HT acoustics and the Benefits of better speakers in stereo and multichannel audio

[Kal Rubinson]

Not everyone has a thick wallet...

That is true. OTOH, what we are discussing is the validity of issues such as 2channel stereo, upmixing, multichannel, etc. as ways to reproduce an original event. Practical considerations such as cost, room space, WAF, etc are personal matters. For example, I would have systems suitable for Atmos/Auro3D were it not for issues such as room space and WAF, particularly the latter.

he spherical array of speakers makes sense if you use one speaker per channel not for some algorythm to generate a mish-mash illusion from a 2, 5 or 7 channel recording.

Aside from explicit statements about upmixing (which I abjure), we are presuming that each speaker is unique for each channel (and vice versa) and no algorithmic mish-mash is being conjured up.

 

[pozz]

...some algorythm to generate a mish-mash fake illusion of immersion from a 2, 5 or 7 channel recording.

Hmm. Just an example:

I set the time at a discussion of upmixing with a presentation of surround/ambience information extraction from stereo, done very well. How well you can hear through a number of iterations at the 20 minute mark.

I can imagine these processes just getting better and better.

 

[Kvalsvoll]

Having seen discussions and some graphs on the preference of speakers and noticing how the gap lessens as we go from mono to stereo , what about multichannel?

Is there any evidence for the audible benefits of supposedly better surround speakers ?

Have there been any studies to see how multichannel might effect our ability to discern the performance differences between speakers in a multichannel (beyond 2) setup ? And how it effects our acoustic experience at home?

I did experiments to investigate effects of different radiation patterns, several years ago, when I designed some surround speakers. While this was intended for movies, it will not be different for music. This is what I found:

- A wide and precisely defined pattern produced a significant improvement in phantom image solidity and immersion compared to the pattern from a normal small speaker. The normal pattern sounded like speakers, while the wide-defined was like sounds appearing totally disconnected from the speakers.
- Too wide, and dipole, gave a diffuse pattern, speakers disappear, but no precsision and solidity in images.
- Accuracy in polar pattern is not very important.
- The wide-defined surround speaker had very good sound quality when used as intended, but did not perform so good when used as ordinary front main speaker.

 

[jhaider]

Is there any evidence for the audible benefits of supposedly better surround speakers ?

With the caveat I've never tried a really bad speaker, I have not found surround speaker quality to be a significant factor in overall sound when listening to music in native or upmixed multichannel, or live sports. (Live sports is the only non-musical A/V content we regularly enjoy that exercises the surrounds.) While using speakers from the same line, or even the same model, is often recommended as "ultimate," I have found no sonic benefit from doing so. The flip side is, a system with disjointed looks may be perceived as sounding disjointed too, and a system that looks impressive will often be perceived to sound impressive! That effect is heightened in a system comprised of freestanding speakers at all positions in an open floor plan room. So it makes sense to use identical or same-line speakers for looks on those grounds.

There are exceptions. For example, if a surround speaker is too close to the listener the sound field can collapse onto the closer speaker. That's when the old fashioned "dipoles" Dr. Toole loathes make practical sense.

The X-factor may be height speakers for immersive setups. While we're not physiologically or psychoacoustically well set up to hear sounds behind us with as much detail and precision as sounds in front of us, current immersive systems include front height speakers. I have not experimented with different quality heights. The selection process for our front and rear heights was 4-pronged with: speakers with

1. with built in wall mounting points,
2. height less than the wall space between the window casings and crown moldings in front of our living room,
3. high quality finish, and
4. an even radiation pattern

The pick ended up being Tannoy Revolution XT Mini. But for the requirement 2, the Revel Performa3 bipoles would have been the likely pick. I wanted to try JBL Control HST, but the wife nixed them on looks. There is also an Infinity HST version with an ugly vinyl wrap, but often an attractively discounted (US) price.

I did experiments to investigate effects of different radiation patterns, several years ago, when I designed some surround speakers. While this was intended for movies, it will not be different for music. This is what I found:

- A wide and precisely defined pattern produced a significant improvement in phantom image solidity and immersion compared to the pattern from a normal small speaker. The normal pattern sounded like speakers, while the wide-defined was like sounds appearing totally disconnected from the speakers.
...
- Accuracy in polar pattern is not very important.

What's the distinction between "wide defined" (compared to a "normal small speaker," which I interpret to mean a standard flat-waveguide 2-way with dispersion disruption at the crossover) and "accurate polar pattern?"

Can you post polar maps showing examples of each?

 

[richard12511]

- Too wide, and dipole, gave a diffuse pattern, speakers disappear, but no precsision and solidity in images.

This has been my experience with Panel speakers.

 

[Sancus]

I'll never understand the fixation on frontal music and replicating perceived "live events". We have whole genres listened to by hundreds of millions of people that are generated electronically, with no natural acoustics involved at all. But moving direct sounds away from the frontal stage is somehow crazy? Pink Floyd would disagree given that 40+ years ago they were doing concerts in surround(quadrophonic) with substantial parts of the recording outside the front channels.

As far as the recording process goes, pretty much all modern albums going back decades are multitracked to sometimes dozens of separate tracks, so mixing them in 2 or 5 or 15 channels is no different in terms of the creativity of the process. At the end of the day, mixing is just as creative a process as playing music.

To me if it sounds engaging I'm happy with it, and I care absolutely nothing for perceived purity in a medium that is entirely manufactured to begin with, which is what all stereo is. The good multichannel recordings out there(and yes, they are relatively few, but that is still in the hundreds or thousands for classical) sound much better than the best stereo recordings to me, so that's why I seek them out. But if I don't like the music or the mix, of course I need not listen to it!

 

[Kvalsvoll]

With the caveat I've never tried a really bad speaker, I have not found surround speaker quality to be a significant factor in overall sound when listening to music in native or upmixed multichannel, or live sports. (Live sports is the only non-musical A/V content we regularly enjoy that exercises the surrounds.) While using speakers from the same line, or even the same model, is often recommended as "ultimate," I have found no sonic benefit from doing so. The flip side is, a system with disjointed looks may be perceived as sounding disjointed too, and a system that looks impressive will often be perceived to sound impressive! That effect is heightened in a system comprised of freestanding speakers at all positions in an open floor plan room. So it makes sense to use identical or same-line speakers for looks on those grounds.

There are exceptions. For example, if a surround speaker is too close to the listener the sound field can collapse onto the closer speaker. That's when the old fashioned "dipoles" Dr. Toole loathes make practical sense.

The X-factor may be height speakers for immersive setups. While we're not physiologically or psychoacoustically well set up to hear sounds behind us with as much detail and precision as sounds in front of us, current immersive systems include front height speakers. I have not experimented with different quality heights. The selection process for our front and rear heights was 4-pronged with: speakers with

1. with built in wall mounting points,
2. height less than the wall space between the window casings and crown moldings in front of our living room,
3. high quality finish, and
4. an even radiation pattern

The pick ended up being Tannoy Revolution XT Mini. But for the requirement 2, the Revel Performa3 bipoles would have been the likely pick. I wanted to try JBL Control HST, but the wife nixed them on looks. There is also an Infinity HST version with an ugly vinyl wrap, but often an attractively discounted (US) price.



What's the distinction between "wide defined" (compared to a "normal small speaker," which I interpret to mean a standard flat-waveguide 2-way with dispersion disruption at the crossover) and "accurate polar pattern?"

Can you post polar maps showing examples of each?

By "Wide defined" I mean here a pattern that covers a wider angle, and then falls off beyond that, and in the case of the S1.2 surround speaker used in the evaluation this pattern covers a wider frequency range down into lower midrange frequencies.

To be able ot make a surround speaker inside reasonable cost and size constraints, some compromises had to be made. One is that the polar pattern is not completely smooth, the frequency response changes when you move around. Due to how our hearing works, this is not so important for a speaker intended to be placed behind or at least 90 degrees off to the sides. Listening confirmed that the benefits of this wide pattern outperforms the point source, even though the point source has a smooth polar response.

 

[Kal Rubinson]

As far as the recording process goes, pretty much all modern albums going back decades are multitracked to sometimes dozens of separate tracks, so mixing them in 2 or 5 or 15 channels is no different in terms of the creativity of the process.

That does not generally apply to recordings of classical music.

 

[Sancus]

That does not generally apply to recordings of classical music.

Yeah that's definitely true, I just didn't go into it to keep that section brief. It seems like classical recording has its own specialized set of recording history and techniques, and I'm sure you're better informed than me on this topic! There seem to be a lot of novel techniques and innovation in the space and it's a complex topic.

Classical recordings deliberately for surround are so much more common than any other genre, too.

 

[Blumlein 88]

My own little experiements indicate like most things audio, FR is important (which contributes to timbre of course). Recording in a space for 5 channels and playing it back, if you curtail the response of the surrounds or swap speakers you hear the sound field change. OTOH, we are less picky as you add channels. This is sort of off the cuff, but maybe on a 10 point scale we rate speaker A as an 8, and speaker B as a 4 in mono. In stereo maybe it becomes an 8.5 and 6 respectively. In 5 channel surround we might rate them a 9 and 7.5 respectively. There still is a quality difference just smaller and it likely will bother us less. So the real question, assuming you have recordings of good quality in both formats, if I have $10K (or $2k) for speakers will I like 5 channels for that money or 2 channels. Will the improving speaker quality offset the fewer channels to make me like it more? I've changed my mind on this over the years, and there are plenty of caveats, but if you have the recordings I think you'd be happier with 5 channels for the given amount of money.

There are limits to this of course. Let us say I purchase 5 LSR305's at a cost of $750. Those alone in a moderate or smaller room could be good. If you've a 5 channel recording you'll like it much better than if you purchased a pair of $750 speakers and listened to the recording in stereo even though the speakers could be of better quality than the LSR305s.

Now about the limits, if you spent $300 for a pair of LSR305's in a room suitable, I don't think you can find 5 speakers for $300 which will equal the stereo version. But from about that price point on up I think MCH can win on satisfaction with good recordings.

 

[Blumlein 88]

J_J has done research into perceptual soundfield reconstruction. It involves 7 microphone recordings for 5 channel playback. And resolves height.
https://www.stereophile.com/content/wheres-real-magazine-we-see-it-february-2001

Unfortunately it belongs to Bell Labs and was never made into anything available to us.

I do so much wish this or some similar version had become the standard. Like stereo it would be a method of playback which could last a long time, and recordings could be made for that destination. Unlike what we have with Dolby and others who are just adding channels, adding processing, not even trying to really do much more than effex (which sometimes is also okay on mch music). You can use gear from circa 1960 today for correct stereo playback. I wish we had a 5 channel version of that. Instead you need to update your AV gear every couple of years or at least every few years. Employing a room for 5 channel playback is rare enough does it really seem like a growth industry to have Auro/Atmos 11.2 or more channels in people's homes?

Plus if 5 channels is enough for perceptual soundfield recreation you don't need the extra channels. Stereo is not enough though it gives a satisfying fascimile for many purposes. Perceptual soundfield recreation like stereo uses some of the ways our ears can be tricked to create what we'll hear as the full soundfield while bypassing the need to actually recreate physically the soundfield at the recording location.

 

[Sancus]

J_J has done research into perceptual soundfield reconstruction. It involves 7 microphone recordings for 5 channel playback. And resolves height.
https://www.stereophile.com/content/wheres-real-magazine-we-see-it-february-2001

Unfortunately it belongs to Bell Labs and was never made into anything available to us.

I do so much wish this or some similar version had become the standard. Like stereo it would be a method of playback which could last a long time, and recordings could be made for that destination.

How many recordings are even attempting to "record a space" outside of classical and some jazz though? In any genre where multi-track is the standard, "the space" is just created at the mixing stage and there's no barrier to producing multi-channel recordings even with very old recordings, so long as they were preserved.

I think the barrier to multi-channel adoption is interest and distribution more than anything else. No streaming service has seriously picked it up(I'm not counting Amazon's "Atmos" that is only distributed for... two channel on mobile devices, lol... if they let this out to AVRs that will be something special) and physical media is at the point where it's niche even for stereo.

In theory, the object-based sound formats make it possible to distribute 1 set of music data for an album and then adapt it to as many channels as you like at the playback stage. This seems like the most logical extension of recording to me, but its adoption speed is heavily restricted by the amount of proprietary IP and industry players all trying to extract money at various stages of the process.

 

[Head_Unit]

I'am very skeptical how much sense it would make to listen to the reverbed sound on back speakers in your room. It's imo fu**ing complicated to get in this way real room Informationen. To be honest, i dont think it's possible.

Once upon a time, I was at the house of a friend who worked for Harman/JBL. He was playing I believe a DTS demo disc of Mahler, sung by a choir on freakishly steep risers. I didn't hear any rear channels, but he stuck his ear to one and said it was working. He switched the track from 5.1 to stereo, and the sound field absolutely collapsed. THAT is what multichannel can do...when done well that is.

In the case of pop music, I was very lucky to hear Fleetwood Mac's Rumours and Emerson Lake and Palmer's Brain Salad Surgery* sitting at the consoles they were mixed at with the guys that mixed them. There was really no comparison, the stereo mixes had much less space and were more crowded. That was particularly true for Rumours where the mixers stated they had cut stuff out originally because they just couldn't cram so much stuff into two channels, and they were tweaking every instrument by 1/4 of a decibel. In 5.1 they could insert back some stuff they cut and let everything breathe. That was my impression also of John's Chicago II mix where the rears were not used much for crazy effects but to open up the stereo field and clarify everything. You can also hear this to some extent on the Chicago Quadio box which is one of the best values EVER in a box set, and due to rare use of the master tapes, sounds fabulous to boot.

*John Kellogg's original DVD-A mix. I am very very curious why a different mix was commissioned later on.

 

[maverickronin]

How many recordings are even attempting to "record a space" outside of classical and some jazz though? In any genre where multi-track is the standard, "the space" is just created at the mixing stage and there's no barrier to producing multi-channel recordings even with very old recordings, so long as they were preserved.

I can understand why some people would want to try and reproduce a live event with a recording but personally, it seems old fashioned to me. It's like comparing a recording of a live play to movie.

I think the barrier to multi-channel adoption is interest and distribution more than anything else. No streaming service has seriously picked it up(I'm not counting Amazon's "Atmos" that is only distributed for... two channel on mobile devices, lol... if they let this out to AVRs that will be something special) and physical media is at the point where it's niche even for stereo.

They just keep adding channels endlessly so they have something new to license every few years. The whole industry is going to implode if it keeps going at this rate.

In theory, the object-based sound formats make it possible to distribute 1 set of music data for an album and then adapt it to as many channels as you like at the playback stage. This seems like the most logical extension of recording to me, but its adoption speed is heavily restricted by the amount of proprietary IP and industry players all trying to extract money at various stages of the process.

This would be absolutely amazing, but it even if it does happen it will always remain niche. There will never be an open and standardized format which will lock out small labels and independents and adoption rates will be stifled by the need for brand new hardware and software.

 

[Jon AA]

In theory, the object-based sound formats make it possible to distribute 1 set of music data for an album and then adapt it to as many channels as you like at the playback stage. This seems like the most logical extension of recording to me, but its adoption speed is heavily restricted by the amount of proprietary IP and industry players all trying to extract money at various stages of the process.

Exactly. But even beyond the problems of adoption...is getting these people to adopt their own technology! Both Atmos and DTS-X promised us this and still advertise on this basis. Neither delivers.

The vast, vast, majority of DTS-X encoded soundtracks contain zero objects. Just channels--wouldn't need any "format" for that at all. Atmos home releases have a limited number of objects they can use and most don't even use those to the extent they could (at least on the ground layer).

 

[Kvalsvoll]

I can understand why some people would want to try and reproduce a live event with a recording but personally, it seems old fashioned to me. It's like comparing a recording of a live play to movie.



They just keep adding channels endlessly so they have something new to license every few years. The whole industry is going to implode if it keeps going at this rate.



This would be absolutely amazing, but it even if it does happen it will always remain niche. There will never be an open and standardized format which will lock out small labels and independents and adoption rates will be stifled by the need for brand new hardware and software.

But it already has (imploded). Home theater has collapsed, only those with dedicated home cinemas are left, the rest throws out surround speakers and installs a soundbar instead. Because we do not want speakers on all walls and then being told we need speakers all across the ceiling as well. And a typical living-room install never delivers the sound experience that was promised, partly due to lack of acoustic treatment.

And to me, it makes more sense to have a decent 2-ch. For music, surround is not necessary, movies is a different story, but if music is the main purpose, a good 2-ch system does the job better than a multich at the same total cost.

Think about it. Music is something that happens up front, rarely are there instruments moving around in the venue. I have some examples, like the Hot Sardine's LIve at Joe's pub, but this is the exception. And even this one manages to move the tuba instrument around me, when the musician walks around in the pub. 2 channels can actually place instruments and sound all over the place, admittedly not with the same precision as a good multich. A q-sound recording can sound better in a good 2-ch than a mediocre surround system.

The only sound from sides and back in normal music productions are reflections that defines the acoustics of the recording venue. And that can be very well presented in a good 2-ch system. The images of instruments and sound up front will have a solidity and presence that most theater systems only can dream of.

2-ch works best seated exactly centered between the speakers, but even here a good system+room will work very well across the whole width of the room, presenting a good sense of sound-stage all over, with no center channel.

 

[Kal Rubinson]

The only sound from sides and back in normal music productions are reflections that defines the acoustics of the recording venue. And that can be very well presented in a good 2-ch system.

You can assert this but it, in my daily experience, is simply not true.

The images of instruments and sound up front will have a solidity and presence that most theater systems only can dream of.

If so, it says that the multichannel system setup is flawed.

 

[jhaider]

By "Wide defined" I mean here a pattern that covers a wider angle, and then falls off beyond that, and in the case of the S1.2 surround speaker used in the evaluation this pattern covers a wider frequency range down into lower midrange frequencies.

Thanks for the model number.

https://www.kvalsvoll.com/Designs/S1_2/S1_2_Folder_1.pdf

Interesting approach. Your speaker is conceptually similar to old Infinity/Revel switchable-pattern designs (seemingly abandoned), though those are wider coverage. However, this speaker seems complex for 100º horizontal coverage. If that's the approximate desired coverage, designs such as Revel M16 (and presumably the identical-in-relevant-part S16 on wall variation) or the better small pro monitors get us there with smoothness to boot. CBT also looks more promising in this role.

 

[DDF]

I have two very good 2 channel systems and a modest (but very well set up) 5.1 rig. The 5.1 set up provides enormous (but different) pleasure and accuracy for chamber music and orchestral recordings.
I attend many classical concerts across many venues and have purposely listened for envelopment in the hall or church and then listened at home. The improvement with 5.1 for me is similar to the improvement gained by adding a good sub. Once taken away (reverting to stereo), its absence is stark and unrealistic. It's unmistakable.
For rock or club jazz, none of this is typically true IME.

 

[Kvalsvoll]

Thanks for the model number.

https://www.kvalsvoll.com/Designs/S1_2/S1_2_Folder_1.pdf

Interesting approach. Your speaker is conceptually similar to old Infinity/Revel switchable-pattern designs (seemingly abandoned), though those are wider coverage. However, this speaker seems complex for 100º horizontal coverage. If that's the approximate desired coverage, designs such as Revel M16 (and presumably the identical-in-relevant-part S16 on wall variation) or the better small pro monitors get us there with smoothness to boot. CBT also looks more promising in this role.

This speaker never made it to become a commercially available product, so you can not buy it, the only ones built are the prototypes now in service in my media room. Just mentioning, so this does not come off as some kind of spam advertisement.

If I were to improve the design, I would omit the switchable pattern and just have one - the wide option. You don't need anything else. Reduces cost, simpler to explain and install.

What is important to understand is that the pattern is not just about angle of coverage. A point source will have a spherical wavefront, while these have a more defined pattern, where sound energy is more even inside the covered area, and falls off steep outside. Radiation control is also present across a wide frequency range.