Speaker Positioning = Frequency or Time Correct Not Both?

[The Computer Audiophile]

I was daydreaming a bit today about speaker positioning, I know ridiculous, and something dawned on me. I don’t know if my thinking is right, so I’d love to have a discussion about it. 
 

My thought is this: When positioning speakers in a room for the best sound, it’s nearly impossible for both the frequency and time domains to be correct simultaneously, because of room issues. 
 

For example, we often place speakers in a rough triangle, then nudge them in all directions until we get the sound we like. Or, we place the speakers in a perfectly measured triangle and leave them. 
 

In the rough triangle, the drivers can’t be equidistant to the listener’s ears. In the perfect triangle, the drivers are equidistant but the chances of the frequency being correct without adjustments is nearly zero. 
 

Without using DSP for time and/or frequency correction, are people only getting one of these right?
 

 

[GregWormald]

Given that the wavelength of a 10,000 Hz tone is less than 1.5 inches, to get the wavelengths correct we would have to sit in precisely the correct spot and not move our heads at all. A ¾ inch movement would result in massive cancellations and wave interferences!

 

Measuring and using DSP would not change the physics.

[fas42]

As in the above post, getting speaker positioning "perfect" is a daydream ... what is almost always overlooked is that the human hearing system is remarkably capable of adjusting for just about everything, if you give it half a chance. The last bit is key - a high percentage of rigs are 'fighting' the ear/brain, by blurring or discarding the fine detail which allows our hearing to make sense of it all.

 

The very good measure is, how true mono material comes across, over stereo speakers: "good enough" reproduction throws up a soundstage which 'follows' you as you move left and right - the ear/brain interprets the phase differences of the sound coming from left and right speakers as you do this as meaning that the sound source is moving in sync with your location.

[STC]

Frequency response is never flat at your ears and it is not meant to be flat as the response changes with the position of the speaker for localization. Unlike a single mic, we have two ears which is about 17 to 20 cm apart. Flat response is not critical or important in stereo configuration. 

[yamamoto2002]

Typical stereo system have only 2 speakers, it can project surprisingly plausible 3D sound stage illusion but it never capture/reproduce physically identical sound space of recorded venue.

 

From my experience, increasing speakers to order of 10 do not improve very much.

sound source positioning becomes more sharp, especially positioning of the sound source close to ears is improved, it can express smaller sound stage than the audio room with improved perceived positional resolution of sound source. But it cannot express larger room sound than audio room.

 

In-ear canal earphones can be perfectly flat and phase/time correct whereever the head rotates/moves. But sound source direction/distance is more ambiguous and relatively less perceptible with earbuds because human auditory cognition uses sound arriving timing difference of left ear and right ear positioning and earlobe shapes to reconstruct 3d sound source positioning. I read somewhere human sound source positioning accuracy is improved by moving / rotating heads slightly while listening

 

In recent years, accelerometer is installed onto some earbuds to get position and direction of ears in 3d space. With this head pose info and HRTF DSP combined, realistic holographic 3D sound stage can be reproduced, in theory.

this tech is still in early stage but it has potential to reproduce sound stage of 2000 seats concerto hall. IMO it is never possible with typical 2 speakers audio room

[fas42]

13 hours ago, yamamoto2002 said:

IMO it is never possible with typical 2 speakers audio room

 

Have a person walk into a totally dark space. And then have some sounds made. Most people will be able to interpret from the echos the general size and space of the invisible area they're facing. Now, replace that environment with an outdoor situation which is black - blindfolding the person would work - and playback recordings made of the indoor sounds just mentioned, over a high quality setup: the individual will hear those indoor spaces. Final step, do the previous step in a normal listening room ... most audio people would say, from experience, that the two acoustics would confuse the listening brain - however, my experience is that the ear/brain is far more capable than that; say, have a person listen to live sounds made simultaneously in two completely different, side by side spaces - he will hear the two spaces as separate events, and have little difficulty on focusing on one or the other. You can often get this on a recording: a solo voice is in a tiny recording booth just behind the speakers, and the backing orchestra stretches out in the distance.

 

Which is what capable replay over stereo does: it projects a completely different acoustic to the ear/brain, which the mind can easily follow as having no connection to the space you physically occupy.

[The Computer Audiophile]

7 minutes ago, fas42 said:

and playback recordings made of the indoor sounds just mentioned, over a high quality setup: the individual will hear those indoor spaces

And the sounds that were appropriately all around the person when indoors, now all come from in front of them, including the rear wall bounce. Talk about disorienting. 

[fas42]

12 minutes ago, The Computer Audiophile said:

And the sounds that were appropriately all around the person when indoors, now all come from in front of them, including the rear wall bounce. Talk about disorienting. 

 

So, you don't believe he would hear those spaces captured - be able to distinguish say between an intimate setting versus an expansive hall?

[The Computer Audiophile]

1 hour ago, fas42 said:

 

So, you don't believe he would hear those spaces captured - be able to distinguish say between an intimate setting versus an expansive hall?

To be honest, it all sounds so strange to me that I really can’t follow your example. 
 

Playing all captured sounds (front, sides, rear) out of the front two channels places the listener outside the listening space peering in. 
 

Anyway, nothing to do with my original post or topic of this thread. 

[fas42]

The topic is about, is there such a thing as "perfect" speaker positioning ... my response is that such is impossible, and in fact is not needed ... I'm happy to have the plane of the speakers be a very large window onto the recording space, because by simply increasing the volume the experience can be as, say, immersive as one would like it to be.

[The Computer Audiophile]

15 minutes ago, fas42 said:

The topic is about, is there such a thing as "perfect" speaker positioning ... my response is that such is impossible, and in fact is not needed ... I'm happy to have the plane of the speakers be a very large window onto the recording space, because by simply increasing the volume the experience can be as, say, immersive as one would like it to be.

The topic is about whether one has to setup speakers for time or frequency, not both.  

[bobfa]

23 hours ago, The Computer Audiophile said:

Without using DSP for time and/or frequency correction, are people only getting one of these right?

For me, I feel that time is more critical.  When the time is out of wack, the illusion of space drops off very quickly.    I try to optimize the distances from the tweeters (big assumption) to the listening space. I am not sure I can quantify where time is a problem.  Is it a rarange of frequencies?  Is it the lizard brain, or is in the frontal lobe?

 

I lean towards nearfield systems for critical listening, and a bias of mine.  

[The Computer Audiophile]

16 minutes ago, bobfa said:

For me, I feel that time is more critical.  When the time is out of wack, the illusion of space drops off very quickly.    I try to optimize the distances from the tweeters (big assumption) to the listening space. I am not sure I can quantify where time is a problem.  Is it a rarange of frequencies?  Is it the lizard brain, or is in the frontal lobe?

 

I lean towards nearfield systems for critical listening, and a bias of mine.  

I hear ya. When the timing of my immersive system was off because of a bad microphone, it drove me nuts. It was off by a millisecond or two at most, but it was immediately obvious. 

[STC]

5 hours ago, The Computer Audiophile said:

And the sounds that were appropriately all around the person when indoors, now all come from in front of them, including the rear wall bounce.

Contrary to the popular myth, stereo recordings do not have the rear wall reverb. If you were to recorded music with the frontal reverb or artificial reverbs like in most recordings, the sound will be more or less like listening with headphones. 
 

3 hours ago, The Computer Audiophile said:

It was off by a millisecond or two at most,

 

It is impossible that you can tell one or 2 ms difference. all delayed sound within the threshold of echo is perceived or fused as single sound to the ear/brain. 

[The Computer Audiophile]

7 hours ago, STC said:

Contrary to the popular myth, stereo recordings do not have the rear wall reverb

If it was captured by the microphones, then it can be played by the stereo. 
 

 

 

7 hours ago, STC said:

It is impossible that you can tell one or 2 ms difference

I encourage you to sit it my listening chair and hear the difference. When one or two speakers are off by one or two milliseconds, the soundstage is far different from when the speakers are time aligned. It isn’t even close. 

[The Computer Audiophile]

On 10/4/2023 at 9:52 PM, STC said:

Frequency response is never flat at your ears and it is not meant to be flat as the response changes with the position of the speaker for localization. Unlike a single mic, we have two ears which is about 17 to 20 cm apart. Flat response is not critical or important in stereo configuration. 

Who’s looking for flat response?

[STC]

26 minutes ago, The Computer Audiophile said:

If it was captured by the microphones, then it can be played by the stereo

Chris, sound is same no matter from which direction it comes. We localize direction based on the HRTF. For sound to be perceived to be coming from the back, the pinna changes the response and you compare the difference and determine the direction. It is possible to create some effect for short period to trick the brain by changing the frequencies to mimic how the sound would be when it is coming from different direction.

 

In recording, the mic does not have pinna. So it just captures whatever sound reaches it . there is no front or back to it. You cannot play the recording in stereo with the front speakers and expect the back reverbs to come from the back. If that is possible, then there is no need for the multichannel rear speakers.

 

Almost all recordings are made either with the placement close to the source or about 1/3 of the critical distance for stereo so that the reverbs is limited to the frontal stage. Just take your phone make random noise with different object and ask others to guess the direction.

 

27 minutes ago, The Computer Audiophile said:

I encourage you to sit it my listening chair and hear the difference. When one or two speakers are off by one or two milliseconds, the soundstage is far different from when the speakers are time aligned. It isn’t even close

Your comment was made referring to your immersive system. For stereo, the front speakers are supposed to be placed at equal distance. If you are having 1 or 2 ms difference then your right and left speakers distance should be at 1 to 2 feet difference. Then it make sense why 1ms is making the difference in your system. 

[The Computer Audiophile]

24 minutes ago, STC said:

Chris, sound is same no matter from which direction it comes. We localize direction based on the HRTF. For sound to be perceived to be coming from the back, the pinna changes the response and you compare the difference and determine the direction. It is possible to create some effect for short period to trick the brain by changing the frequencies to mimic how the sound would be when it is coming from different direction.

 

In recording, the mic does not have pinna. So it just captures whatever sound reaches it . there is no front or back to it. You cannot play the recording in stereo with the front speakers and expect the back reverbs to come from the back. If that is possible, then there is no need for the multichannel rear speakers.

 

Almost all recordings are made either with the placement close to the source or about 1/3 of the critical distance for stereo so that the reverbs is limited to the frontal stage. Just take your phone make random noise with different object and ask others to guess the direction.

 

Your comment was made referring to your immersive system. For stereo, the front speakers are supposed to be placed at equal distance. If you are having 1 or 2 ms difference then your right and left speakers distance should be at 1 to 2 feet difference. Then it make sense why 1ms is making the difference in your system. 

I think we are talking about different things. A single mic certainly can’t detect direction. I’m talking about a multi-mic recording capable of capturing a real space. 

[Confused]

Some of the words in the OP certainly relate to my system. My room is rectangular, but it does have some asymmetry. On one side there is a window and curtains, but not on the other side, which does have a door further down, and so on. 

 

The speakers are positioned to create a triangle, but if I measure in the theoretical "sweet spot", there are some small differences in the frequency response curves between left and right.

 

With a room with even small asymmetry, I do not believe that any speaker positioning would achieve identical frequency response curves between left and right. You might get close, but an exact match would be near to impossible.

 

If nothing else, the above makes me happy that DRC seems to work well for my system and ears. 

[Confused]

 

On 10/5/2023 at 3:50 AM, fas42 said:

The very good measure is, how true mono material comes across, over stereo speakers: "good enough" reproduction throws up a soundstage which 'follows' you as you move left and right - the ear/brain interprets the phase differences of the sound coming from left and right speakers as you do this as meaning that the sound source is moving in sync with your location.

[STC]

50 minutes ago, The Computer Audiophile said:

I think we are talking about different things. A single mic certainly can’t detect direction. I’m talking about a multi-mic recording capable of capturing a real space. 

Multi mic recordings meant to be used in multi channel production and depending on the reverbs of the front channel, the other mics captured reverbs maybe used to have the correct reverbs for the front speakers. 
 

I think I posted here somewhere how the 2L Cube mics used for different formats including stereo. One mic for one speaker. The capture for Aura 3D but for stereo they only use what was captured by the two speakers.  You cannot have rear reverbs direction from front speakers. Even when you mix them, which happens often it is still the frontal reverbs. 
 

[bobfa]

Interaural time difference. (2023, July 17). In Wikipedia. https://en.wikipedia.org/wiki/Interaural_time_difference  While I hate quoting secondary sources, the above article has good references.  The 100ms ITD is interesting.  I read somewhere but cannot find it right now that a trained ear can hear timing differences in the microsecond ranges.  I am not sure.

 

 

I go back to mechanical speaker designs like the Theil speakers that align the drivers with a sloped cabinet or modern concentric designs that do the same.  These are critical to making the speaker sound good.  

 

All this makes me think we can hear minute differences in arrival time.  I think we sense this in the sound stage and instrument location.  

 

 

[STC]

33 minutes ago, bobfa said:

Interaural time difference. (2023, July 17). In Wikipedia. https://en.wikipedia.org/wiki/Interaural_time_difference  While I hate quoting secondary sources, the above article has good references.  The 100ms ITD is interesting.  I read somewhere but cannot find it right now that a trained ear can hear timing differences in the microsecond ranges.  I am not sure.

 

 

I go back to mechanical speaker designs like the Theil speakers that align the drivers with a sloped cabinet or modern concentric designs that do the same.  These are critical to making the speaker sound good.  

 

All this makes me think we can hear minute differences in arrival time.  I think we sense this in the sound stage and instrument location.  

 

 

I am afraid you misunderstood the 100ms. The 100ms is the reference to the duration of the white noise to measure the ITD. ITD for human only matters for not more than 700 μs depending on your head size. 

[bobfa]

Thanks this was confusing to me!

[Nikhil]

In my opinion you can get Time and Frequency right for a simple 2 way mechanical system. If by frequency you mean phase that is.  You will need to have separate enclosures for the tweeter, mid range and bass drivers (in a 3 way system).

 

In a 2 way system you will need to move the tweeter back a few inches until the sound appears to emanate from the mid range driver. Use a familiar track with something like male vocals to dial this in.  Once you get the sound "seeming" to emanate from the mid range driver you have both time and phase alignment.   In a 3 way - you will have to move either the bass unit or the tweeter/mid range "locked" pair.  Once you get the sound emanating as a whole from a fixed point in the setup you have time + phase alignment for that speaker.  In reality this is a little more complex because you need to move the drivers in a 3D space and is why manufacturers like Wilson Audio have their speaker setup the way they do.  But they are essentially doing the same thing.

 

You then have to setup the second speaker in a similar fashion for a speaker pair in a stereo setup.  The speaker pair would then have to be positioned in the room as per the usual speaker setup procedure taking into account room modes etc.  But as you can guess the number of variables increases immensely.

 

So to answer your question - trying to do this without DSP is almost impossible! 😂😂 

But my point is that you need to time align your drivers first before you attempt room positioning.    

 

 

 

.