having several channels similar to 5.1, 7.1, etc, any timing errors can be calibrated with the software similar to speaker calibration. It would be great that in addition to the "head shape", individual anatomy of the ear channels, as well as individual physiology and preferences of the sound perception can be selectively adjusted.

I did not have any experience with "Smyth Realiser (?). I am sure that the software can be reasonably implemented at the DAC-transducer level, if there will be a demand/market for spacially-resolved sound reproduction.

Right now, I simply enjoy my all-BA IEMs, in particular exploring music (and different sound implementation) of Decca Analog collection, along your recommendations in classical music

 

[bigshot]

I don't think you understand what multichannel is. It isn't just multiple speakers... it's is 4 or more separate and discrete tracks channeled to different parts of the room, both front and rear. It's surround sound that interacts with the listening room to create a 360 degree sound field. You can't do that with IEMs, even ones with multi driver designs. We aren't talking about stereo. It's multitrack audio designed to create a sound environment.

 

[PhonoPhi]

5.1 headphones exist for a long time. Making multichannel IEMs may be impractical, but the boundary is mute between small headphones and larger planar IEMs, and some drivers are very small now.

So I am not sure what I need to understand. Please tell me. I'll patiently learn.
I actually have one of those budget 5.1 headphones - poor implementation, now disassembled at my desk...but I was curious at one point.

 

[bfreedma]

5.1 headphones exist for a long time. Making multichannel IEMs may be impractical, but the boundary is mute between small headphones and larger planar IEMs, and some drivers are very small now.

So I am not sure what I need to understand. Please tell me. I'll patiently learn.
I actually have one of those budget 5.1 headphones - poor implementation, now disassembled at my desk...but I was curious at one point.

Using fairly basic DSP to simulate multichannel via headphones doesn’t present surround nearly as comprehensively as discreet channel multichannel audio via speakers. The Smyth Realizer is the only decent simulation via headphones I’ve experienced. That said, it still, for me, was not close to a well setup MCH speaker setup.

 

[bigshot]

If you’ve never experienced multichannel sound, the only way I can explain it to you is to say that it is as much an improvement over stereo as stereo is over mono.

I think the reason there are long discussions on spatiality of sound is because many headphone users just can’t conceive of what spatial sound sounds like, since they’ve never experienced it themselves. If you’ve heard it done properly, it’s a self evident improvement.