Now things have become too detailed for me to explain them in words. 
A picture might be better. This is where I "see" the phantom sources:
"Dry" is fixing the sound to the speaker. "Wet" moves the phantoms away from the speaker, but also increasingly to the center. Distances to the front wall or behind the front wall are only descriptive - "more wet" is farther behind the wall than shown, but I can't give an exact distance.
Is this roughly in line with other's findings?
I don't experience any remarkable changes in the heights.
Elias,
with SSSx5 - do you hear any comparable change in the positions of the phantom sources between "at the side walls" and "at the center"?
A picture might be better. This is where I "see" the phantom sources:"Dry" is fixing the sound to the speaker. "Wet" moves the phantoms away from the speaker, but also increasingly to the center. Distances to the front wall or behind the front wall are only descriptive - "more wet" is farther behind the wall than shown, but I can't give an exact distance.
Is this roughly in line with other's findings?
I don't experience any remarkable changes in the heights.
Elias,
with SSSx5 - do you hear any comparable change in the positions of the phantom sources between "at the side walls" and "at the center"?
Attachments
I had not been following this thread, interesting test - in a mixdown sort of way. I was going to draw a chart but find that Rudolph's is very close to what I hear. The wet and more wet brings the voice 1/2 way between center and right speaker. Piano less distinct, but same move toward center. "Dry less spread" sounds like a good home recording, very close mic'd.
Multitap is like a home recoding with the mics farther back. Not unpleasant. Sounds stick pretty close to the speakers.
On headphones it's easy to hear when the vocal track comes in and out, it's turned on and off. Not obvious over speakers. Also on the wet tracks can clearly hear voice reverb without voice on the left thru headphones, it is better blended on speakers, more whole.
The piano seems to have some ambient effect built into it, it's not 100% dry. Don't know if that's the reason, but the reverb on the piano is not as obvious as on voice.
Thanks for the fun test.
Multitap is like a home recoding with the mics farther back. Not unpleasant. Sounds stick pretty close to the speakers.
On headphones it's easy to hear when the vocal track comes in and out, it's turned on and off. Not obvious over speakers. Also on the wet tracks can clearly hear voice reverb without voice on the left thru headphones, it is better blended on speakers, more whole.
The piano seems to have some ambient effect built into it, it's not 100% dry. Don't know if that's the reason, but the reverb on the piano is not as obvious as on voice.
Thanks for the fun test.
Yes, that is strange. They are either using a noise gate or just chopped out the gaps between phrases.
This was fun. I think it showed what we expected, that very dry recordings can image at the speakers and added reverb (and mixing to center) can get the image away from the speakers. Still, some had a distinctly different experience.
David S.
Late for the party
I did theose wet/dry tests last night with my 3-ch DML mockups (stereo to 3-ch conversion by passive linear matrix).
Track after track, except for the last one (dry less spread), the images I've heard were all at the same positions -- at about 25% & 75% points laterally, 1m or so behind the surface of speakers for the singer, and 1 more meter back for that (toy) piano.
In the 'dry less spread', 2 images are much closer laterally, also a step closer to me. The singer is almost at the center while the piano is slightly to the left. In addition, the (toy) piano in this track came with a better sense of size and weight behind the singer.
Well, those are not precise descriptions. In the "mix of wet/dry", with the reverberations come and go alternatively, the 2 images seemed to move back and forth a little. Just a little, not even a small step.
Images and senses of spaces are all clearly detached from all 3 large panel-type speakers and walls of room. So I guess it's not too shabby in this disappearing game.
The major differences between wet and dry seem only in the amount of reverberations. With the wet and more wet, the reverbs more or less render some senses of spaces, but not real or natural enough to make believable illusion, like CGs, kind of. A CG must be very very careful to every tiny bit of details (or lack of them naturally) and also very very careful in the overall balance to be mistaken for a real photo.
The senses of spaces in the wet mix is very close to real by a small gap, like a very good CG. While the more wet is indeed too exaggerated to be true. Maybe the effect could be true in a huge cave I've never been, but OTOH a soprano and a piano in that kind of location is unreal practically.
If they were not in such tests, all of them are actually good enough to listen to, even the not-so-natural multitap is not any worse than those regular pop recordings.
By comparison, I prefer the dry and less dry in my system (and very live room). Maybe it's because the sense of space is similar to my real space (but double in the depth), so the images seem more real as I'm in the space with them.
The last (dry less spread) is my favorite. Maybe because it's what I've been used to in such music.
I did theose wet/dry tests last night with my 3-ch DML mockups (stereo to 3-ch conversion by passive linear matrix).
Track after track, except for the last one (dry less spread), the images I've heard were all at the same positions -- at about 25% & 75% points laterally, 1m or so behind the surface of speakers for the singer, and 1 more meter back for that (toy) piano.
In the 'dry less spread', 2 images are much closer laterally, also a step closer to me. The singer is almost at the center while the piano is slightly to the left. In addition, the (toy) piano in this track came with a better sense of size and weight behind the singer.
Well, those are not precise descriptions. In the "mix of wet/dry", with the reverberations come and go alternatively, the 2 images seemed to move back and forth a little. Just a little, not even a small step.
Images and senses of spaces are all clearly detached from all 3 large panel-type speakers and walls of room. So I guess it's not too shabby in this disappearing game.
The major differences between wet and dry seem only in the amount of reverberations. With the wet and more wet, the reverbs more or less render some senses of spaces, but not real or natural enough to make believable illusion, like CGs, kind of. A CG must be very very careful to every tiny bit of details (or lack of them naturally) and also very very careful in the overall balance to be mistaken for a real photo.
The senses of spaces in the wet mix is very close to real by a small gap, like a very good CG. While the more wet is indeed too exaggerated to be true. Maybe the effect could be true in a huge cave I've never been, but OTOH a soprano and a piano in that kind of location is unreal practically.
If they were not in such tests, all of them are actually good enough to listen to, even the not-so-natural multitap is not any worse than those regular pop recordings.
By comparison, I prefer the dry and less dry in my system (and very live room). Maybe it's because the sense of space is similar to my real space (but double in the depth), so the images seem more real as I'm in the space with them.
The last (dry less spread) is my favorite. Maybe because it's what I've been used to in such music.
CLS, I'd guess you do have a very live space of you liked the dry files. In my room the centered one isn't bad, but it's still very dry and somewhat unnatural. It ought to sound great in a big, live space.
There is an old Chesky Demo CD where they do some real and electronic reverb. It's interesting to hear the difference. FWIW, I've always found the Lexicon stuff to sound the most natural.
There is an old Chesky Demo CD where they do some real and electronic reverb. It's interesting to hear the difference. FWIW, I've always found the Lexicon stuff to sound the most natural.
Yes, my space is very live. (And the DML panel-type speakers also tend to sound more live then others.)
The dry ones do sound dry on the 2 images, but they are still slightly behind the speakers and there's also a sense of space behind them. Put it visualized, it seems the images are more focused and sharp, standing in the front part of the space near me. So I 'see' them more clearly, thus more real. The audiophile term is 'touchable'
OTOH, the images of the wet ones are somewhat surrounded by a fuzzy halo, so not as sharp and clear as dry ones.
And there's a proportion aspect in these things: how clear the images should be in their presented locations related to our other experiences. As mentioned I heard (saw) them more or less the same distance from me. In such (sense of) distance, seeing them clearly is more real to me.
Also, I noticed the recording level is quite low. Don't know if it's relevant. In my impressions, recordings of lower average level tend to sound more laid-back. Maybe it's why the dry ones didn't put too much pressure on me.
The dry ones do sound dry on the 2 images, but they are still slightly behind the speakers and there's also a sense of space behind them. Put it visualized, it seems the images are more focused and sharp, standing in the front part of the space near me. So I 'see' them more clearly, thus more real. The audiophile term is 'touchable'
OTOH, the images of the wet ones are somewhat surrounded by a fuzzy halo, so not as sharp and clear as dry ones.
And there's a proportion aspect in these things: how clear the images should be in their presented locations related to our other experiences. As mentioned I heard (saw) them more or less the same distance from me. In such (sense of) distance, seeing them clearly is more real to me.
Also, I noticed the recording level is quite low. Don't know if it's relevant. In my impressions, recordings of lower average level tend to sound more laid-back. Maybe it's why the dry ones didn't put too much pressure on me.
Last edited:
While he does have a live space, I don't think that is why the sound is "detached".
Using a 3 channel system, or something like Elias's single speaker, center is more "centered" in the range of 2kHz up (i.e. it "pulls" the L and R channels away from L & R). When you do this the monophonic emphasis increases and depth enhances.
This is a phenomenon related to "head-shading". Above 2kHz direct sound cross-correlation *decreases* (or L & R have a lower correlation), because the head "blocks" the sound from the right speaker traveling to the left ear (and vice versus).
An additional component to this is the "angle" of the "stereo triangle". Basically the further apart you move speakers L & R (without a center speaker), the more head-shading happens, and the lower the cross-correlation. It also happens more a bit lower in freq.. (still, unless the speakers are very close together you get this decreased cross-correlation.)
It doesn't get really separated for most people until around 5 kHz with a typical stereo separation, and at 7-7.5kHz things go "omni" again briefly with higher correlation. (..this is due to an inner-ear effect.)
Unfortunately we tend to "latch" onto these higher freq.s for positioning information - so even if a harmonic at a lower amplitude is extending to 5 kHz then we have that decreased cross-correlation. (i.e. the lower level harmonic "pulls" the position of the image to the left or right and results in a lower monophonic emphasis.)
Note: If you look at Rudolf and Helmuth's pic's - Helmuth has greater separation of L & R relative to Helmuth (..than Rudolf's L & R relative to Rudolf). Because of the lower cross-correlation for Helmuth, the image is slightly in front of the speakers, whereas Rudolf has the image directly at his speakers position. If Rudolf were to move L & R closer together, then he should have the "dry" recording's image move back from L & R marginally.
Last edited:
These extracts are a developer's tool. I had already some samples of this kind but never done in an anechoic room with such an easy message to analyse. They are also very useful in double mono. I would like to thanks Dave to have put them in the market.
Because this rises up the classical question : if this sounds too beautiful, isn't the compound room/system adding too much of it's own ?
I have been making some modifications, that render them from full center to full sides. The choice of the reason is in between.
Because this rises up the classical question : if this sounds too beautiful, isn't the compound room/system adding too much of it's own ?
I have been making some modifications, that render them from full center to full sides. The choice of the reason is in between.
Ha, I knew it
Yes we did find it funny I listened to the tracks on a cheap pair of logitech USB headphones tonight. What a difference! The dry track became totally planted in each ear, very nu-listenable. The piano sounded more like a mandolin (slight exageration) than a piano, and interestingly the wet track sounded a lot better than the dry. The difference between wet and dry on the headphones was much greater than on the speakers.
I also noticed what pano said about the multitap with a small component of the singers voice in my left ear.
Tony.
I'm a bit late to the party, I've been watching the thread but haven't had time to sit down and do my own listening test, so here are my results.
Mozart dry - the singer in the right channel initially appears to be about 1 foot to the right of the speaker if I face directly ahead, however if I turn my head around a bit while listening localisation then snaps to being in line with the middle of the speaker.
Somewhat puzzling, I repeated the test several times and each time had the same ambiguity of position, I suspect its due to a fairly strong sidewall reflection of the right sidewall which is only 0.7m from the centre of the speaker, "stretching" the apparent source location towards the wall, however head turning must allow me to more accurately pinpoint the location.
Depth I found ambiguous - probably at or just behind the speaker, but there were no strong depth cues, all I could tell for sure was direction.
The piano on the left also sounded about 1 foot to the left of the speaker, and depth was definitely behind the speaker maybe a foot or two, in fact it sounded like it was coming from right in the corner of the room, not from the front of the speaker.
Although both sounded more or less in line with each speaker I didn't find them to localize strongly to the front of the speakers, depth was either ambiguous or somewhat behind, not quite what I was expecting for a dry recording.
Mozart wet - sounds like the acoustics of a large hall. The voice shifted to about halfway between right speaker and centre azimuth wise, and beyond the front wall of the room depth wise. Likewise the piano moved to half way between the speaker and centre line for azimuth, and further away to beyond the front wall of the room, approximately the same distance as the voice. Neither speaker is localised at all on this recording - both sound sources hang in space quite convincingly some distance beyond the actual boundary of the room.
Mozart multi-tap - sounds like the acoustics of a small live room, both sound sources azimuth are more or less in line with the speakers but sound like they're coming from the corners of the room behind the speakers, not from the speakers themselves. In other words the sound doesn't appear to localize at the speakers, it just comes from the same direction.
Dry less spread - sounds more or less mono to me - pretty much a phantom mono channel with maybe a tiny bit of separation in the middle. The apparent source hovers in the air slightly above the speakers and slightly behind them, pretty much what I get on any phantom mono signals.
Morzart more wet - sounds much like mozart wet but a bit further away behind the front wall. Azimuth is still between centre and each speaker, but perhaps more towards the speaker than it was with less reverb.
Since many of us found that the wet files shifted the singer and piano in toward center, I have to ask: Dave, did you pan the channels any on these tracks? Or is it reverb only?
I have not looked at the tracks in an editor to see.
EDIT: OK, there is about 6dB of crosstalk in the wet file. Hard to tell if that's reverb only, or a mix of dry and wet. That would help shift the image toward the center.
I have not looked at the tracks in an editor to see.
EDIT: OK, there is about 6dB of crosstalk in the wet file. Hard to tell if that's reverb only, or a mix of dry and wet. That would help shift the image toward the center.
Yes, I used a "pan" feature in "wet" and "more wet". The goal was to create soundspaces that sounded most natural (at the time on headphones) so I didn't want extreme seperation, more of a sense of two performers in front of us on a stage. Cool Edit Pro has a pan scheme with -100 to 0 to 100 for left into left, left into right, right into right, etc. I think the settings were 100(%) left into left, 40(%) left into right (the blending part) and the opposite for the other channel. Note, the minus numbers would allow out of phase expanding, and, I don't know for sure if the numbers are real percentage or some sort of log taper.
The "dry" was no pan at all to maximize the in-your-ear or in-your-speaker effect, and the "less seperation" was, I think, 70% on the cross feed aspects.
Also, the reverb itself is proper stereo reverb, so it tends to mimic reflections from all sides, hence it diminishes seperation a little.
Finally, the multitap had an option of "add to same channel" or "add to opposite channel" and I chose the later. Otherwise it was an even odder dual mono effect.
Regards,
David
I have thin curtains in front of my speakers and with 'dry', I didn't notice that the performers coincided with the speakers until I pulled back the curtains and watched, FWIW.
The 'wet' sounded like a hall and it did narrow the separation however with curtains in place and no visual reference, it wasn't pulling anything off the speakers as such, but rather just narrowing the separation. My experience was similar to Rudolph.
This reminds me that I dislike simple reverberation effects. I also dislike the peaky quality of a natural hall of that nature, but that's another issue. With regards to reverb I can cite an example in 'Baker Street', where the vocals have been overdone with reverb and they now sound out of place and less distinct, with the effect that they 'seem' less present.
As Pano says, the piano has something going on and it doesn't join in as readily. It seems less distinct which would have had me reaching for the level at first glance, but I suspect Pano may be right about it not being clean. Could it possibly be a stereo unit with one speaker at each end?
I preferred the sound of 'dry' to 'wet' due to the tonal balance/colouration as well as lack of 'confusion' in the multitude of reflections. It is very dry which isn't great, but I don't seem to dislike it as much as some people would.
Multitap struck me as quite impressive at first, although I could see myself beginning to hate the effect if I listened for too long.
The 'wet' sounded like a hall and it did narrow the separation however with curtains in place and no visual reference, it wasn't pulling anything off the speakers as such, but rather just narrowing the separation. My experience was similar to Rudolph.
This reminds me that I dislike simple reverberation effects. I also dislike the peaky quality of a natural hall of that nature, but that's another issue. With regards to reverb I can cite an example in 'Baker Street', where the vocals have been overdone with reverb and they now sound out of place and less distinct, with the effect that they 'seem' less present.
As Pano says, the piano has something going on and it doesn't join in as readily. It seems less distinct which would have had me reaching for the level at first glance, but I suspect Pano may be right about it not being clean. Could it possibly be a stereo unit with one speaker at each end?
I preferred the sound of 'dry' to 'wet' due to the tonal balance/colouration as well as lack of 'confusion' in the multitude of reflections. It is very dry which isn't great, but I don't seem to dislike it as much as some people would.
Multitap struck me as quite impressive at first, although I could see myself beginning to hate the effect if I listened for too long.
This thread has helped me to settle on what I like (thanks Dave in particular for setting up the experiment). I definately prefer a "They are here" presentation than a "You are there" performance with most music. Classical music is different and I like the ambience of the original setting, but otherwise I want the performers here but they can keep their rooms, their studios and their artificial rooms behind.
@ 454Casull, I was listening to (I think) the end of 'Tunnel of Love' but can't remember and there was a piano there on the left that seemed to go from 1/4 left and 5 feet high behind the front wall at the bass end to about 4 feet high just behind the speaker at the treble end like as if the mic were at a standing height facing the pianist angled to the left, but I'm not sure how much I can make of what I heard. I also recall there's a particular Neil Young song where it's obvious that he is sitting with his guitar on his knee.
@ 454Casull, I was listening to (I think) the end of 'Tunnel of Love' but can't remember and there was a piano there on the left that seemed to go from 1/4 left and 5 feet high behind the front wall at the bass end to about 4 feet high just behind the speaker at the treble end like as if the mic were at a standing height facing the pianist angled to the left, but I'm not sure how much I can make of what I heard. I also recall there's a particular Neil Young song where it's obvious that he is sitting with his guitar on his knee.
Check your tweeters on axis response for a peak in the ~8Khz region if you consistently notice imaging well above the speaker.
See the attached image that shows what frequency ranges localize offsets in which axis - oben is german for "above", and the graph shows that a fairly narrow frequency region around 8Khz is responsible for elevation perception.
I had always noticed my ribbon tweeters imaged quite a bit higher than their physical location (sometimes nearly as much as you describe) and a careful measurement showed a 1/3rd octave 1dB bump around 8Khz was present - I EQ'ed it out and presto, the apparent image location has moved down quite a bit and is now only slightly higher than the tweeters. (Quality and tonal balance has improved significantly as well despite the fairly small error)
Note: even if your speakers are flat in this region the music itself may have some emphasis at 8Khz, and the cumulative effect of the two if the speakers also have a bump there could be excessive and cause an unusually high apparent image source location.
The graph was posted in another thread, I've lost track of which thread it was and who posted it, but thanks to whoever it was, as I've found it very accurate and enlightening.
(English translation: Green - in front, Red - behind, Blue - above)
Attachments
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.