Soundstage and imaging

[STC]

edit. hmm, if speakers are all around its not possible to switch distance I think, either the speakers are close enough, or far enough, moving closer to front speakers would make you further from back speakers. Have you experimented with this, does it make any difference?

Use time delay.

I only have limited exposure Griesinger's work and limited to the envelpment and therefore I cannot comment on the rest.

Cheers!

 

[tmuikku]

Hi, but the delay does not change amount of sound in room which is constant unless acoustic properties of the room or the speaker(s) was manipulated Amount of direct sound can be changed with toe-in and moving closer / farther away from listener. Some people might have nice room acoustics so it's good sound pretty much everywhere, while others with poor rooms could perceive great change when brain does its work. Well, just trying to give insight for everyone to experiment with their setups and perception.

Thanks for your comments and insight so far! I wish you fun time with music!

 

[goat76]

One straightforward method for finding the optimal distance between loudspeakers, also optimal toe in that I would suggest is by using transients such as this:

To elaborate, note that this kind of waveform contains pitch down sinewave starting from about 17kHz to about 44Hz. However, if you look closely at the time related to the mouse cursor, it appears that the signal pitches down from 17kHz to about 90Hz or so in 2-3 milliseconds. This goes down way too fast for us to perceive it as anything but a "click", with a substantial weight of mid-bass.

Optimal setup should have a very clear and pinpoint focus on that. A "click" should be very tiny, floating exactly in the middle of phantom center, augmented by very focused and strong mid-bass. The rest of the waveform contains low frequencies and on a system which is capable enough, you should get a very distinct and focused "kick", followed by envelopment of lowest of frequencies and soft decay. Low frequencies are long in duration in comparison but should never mask the higher frequencies. Subjectively, it should be perceived all at once, but high to mid frequencies should never be masked or loose focus. Also, they should be loud enough in comparison to lows.

You can start wider with no toe in, and then gradually decrease the distance and introduce some degrees of toe in. This will depend on speaker directivity. The "sweetspot" is usually where mid-bass is strongest with the same level of amplification.

It is hard to describe, but once you get it right, what you hear can be an "aha" moment. On the headphones, the pinpoint click would obviously be right in the middle of your head. On loudspeaker system, it should be even more distinct right in the middle of the phantom center. To test the image stability, you can move yourself out of the comfort zone of your MLP and see if you loose any focus.

To me, this method is all in one to show what your loudspeakers are doing and what your room is doing. It shows direct to reverberant ratio in a very distinct and simultaneous manner.



I find it funny that usually there's so much gear in between the speakers that they can't be setup any closer, form over function

Another and more controlled way to check the phantom center than relying on unknown recordings of full mixes is to use a DAW of some sort (there are free ones out there) and download a few sound samples recorded in mono, like kick drums, snare drums, vocals, and other sounds that are typically panned to the center in most mixes. Free samples can be found here: https://freesound.org/search/?q=kick , just search for different types of samples and make sure they are mono recordings.

In the DAW it's easy to listen to how those samples sound panned hard to one of the loudspeakers, and then compare how that exact same sound sample sounds when panned to the phantom center. The phantom-centered sound should sound almost as distinct as the hard-panned sound and not apparently wider or washed-out, it will probably sound a little less focused because of the stereo fault and the cancellations that occur by the crosstalk from the two loudspeakers, but try to concentrate on the width and the pinpoint localization of the sound sample. The important thing here is that you can be sure that this mono sound sample is exactly the same no matter where it's panned in the stereo field, and should therefore have the same width no matter where it is located in the mix.

A few years ago I made a test file with a snare drum playing the first four beats in the middle, the next four beats to the left, and the last four beats to the right. This can be used to evaluate the phantom center, but it's probably even better if you download a few different sound samples of a variety of sound objects to get an even better idea of the center image.

Anyway, here is my sound file of that snare drum: https://www.dropbox.com/s/1o5t4wifgafv33c/Snare Beats Center Left Right.wav?dl=0

 

[STC]

Hi, but the delay does not change amount of sound in room which is constant unless acoustic properties of the room or the speaker(s) was manipulated Amount of direct sound can be changed with toe-in and moving closer / farther away from listener. Some people might have nice room acoustics so it's good sound pretty much everywhere, while others with poor rooms could perceive great change when brain does its work. Well, just trying to give insight for everyone to experiment with their setups and perception.

Thanks for your comments and insight so far! I wish you fun time with music!

As I mentioned earlier, the room’s RT is low. It is probably 0.2s. The last measurement was with partial damping it was 0.28 and since I have covered the entire room except the floor but didn’t measure. LF is not problem as this is a room in room with 1 foot thickness filled with 4inch Rockwool and 4 inch void before another 4 wall. Practically, even with the large speakers and two subwoofer the bass is clean. Even in concert halls you have reflection from objects and persons immediate to you but the RT still continues for about 2s. So kind of same principle.

Mine is a mini copycat version of the Ambiophonics institute but now rather handicapped with the retirement of Ralph who mentored me over 10 years. I still think improvements could be made but I need to utilize twice the impulse responses I have and even with 56 impulse responses the CPU is already more than 80 %. I can only move to the next step when I can get a CPU thrice the power of my current one at the price point I could afford.

 

[tmuikku]

Nice, yeah perhaps it works very well and not particularly subject to this critical distance thing. I can only imagine the experience

 

[audiofooled]

Another and more controlled way to check the phantom center than relying on unknown recordings of full mixes is to use a DAW of some sort (there are free ones out there) and download a few sound samples recorded in mono, like kick drums, snare drums, vocals, and other sounds that are typically panned to the center in most mixes. Free samples can be found here: https://freesound.org/search/?q=kick , just search for different types of samples and make sure they are mono recordings.

In the DAW it's easy to listen to how those samples sound panned hard to one of the loudspeakers, and then compare how that exact same sound sample sounds when panned to the phantom center. The phantom-centered sound should sound almost as distinct as the hard-panned sound and not apparently wider or washed-out, it will probably sound a little less focused because of the stereo fault and the cancellations that occur by the crosstalk from the two loudspeakers, but try to concentrate on the width and the pinpoint localization of the sound sample. The important thing here is that you can be sure that this mono sound sample is exactly the same no matter where it's panned in the stereo field, and should therefore have the same width no matter where it is located in the mix.

A few years ago I made a test file with a snare drum playing the first four beats in the middle, the next four beats to the left, and the last four beats to the right. This can be used to evaluate the phantom center, but it's probably even better if you download a few different sound samples of a variety of sound objects to get an even better idea of the center image.

Anyway, here is my sound file of that snare drum: https://www.dropbox.com/s/1o5t4wifgafv33c/Snare Beats Center Left Right.wav?dl=0

That's very nice too. Yeah, there are many ways to have a more controlled test by knowing more precisely what to listen for. In my example, first portion of the video deals with very dry and quick enough transient containing almost full range of frequencies. Quick enough that it has precedence effect over the reflections so it should sound as dry as with headphones.

If your loudspeakers are too close to the boundaries and too wide apart there will be discreet phase differences and some cancellation, very audible as changes in timbre and loss of image focus. It also reveals room modes at lower frequencies, as well decay or possible ringing problems. In other words, what's still going on when the signal already stopped. It can also reveal the optimal toe in angle with regards to loudspeaker directivity. In my experience, when the setup is done right this way, this ensures that ITD's and ILD's in recordings will be presented as correctly as possible and that everything summed to mono will have good image focus and stability.

Your example of snare drum also contains some reverb, so that can be used to assess some of the late reflections too.

 

[Axo1989]

I think there could be miscommunication here. My understanding of 3D and envelopment is different. For me 3D is to mean more realistic lifelike experience of imaging of the sound in space from stereo recordings. Once done correctly with crosstalk cancellation the listener hear the difference like in this example when played over loudspeakers.

And envelopment is something that stereo cannot reproduce in full as it was not captured in the recordings except the limited frontal reverbs. I use actual concert hall or space impulse response to create the envelopment via 26 speakers from 45 degrees onwards towards the back. It creates the artificial envelopment from anywhere from 0.3 to 4.4 second RT with different character like a concert hall or hard surface room like here

Ok I finally watched the second video. Really excellent.

As you would expect, with headphones I get envelopment, with loudspeakers at my usual 2.5 metre LP it's mostly at the front. Moving forward to approx 1.5 metres there's noticeably more envelopment (especially if I close my eyes and imagine headphones). However, despite the homilies from our gracious host and our luminary Toole et al, I most like the anechoic chamber rendition.

Note also that my room is fairly dry (around 200 ms > 125 Hz using the RT60 approximation) and currently lacks anything much in the way of direct rear or any rear side walls (also, fairly breezy) so probably an extreme case. If sound was coming from the back, I'd worry.

Going to watch their 'how we made it' video next.

 

[STC]

The important thing here is that you know that this mono sound sample is the same no matter where it's panned in the stereo field, and should therefore have the same width no matter where it is located in the mix.

Have you tried with pink noise? Somehow I feel there will always be tonal changes. At least, to my ears.

Ok I finally watched the second video. Really excellent.

As you would expect, with headphones I get envelopment, with loudspeakers at my usual 2.5 metre LP it's mostly at the front. Moving forward to approx 1.5 metres there's noticeably more envelopment (especially if I close my eyes and imagine headphones). However, despite the homilies from our gracious host and our luminary Toole et al, I most like the anechoic chamber rendition.

Note also that my room is fairly dry (around 200 ms > 125 Hz using the RT60 approximation) and currently lacks anything much in the way of direct rear or any rear side walls (also, fairly breezy) so probably an extreme case. If sound was coming from the back, I'd worry.

Going to watch their 'how we made it' video next.

There’s is no correct answer. It depends on the genre. See PM I have sent you as this is not relevant to OP.

 

[goat76]

Have you tried with pink noise? Somehow I feel there will always be tonal changes. At least, to my ears.

The answer to the tonal changes is in the bolded part below that you cut out from my reply.

It's possible to reduce that tonal change caused by crosstalk cancellation with EQ compensations for the phantom-centered sound, at least for the reproduction via two loudspeakers in a stereo configuration. For headphone listening such compensation will lead to an over-compensation.

In the DAW it's easy to listen to how those samples sound panned hard to one of the loudspeakers, and then compare how that exact same sound sample sounds when panned to the phantom center. The phantom-centered sound should sound almost as distinct as the hard-panned sound and not apparently wider or washed-out, it will probably sound a little less focused because of the stereo fault and the cancellations that occur by the crosstalk from the two loudspeakers, but try to concentrate on the width and the pinpoint localization of the sound sample. The important thing here is that you can be sure that this mono sound sample is exactly the same no matter where it's panned in the stereo field, and should therefore have the same width no matter where it is located in the mix.

 

[STC]

The answer to the tonal changes is in the bolded part below that you cut out from my reply.

It's possible to reduce that tonal change caused by crosstalk cancellation with EQ compensations for the phantom-centered sound, at least for the reproduction via two loudspeakers in a stereo configuration. For headphone listening such compensation will be an over-compensation.

I am still blur why you say there is cancellation of the crosstalk. How is the cancellation implemented here? Unless I overlooked that you are actually referring to crosstalk cancelled playback . ( Note - perhaps crosstalk minimized because I doubt it could be cancelled completely during playback via loudspeakers unless the speakers are beam forming types).

 

[goat76]

I am still blur why you say there is cancellation of the crosstalk. How is the cancellation implemented here? Unless I overlooked that you are actually referring to crosstalk cancelled playback . ( Note - perhaps crosstalk minimized because I doubt it could be cancelled completely during playback via loudspeakers unless the speakers are beam forming types).

Okay, I see now what may be the reason for the misunderstanding. Sorry.

What I was trying to say is that the tonality change of the phantom-centered snare drum is caused by the crosstalk that occurs for the phantom-centered snare drum when the playback system is two loudspeakers in a stereo configuration. This causes cancellations for some frequencies that affect the tonality of the sound.



Anyway, the change in tonality is expected for the phantom image and should be ignored by the listeners who use the track for optimizing their loudspeaker positions for a better phantom center. The listener should instead concentrate more on getting a distinct-sounding phantom image with an accurately perceived width to the sound source, and the snare drum hard-panned to either the left or the right speaker can be used as a reference point of how wide that sound source should sound like.

As I see it, with a fully optimized phantom image all pieces of the stereo image "puzzle" will fall into the right places, and it may be even more important for the finer details like the soundstage and imaging than what it is for the main sound sources in the mix. It's almost like adjusting a pair of binoculars to get a unified image.

 

[Tangband]

I think people should try your suggestion and set up their speakers with a way smaller listening triangle than what they probably are used to, it makes a significantly different when it comes to hearing the image and depth of what's actually in the recording when the ratio of the direct sound is optimized vs the reflections from the listening room.
I have come to the same conclusion as you, not by doing it exactly like your test but simply by shrinking my listening triangle to the point where I hear more "into" the recordings than listening to my own room.

Finding the right distance between the loudspeakers is also highly critical, too large a distance and you end up with a washed-out sounding phantom image. When you find an optimal distance between the two speakers, the phantom center can sound almost as solid as if a physical center speaker is used in the setup. And with that, you will know that you have "anchored" the sound from the two loudspeakers in the stereo setup and that they are reproducing a unified stereo image, and by that, pretty much everything including the finer details that paints the three-dimensionality (if the recording contains such information) will also be optimized and have better "focus".
When the "right" distance has been found for optimizing both the direct sound ratio and a distinct-sounding phantom image is found, toe-in can be used to fine-tune it even further.


When looking at pictures of people's HiFi systems, many people seem to have a somewhat large listening triangle where the ratio of the direct sound is probably pretty low vs their room reflections. It's as if they let the room size dictate how far away they sit from their speakers, almost as if it's a rule that a fairly large room is the deciding factor of how large the listening triangle ends up being, and that the listening position must be right against the wall on the opposite side of the room.

Agree on the importance of correct setup. I would go even further and say that when you find the optimal listening spot and correct loudspeakerplacement the soundstage will be better using two good speakers than also using a center speaker
( In the sweet spot )

 

[Suono]

The phase, the phase is the most important thing to get a solid and focused image. Any phase distortion, recording, diffusers(polar diagram) room reflections, can negatively affect.

 

[Deleted member 21219]

The phase, the phase is the most important thing to get a solid and focused image. Any phase distortion, recording, diffusers(polar diagram) room reflections, can negatively affect.

What about the humongous amounts of phase distortion that musicians introduce in the studio? What about the diffusers that are used in many recording venues? What about the room reflection that are inherent in many recordings, such as concert halls?

Jim

 

[Emlin]

What about the humongous amounts of phase distortion that musicians introduce in the studio? What about the diffusers that are used in many recording venues? What about the room reflection that are inherent in many recordings, such as concert halls?

Jim

It’s not really about phase, but the accurate reproduction of the reverb on the recording.

 

[Deleted member 21219]

It’s not really about phase, but the accurate reproduction of the reverb on the recording.

I see. And will that be discernible in a blind A/B comparison?

Jim

 

[Emlin]

I see. And will that be discernible in a blind A/B comparison?

Jim

Yes, because reverb is your brain's first go to when it tries to work out the size of the cave you are in and where the sounds of any predators within it originate. In the dark.

 

[Tangband]

The phase, the phase is the most important thing to get a solid and focused image. Any phase distortion, recording, diffusers(polar diagram) room reflections, can negatively affect.

If that was true we all should listen to fullrange drivers without any crossovers

 

[Suono]

Che dire dell'enorme quantità di distorsione di fase che i musicisti introducono in studio? E i diffusori utilizzati in molti luoghi di registrazione? Che dire del riflesso della stanza che è insito in molte registrazioni, come le sale da concerto?

Jim

Preserving stereo or iin general stereo or microwave shooting, I meant from this moment on

 

[Suono]

Se ciò fosse vero, dovremmo tutti ascoltare driver fullrange senza crossover

È essenziale che i conducenti siano coerenti nelle emissioni