I recently built a new set of speakers (stereo, not surround) and I've noticed that on some recordings the soundstage is noticeably wider than the speaker placement. What I mean is while listening with my eyes closed if I point to where an instrument appears to be that point may be substantially farther to the left or right than the speakers.
Also there is more depth than before, though this is easier for me to understand in terms of creating that on a recording.
Some recordings have a wider/deeper stage than others so obviously the way the music is recorded controls this to a great extent.
So can someone explain how the soundstage can appear wider than speaker placement? Or maybe point me to a website that explains it in layman's terms?
Also there is more depth than before, though this is easier for me to understand in terms of creating that on a recording.
Some recordings have a wider/deeper stage than others so obviously the way the music is recorded controls this to a great extent.
So can someone explain how the soundstage can appear wider than speaker placement? Or maybe point me to a website that explains it in layman's terms?
It's a good question, and no-one knows all the answers.
Panning across the soundstage can be accomplished in one of two ways, panning by level and panning by phase error.
Most studio recordings are panned by level (aka Pan-pot-mono) which will normally confine the soundstage to inside the speakers separation.
Natural sound has a phase error (a delay time) between your left and right ear, as well as level differences, which means that sounds can come from beyond the width of the speakers, and if the phase error approaches 180', sounds appear to come from all around at once.
Depth is usually perceived by the spectral imbalance of the direct path sound. Height appears to be determined by the bass content of the first reflected path sound, but AFAIK no-one has conclusively determined this or any other mechanism for judging height in a recording.
Panning across the soundstage can be accomplished in one of two ways, panning by level and panning by phase error.
Most studio recordings are panned by level (aka Pan-pot-mono) which will normally confine the soundstage to inside the speakers separation.
Natural sound has a phase error (a delay time) between your left and right ear, as well as level differences, which means that sounds can come from beyond the width of the speakers, and if the phase error approaches 180', sounds appear to come from all around at once.
Depth is usually perceived by the spectral imbalance of the direct path sound. Height appears to be determined by the bass content of the first reflected path sound, but AFAIK no-one has conclusively determined this or any other mechanism for judging height in a recording.
The sound engineer who does the recording at the studio, makes recording of each instrument and vocals initially at a seperate channel. Then he combines in a special way those channels, to produce the final 2 channel (stereo) recording.
Lets say that the singer is located exactly in the middle. His/her voice will be recorded to the final stereo recording with the same level (left and right speaker will produce exactly the same spl)
and with no phase delay (at the same time)
Lets say the sound engineer decides to (audibly) put (for example) the violin in a specific spot of the soundstage, left of the singer, and 1 meter deeper (1meter biggest distance from the listener)
He will transfer the initial violin recording, to the final stereo recording by adjusting the balance potentiometer of the violin channel more to the left (making our left speaker playing the violin more loud than the right one,
while the singer will be played at the same level left and right). He will also adjust the left delay potentiometer in a value that corresponds in milliseconds to the 1 meter of longer distance of the volin, compared to the singer (so the left speaker will reproduce the violin some milliseconds later than the right speaker wich is in phase with the centered vocals)
By this way, the sound engineer places each instrument at the desired locations.
This is the simplified technique, but there are more parameters that the sound engineer takes advantage of. Room reflections, echos, can be entered to give a larger room perspective. So in the recording among with the instruments and vocals info, there is room info that can be entered to give the impression of the supposed room that the band is playing.
This is done, while the sound engineer has in mind that we will put our two speakers in a proper position, in front of us (like a band always is) and in a symmetric position. Thats why the speaker placement is critical. The listener's distance from the left and right speaker should be equal and the cables' polarities should be correct (so the time delay trick works well) and the way we put the speakers in our room, decides the way the soundstage will appear to us (keeping the speakers 1 meter from the walls will give a larger and deeper soundstage - due to the wall's reflections, than keeping the speakers 20cm from the walls). There are some basic rules on speaker positioning, but on each room the best result can only be achieved with many hours of experimenting and listening.
Lets say that the singer is located exactly in the middle. His/her voice will be recorded to the final stereo recording with the same level (left and right speaker will produce exactly the same spl)
and with no phase delay (at the same time)
Lets say the sound engineer decides to (audibly) put (for example) the violin in a specific spot of the soundstage, left of the singer, and 1 meter deeper (1meter biggest distance from the listener)
He will transfer the initial violin recording, to the final stereo recording by adjusting the balance potentiometer of the violin channel more to the left (making our left speaker playing the violin more loud than the right one,
while the singer will be played at the same level left and right). He will also adjust the left delay potentiometer in a value that corresponds in milliseconds to the 1 meter of longer distance of the volin, compared to the singer (so the left speaker will reproduce the violin some milliseconds later than the right speaker wich is in phase with the centered vocals)
By this way, the sound engineer places each instrument at the desired locations.
This is the simplified technique, but there are more parameters that the sound engineer takes advantage of. Room reflections, echos, can be entered to give a larger room perspective. So in the recording among with the instruments and vocals info, there is room info that can be entered to give the impression of the supposed room that the band is playing.
This is done, while the sound engineer has in mind that we will put our two speakers in a proper position, in front of us (like a band always is) and in a symmetric position. Thats why the speaker placement is critical. The listener's distance from the left and right speaker should be equal and the cables' polarities should be correct (so the time delay trick works well) and the way we put the speakers in our room, decides the way the soundstage will appear to us (keeping the speakers 1 meter from the walls will give a larger and deeper soundstage - due to the wall's reflections, than keeping the speakers 20cm from the walls). There are some basic rules on speaker positioning, but on each room the best result can only be achieved with many hours of experimenting and listening.
I have also noticed that the songs where I get the most lateral spread are from tracks with odd sound effects or weird phaseyness added to synthesisers. Normal recordings dont usually extend pass the boudaries of the loudspeakers but some, these electronic ones do.
Unfortunately its probably only a select number of sound engineers who will actually engineer in phase difference between the two channels to create a more realistic sound stage. Also normally when stereo mixes are produced most of the time I think they use total pan, ie sound either from both speakers or the left or the right. Obviously there are some inbetweens, but just panning is not suffucient to create realistic sound stages.
I always doubt hifi reviewers who say their speakers create a soundstage that extends several feet backwards and sidewards from the speakers. David Allcock seems to be one who regularly states this, the soundstage extends into my garden and into the next rooms! This is fabulous for dave if this is how he percieves the sounds, as out brains also have as much to do with creating a sound stage as the sound system does. But I bet if I were to sit exactly where he is in the room with exactly the same material I wouldnt hear what he hears.
I find it very difficult to hear stage depth. This is not music depth. I cane easily hear from one piece of music to the next a difference in stage depth. Ie one piece plays behind the speakers plane, whereas another plays infront. I just find it hard to hear depth between individual intruments, within the stage.
I know what I should do for my career, instead of trying to be a loudspeaker designer, be a recording engineer with a difference I use PHASE MWHAHAHAHAH eat that sony!
Unfortunately its probably only a select number of sound engineers who will actually engineer in phase difference between the two channels to create a more realistic sound stage. Also normally when stereo mixes are produced most of the time I think they use total pan, ie sound either from both speakers or the left or the right. Obviously there are some inbetweens, but just panning is not suffucient to create realistic sound stages.
I always doubt hifi reviewers who say their speakers create a soundstage that extends several feet backwards and sidewards from the speakers. David Allcock seems to be one who regularly states this, the soundstage extends into my garden and into the next rooms! This is fabulous for dave if this is how he percieves the sounds, as out brains also have as much to do with creating a sound stage as the sound system does. But I bet if I were to sit exactly where he is in the room with exactly the same material I wouldnt hear what he hears.
I find it very difficult to hear stage depth. This is not music depth. I cane easily hear from one piece of music to the next a difference in stage depth. Ie one piece plays behind the speakers plane, whereas another plays infront. I just find it hard to hear depth between individual intruments, within the stage.
I know what I should do for my career, instead of trying to be a loudspeaker designer, be a recording engineer with a difference I use PHASE MWHAHAHAHAH eat that sony!
Well, I can't claim the soundstage goes into the next room! It does however extend about 18 inches wider than the speakers. Again it isn't that wide on all recordings but on a couple of recordings it is very noticeable. Neither has weird synth stuff but neither is a live recording either.
As for depth, it doesn't seem to extend behind the speakers but comes out in front. It isn't huge but it isn't that flat wall of sound I'm used to. I actually made recordings with depth while in college and I made recordings where I placed instruments where I wanted left to right but I was never able to extend them beyond the speakers, but then again I never thought of trying!
Sherman
As for depth, it doesn't seem to extend behind the speakers but comes out in front. It isn't huge but it isn't that flat wall of sound I'm used to. I actually made recordings with depth while in college and I made recordings where I placed instruments where I wanted left to right but I was never able to extend them beyond the speakers, but then again I never thought of trying!
Sherman
Just a few points based on my experiences with this issue.
Firstly, I believe that phase problems introduced by crossover units interfere with an accurate soundstage and accurate imaging. Vibrations from the speaker enclosure are also a dead giveaway.
However, reflections from walls, floor and ceiling can also completely distort the soundstage picture. These can be minimized with placement and toe-in but they are related to the dispersion characteristics of the drivers used. Another negative for crossovers is that the dispersion characteristics change as you move from woofer to tweeter and this messes up the soundstage.
The point about the sound seeming to come from beyond the speakers is a different, though interesting, issue. Headphones always give a soundstage that goes way beyond the 'speakers'. I'd bet that if you position your speakers so that they're 18" apart, toed in and 6" in front of your head, your soundstage would be way beyond the speakers too.
Firstly, I believe that phase problems introduced by crossover units interfere with an accurate soundstage and accurate imaging. Vibrations from the speaker enclosure are also a dead giveaway.
However, reflections from walls, floor and ceiling can also completely distort the soundstage picture. These can be minimized with placement and toe-in but they are related to the dispersion characteristics of the drivers used. Another negative for crossovers is that the dispersion characteristics change as you move from woofer to tweeter and this messes up the soundstage.
The point about the sound seeming to come from beyond the speakers is a different, though interesting, issue. Headphones always give a soundstage that goes way beyond the 'speakers'. I'd bet that if you position your speakers so that they're 18" apart, toed in and 6" in front of your head, your soundstage would be way beyond the speakers too.
I've tried that, not with particularly great speakers, but they imaged consistently enough. The headphone effect has a lot to do with almost no sound from the right channel reaching the left ear and vice versa.7V said:
You also loose most of the pinna effects, the head transfer function(s - being complex it is easier to calculate as several paths rather than one composite path) and the room interaction.
Not that I would forsake speakers for headphones unless forced to.😀
Good point. I guess you'd have to increase the distance the speakers were apart and in front for the experiment to be valid.Mudge said:
My thought was that if you can't 'hear' the speaker cabinets (vibrations or resonances), why should the sound not seem to come from beyond the speakers?
Steve,
If there's no intrinsic phase difference between signal L and signal R, then the soundstage through speakers will depend on the differing arrival times at each ear of each signal and the difference in level. In effect the brain can only calculate soundstage width from the sound being entirely in one speaker or entirely in the other, any condition in between is possible, but nothing beyond that.
When you switch to headphones, there's no difference in arrival time, because the sound doesn't reach your other ear in any comparable signal strength. That part of the auditory complex switches off and level and spectral balance play an increased part in determining soundstage width. Hence why I prefer speakers to headphones, music really needs to be mixed with headphones in mind to sound right through them.
EDIT: I should add that the 'depth' of a sound in one ear only is the determining factor for soundstage width in headphones.
If there's no intrinsic phase difference between signal L and signal R, then the soundstage through speakers will depend on the differing arrival times at each ear of each signal and the difference in level. In effect the brain can only calculate soundstage width from the sound being entirely in one speaker or entirely in the other, any condition in between is possible, but nothing beyond that.
When you switch to headphones, there's no difference in arrival time, because the sound doesn't reach your other ear in any comparable signal strength. That part of the auditory complex switches off and level and spectral balance play an increased part in determining soundstage width. Hence why I prefer speakers to headphones, music really needs to be mixed with headphones in mind to sound right through them.
EDIT: I should add that the 'depth' of a sound in one ear only is the determining factor for soundstage width in headphones.
A week or so ago I was listening to a CD of The Shadows, the track was "This 'ol House" IIRC. Right at the very beginning before the music starts you can hear distant voice going "a-one, a-two, a-one-two-three" to get the band starting in time. From where I sit the speakers are at 10:45 and 1:15 o'clock position. On my system this voice appears right over to the left at about the 9:30 position! I played it again and again to just make sure I wasn't imagining it.
I moved my speakers, placing them on either side of my entertainment center which is in a corner of the room. The fronts of the speakers are parallel but the wall behind each is at an angle. They are closer together than they were before.
I listened to several recordings again last night and the "wider than speaker placement" soundstage is still there in this new position. The effect is a little different due to the closer speaker placement I guess. A guitar which seemed to come from 18" or so to the left of the left speaker now seems to be a little closer to the speaker, maybe 8 inches or a foot to the left.
I had my kids and wife listen to the recordings where the effect was most obvious to me and asked them to close their eyes and point to the instruments. All three of them identified the same effect.
After listening more and thinking about it a bit I can only come up with this- we hear "width" in a recording by both comparative loudness from each speaker and the timing of the sound reaching the ears.
Also the speakers themselves seem to have something to do with it since I didn't notice this on the same recordings with my previous speakers. I have two other sets of speakers I think I'll try just to see what the effect is with those.
I listened to several recordings again last night and the "wider than speaker placement" soundstage is still there in this new position. The effect is a little different due to the closer speaker placement I guess. A guitar which seemed to come from 18" or so to the left of the left speaker now seems to be a little closer to the speaker, maybe 8 inches or a foot to the left.
I had my kids and wife listen to the recordings where the effect was most obvious to me and asked them to close their eyes and point to the instruments. All three of them identified the same effect.
After listening more and thinking about it a bit I can only come up with this- we hear "width" in a recording by both comparative loudness from each speaker and the timing of the sound reaching the ears.
Also the speakers themselves seem to have something to do with it since I didn't notice this on the same recordings with my previous speakers. I have two other sets of speakers I think I'll try just to see what the effect is with those.
Yesterday I accidently inverted the connectors at one of my speakers. Boy did I get some pretty sound effects beyound the speaker base! 😀
So phase issues seem to be dominant here.
The recording featuring the most prominent space effects in my CD collection is "Andromeda Heights" by Prefab Sprout.
From my listening position the speakers are 60° apart, but on some tracks of this CD I can easily locate voices, instruments etc. in a 120° angle. I have experienced this effect with three very different speaker sets, so I don´t think it depends much on the speakers.
By the way: The recording only sounds "phasey" while the sound sources are moved through space. At a stable position (even far beyond the speaker base) there is not much difference compared to the same instrument or voice on axis.
Apart from such special recordings an exceptionally wide soundstage IMHO could indicate phase problems with the crossovers.
Rudolf
So phase issues seem to be dominant here.
The recording featuring the most prominent space effects in my CD collection is "Andromeda Heights" by Prefab Sprout.
From my listening position the speakers are 60° apart, but on some tracks of this CD I can easily locate voices, instruments etc. in a 120° angle. I have experienced this effect with three very different speaker sets, so I don´t think it depends much on the speakers.
By the way: The recording only sounds "phasey" while the sound sources are moved through space. At a stable position (even far beyond the speaker base) there is not much difference compared to the same instrument or voice on axis.
Apart from such special recordings an exceptionally wide soundstage IMHO could indicate phase problems with the crossovers.
Rudolf
It really depends on the speakers used, but with floorstanders and also with smaller two-ways you must not cram them into the walls or near furniture and so on. They need room to "breathe"!
My dipole floorstanders are placed in a 20ft/12ft/9ft room 3ft from the back wall and 4ft from the side walls to achieve the best 'soundstage' and I managed to get to this position in about 8 months of moving speakers around.
My dipole floorstanders are placed in a 20ft/12ft/9ft room 3ft from the back wall and 4ft from the side walls to achieve the best 'soundstage' and I managed to get to this position in about 8 months of moving speakers around.
Rudolf said:
After reading your post I double checked my cabling (and checked the cable continuity with a meter in just in case). All is well there. Also the speakers I'm now using do not have a "crossover" per se but just a high-pass filter cap for the tweeters so its only a 6db per octave roll off and shouldn't have any phase issues. Just in case I reversed the phase of the tweeters and all I got was slightly muddier upper mids and highs.
🙁
After listening more I do think you are right that this effect is phase related. But I think in this case it is in the way the recordings were made rather than a speaker issue. I tried a couple of different sets of speakers and with three sets the effect was there and with one it wasn't. I've also moved the speakers around a bit and placement does affect it but doesn't remove it unless I'm very close to the speakers.
I guess the final word is that some recordings have this effect and some don't and some speakers reproduce the effect better than others.
Speakers with very good phase response are a complete pain if you're the owner of some recordings which were 'stereo-ised' from a mono master. Through ordinary consumer grade crap they sound okay, but use something decent and you wind up listening to a bizarre tug of war between level (identical) and phase (constant difference). Words can't describe how awful these abominations can sound.🙄
This is my feeling as well.
I was once appalled by listening to a specific CD for the first time at night. It was a cover version of "Spooky" (what a coincidence !) by Christy Baron. On this track there is some percussion instrument sounding like laughter. This one sounded like coming from the window (i.e. 90 degrees to the righ !!!), giving an instant impression of someone lurking around outside !
Regards
Charles
Soundstage is created by the brain.
If you want a neat demo of the sorts of phase tricks which engineers can use to image outside of speaker positions, listen to the Q-sound encoded "Amused to Death" by Roger Waters. With point-source speakers (like minimonitors), some of the sound effects placements can be startling.
If you want a neat demo of the sorts of phase tricks which engineers can use to image outside of speaker positions, listen to the Q-sound encoded "Amused to Death" by Roger Waters. With point-source speakers (like minimonitors), some of the sound effects placements can be startling.
Someone once called that one "cinema for the blind". Unfortunately I haven't been able to find that one so far.
Regards
Charles
SY said:
When I was in college the guy in the room across the hall had a real quadraphonic system. Four amplifiers, four equal speakers, a four track reel-to-reel tape deck and more. It was a pretty cool system for the day.
He once made a recording using his portable reel-to-reel (stereo) which he then mixed into real quad. He recorded an oncoming freight train placing one mike on a stand about 8 feet off the ground and the other mike was about four feet off the ground.
He paid a guy at a junk yard to pick up a couple of tons of scrap metal with one of those big claw machines and drop it. He recorded that in stereo as well. He added other bits of recordings he made and mixed it all together. The final effect was like standing on a railroad track with a train coming fast. As the train approached there was a voice yelling something like "get off the track" and finally a huge crash (the metal dropping).
He would get someone to stand in the middle of his room with a blind fold on and play back the recording. I tried it and saw at least a dozen others do it as well. Not one of us could stand still at the end, even after doing it a couple times!
The thing was that he had the speakers suspended in nets in the corners of the room so the soundstage couldn't be "wider" than the speakers but it certainly had huge depth! That train really seemed to be coming from a mile away and getting closer and closer.
- Status
- Not open for further replies.