Ideal speaker for reproducing voice

[ajd578]

I've read that no speaker can be perfect - every speaker is an optimization to a specific goal, with necessary compromises.

So if the goal was to have a speaker through which playback of a vocalist *alone* was indistinguishable from a live, unamplified singer, to a blinded listener, what measurement characteristics would that speaker have, and how would it be best approximated with real components?

 

[fpitas]

It would measure as Amir's ideal speaker, in every respect. That's the idea behind his evaluation.

 

[ajd578]

It would measure as Amir's ideal speaker, in every respect. That's the idea behind his evaluation.

I'm skeptical of that. What about horizontal and vertical dispersion of sound coming from a person, vs. typical speaker enclosures? Or number of drivers (size of sound source), or crossovers/phase?

 

[fpitas]

I'm skeptical of that. What about horizontal and vertical dispersion of sound coming from a person, vs. typical speaker enclosures? Or number of drivers (size of sound source), or crossovers/phase?

You'll have to sort all that out for yourself. The dispersion is part of the existing testing.

 

[ajd578]

You'll have to sort all that out for yourself. The dispersion is part of the existing testing.

Yeah dispersion is measured, but not with this goal in mind, as far as I know.

 

[fpitas]

Yeah dispersion is measured, but not with this goal in mind, as far as I know.

OK. Sounds like you already have a lot of pre-conceived notions.

 

[ajd578]

OK. Sounds like you already have a lot of pre-conceived notions.

No, just a higher evidence threshold for accepting an explanation than what you've offered.

 

[fpitas]

No, just a higher evidence threshold for accepting an explanation than what you've offered.

Great. Bye.

 

[theREALdotnet]

It would measure as Amir's ideal speaker, in every respect. That's the idea behind his evaluation.

That seems like a fairly limited specification. In essence, Amir’s ideal speaker has a flat frequency response and smooth directivity all around. Nothing else matters. Compression, phase, time alignment, stereo performance, pair matching and other parameters are not evaluated. It’s because these things are not mentioned (or are dismissed) in the scripture of St Toole of Harman.

Ok, enough with the facetiousness, I do think that current and future work into characterising speakers will yield useful information on what makes a good speaker, for different applications and listener preferences.

But back to the OP’s question (and to facetiousness), here is an easy test of whether a pair of speakers will excel at reproducing the human voice: turn them around, and if it says “LS3/5a” on the back of both, you have speakers that will do this with startling, almost spooky realism.

 

[JSmith]

indistinguishable from a live, unamplified singer, to a blinded listener

Yeah that 'aint going to happen anyway.


JSmith

 

[theREALdotnet]

Yeah that 'aint going to happen anyway.

I agree, not in an A/B comparison anyway. But able to fool a listener into thinking there is a live, unamplified singer in the room, without direct comparison? Absolutely.

 

[Thomas_A]

For directivity of the human voice.

Facts about speech intelligibility: human voice frequency range

Recording or amplifying the voice is a technical challenge. In this article we present facts on speech intelligibility, including the voice as a sound source.

www.dpamicrophones.com

 

[Plcamp]

For voice, I would expect a physically large single horn covering the range of about 300 to 5 kHz would produce best possible imitation of reality

 

[CREMA]

Speakers with narrow directivity and linear response are needed. The initial reflection and reverberation in the listening room must also be very low or controlled.

When a harmful level of comb-filtering occurs, the feeling of the voice collapses.

 

[DSJR]

That seems like a fairly limited specification. In essence, Amir’s ideal speaker has a flat frequency response and smooth directivity all around. Nothing else matters. Compression, phase, time alignment, stereo performance, pair matching and other parameters are not evaluated. It’s because these things are not mentioned (or are dismissed) in the scripture of St Toole of Harman.

Ok, enough with the facetiousness, I do think that current and future work into characterising speakers will yield useful information on what makes a good speaker, for different applications and listener preferences.

But back to the OP’s question (and to facetiousness), here is an easy test of whether a pair of speakers will excel at reproducing the human voice: turn them around, and if it says “LS3/5a” on the back of both, you have speakers that will do this with startling, almost spooky realism.

LS3/5A's are actually not good these days - uneven response deliberately tuned for the OB task they were primarily used for and making the best of what are now outdated original drivers They were always a loss leader for their makers but of course now, the wealthier retro market seems to keep the blasted things going. Silly money for the standard they set today.

I still maintain that good speaker design is an art as well as a science. Get the measurements right and then listen, listen and listen some more! Fine tune half a dB here and there for best subjective results, but only once the fundamentals are done.

 

[Thomas_A]

Speakers with narrow directivity and linear response are needed. The initial reflection and reverberation in the listening room must also be very low or controlled.

When a harmful level of comb-filtering occurs, the feeling of the voice collapses.

Why is that? A natural voice have wide directivity.

 

[ajd578]

Speakers with narrow directivity and linear response are needed. The initial reflection and reverberation in the listening room must also be very low or controlled.

When a harmful level of comb-filtering occurs, the feeling of the voice collapses.

I don't see why reflections should be low - are you saying you need a specific room to achieve this? A person sounds like a person regardless of the room they are in.

 

[kongwee]

I can't remember how many attributes need to be ideal. However, you just need flat FR, low SINAD and high Klippel Near Field Scanner ASR rating.

 

[Koeitje]

I'm skeptical of that. What about horizontal and vertical dispersion of sound coming from a person, vs. typical speaker enclosures? Or number of drivers (size of sound source), or crossovers/phase?

Your voice is a point source, so you want smooth directivity. Like with any loudspeaker.

 

[theREALdotnet]

LS3/5A's are actually not good these days - uneven response deliberately tuned for the OB task they were primarily used for and making the best of what are now outdated original drivers They were always a loss leader for their makers but of course now, the wealthier retro market seems to keep the blasted things going. Silly money for the standard they set today.

Yep, I’m quite familiar with all these misconceptions

I would dearly love to be able to replace them, especially considering the prices of the modern versions. So far I have come up empty, and not for lack of trying.

With regard to retro charm, yes, that’s a problem because it drives up prices. However, they were special even “back then” and exceeded their design mandate in unexpected ways. This is what Peter Aczel wrote in 1977:

Meanwhile, we suggest you carry a pair of Rogers LS3/5A’s around with you (in your car trunk, in an overnight bag, or whatever) when you go hi-fi visiting. After listening to your friend’s (or dealer’s) $2500 speaker system, you connect your little shoeboxes and you’ve got an enemy for life.

But let’s not derail this thread, perhaps one day I’ll start a separate one.