Blue or red pill?

[fas42]

37 minutes ago, dalethorn said:

I've been making ad-hoc binaural recordings with the Sennheiser-Apogee Ambeo headset, and each recording sounds way different depending on numerous factors affecting the rooms I record in, or not.  It's clear to me that a quality recording setup that addresses ambient noise, reflective surfaces, ideal listening location, etc. is vastly more important than the 'HRTF' and related properties that headphone listeners read so much about.  I've been reviewing headphones for a few years, but haven't attempted "quality" recordings until a few weeks ago.  I think if the average headphone user could use the Ambeo headset for a few weeks, they'd appreciate their audiophile music recordings a lot more.

 

Recordings sound enormously different from each other - 'quality' recordings are boring, generally - for the same reason that a super smooth, straight highway is far less interesting that a winding country road, for an enthusiastic driver - they are both means for people to progress quickly between two points, but I know which I prefer ...

 

Which means the cues that provide depth information are not "standard" in their quality, and level. They are always there, but vary in their audibility - between every recording. The higher the "resolving power" of the playback rig, the more likely a particular recording will show obvious depth - ultimately, it becomes impossible to find a recording that doesn't present a 3D picture in some manner.

[mansr]

1 minute ago, fas42 said:

Recordings sound enormously different from each other - 'quality' recordings are boring, generally - for the same reason that a super smooth, straight highway is far less interesting that a winding country road, for an enthusiastic driver - they are both means for people to progress quickly between two points, but I know which I prefer ...

Have you heard any Sound Liaison recordings? I consider them good quality and not the least bit boring.

[adamdea]

1 hour ago, Audiophile Neuroscience said:

@adamdea thanks, you have provided quite a few references. Please give me a little time to digest them,cheers.

Certo 

[fas42]

2 minutes ago, mansr said:

Have you heard any Sound Liaison recordings? I consider them good quality and not the least bit boring.

 

No, I'll check them out ...

[Audiophile Neuroscience]

13 minutes ago, mansr said:

Not in the slightest. From a stereo recording, it is impossible to calculate the distance to a sound source. It is also impossible to calculate the distance a listener will perceive it to be at. The two values need not even be the same. In fact, they are not even likely to match. For left/right direction it is an entirely different situation. If the microphone pickup pattern is known, the direction of a sound can be calculated precisely, and the perceived direction can be predicted with decent accuracy although it depends to an extent on both the room and the individual. This has been my position all along.

 

Would it be a fair compromise to say that in 2 channel recordings there are accurate lateral distance cues (for objective reaons you explained) and that for depth and height, far less accurate, less precise and ambiguous cues varying with the skill of the recordist?

[sandyk]

1 hour ago, dalethorn said:

I've been making ad-hoc binaural recordings with the Sennheiser-Apogee Ambeo headset, and each recording sounds way different depending on numerous factors affecting the rooms I record in, or not.  It's clear to me that a quality recording setup that addresses ambient noise, reflective surfaces, ideal listening location, etc. is vastly more important than the 'HRTF' and related properties that headphone listeners read so much about.  I've been reviewing headphones for a few years, but haven't attempted "quality" recordings until a few weeks ago.  I think if the average headphone user could use the Ambeo headset for a few weeks, they'd appreciate their audiophile music recordings a lot more.

 Dale

 Have you tried playing those binaural recordings via speakers as well, and if so, did you obtain any Height Cues from them ?

Many people when first playing around in this area, not just with Binaural recordings , often record a thunderstorm as well, did you also try this ?

 

Alex

[gmgraves]

17 hours ago, fas42 said:

I'm not worrying about vertical positioning - and they can move sounds back and forth, by the application of reverb. The latter can create deeply cavernous spaces, with ease - yes, totally 'fake', but one's hearing can pick up the meaning, with no problems ...

 No, you can move an ENTIRE ensemble nearer or further by the ratio of direct sound to reverb, but you can't move individual instruments in an ensemble using reverb. 

 

17 hours ago, fas42 said:

It does depend on the playback, because if this is not of sufficient quality the cues are too confusing, too subtle; it just sounds like a mess. One reason pop recordings can be unpleasant to listen to is because the depth information is too tangled - presented at a high enough quality level, the mind can untangle what's going on, and everything makes sense. This is so reliable that I use complex pop recordings to assess unknown rigs - if what I hear is a mess, it means that the playback is doing too much damage to the cues, and it's a fail - for that system.

Simply poppycock. if the imaging cues are in the stereo recording, the cheapest, simplest stereo system will replay them in all their dimensions! My desktop system images beautifully and it consists of a $400 hybrid tube/SS amplifier and a $200 dollar pair of bass reflex "mini-monitors"!

 

17 hours ago, fas42 said:

Consider a jazz trio: the three are complete separated, use headphones to synchronise their playing with each other; the left player is set up in a huge hall, and recorded; the centre one is in a normal sized room, and mic'ed the same way; and then the right player is in a tiny recording booth, and similarly captured. The three tracks are mixed, with the lateral positions adjusted to give that left, centre and right, but no other 'fixing'.

 

In the playback, what will a listener hear?

 

Three channel mono.  A group of instruments in the left channel, a group of instruments in the right channel, and a group of instruments and/or the vocalist in the phantom center channel! In you're rather extreme example, the center and right channels will have artificial reverb added to match that of the musician(s) in the left channel. But this is totally irrelevant to my missive. I'm not talking about ambience our reverb, I'm talking about "on-stage" imaging.

[gmgraves]

14 hours ago, adamdea said:

I think this exchange is becoming futile because you are refusing to approach this from the starting point of how you actually localise sounds in the real world. Ignoring non auditory information (which is significant) this involves your brain analysing the differences in patterns of sound between the ears. Aside from general inter-aural time and ampitude differences there are spectral differences caused by the shape of your neck head and pinna. 

Those differences are crucial to much of our ability to localise sound. They are not encoded in an ordinary stereo recording. When the sound comes from two speakers not from an actual single point the path to your ears is different. Ergo the spectral cues even from sounds on the median axis are not regenerated. This is not debatable. 

Any person starting with a very basic knowledge of perception knows thus. 

 

So without a dummy head the spectral cues are not all recorded. Sensible discussion can only start by considering the importance of what is and is not recorded 

Well, having made literally hundreds of recordings of symphony orchestras, and many more recordings of smaller ensembles, I can tell you with some authority that you are wrong. In a real stereo recording, those cues are recorded and any stereo playback system can "reconstruct" them in your listening room. There are caveats here, though. Room acoustics can mess up those imaging cues (although the room would have to be pretty bad to do that) and, of course, small, more "point source" speakers tend to image more precisely than do many large speakers, but even big speakers can image well too - it depends on how they were designed. 

[semente]

48 minutes ago, Audiophile Neuroscience said:

 

Would it be a fair compromise to say that in 2 channel recordings there are accurate lateral distance cues (for objective reaons you explained) and that for depth and height, far less accurate, less precise and ambiguous cues varying with the skill of the recordist?

 

In pure two channel orchestral recordings you can hear the tympani in the back, and this lead to the unpleasant practice of spot-mic'ing them. You know it's in the back because there's a lot of diffuse sound coming through, sometimes with a bit of delay, and the sound is lower in level than that of the string sections.

 

I have just turned off my laptop but tomorrow I will try to find a good example, probably from Dorian.

 

As for height, I can't figure out what kind of mechanism would produce an impression of vertical position without using another pair of speakers place vertically above the regular pair and recording an extra pair of channels...

[gmgraves]

11 hours ago, mansr said:

You claimed that two regular microphones capture all the positioning cues a person would hear. Since these include HRTF effects, it follows that the microphones somehow, magically account for this. Binaural heads are an attempt to do exactly this, and it works to an extent. However, you insist that even a pair of plain cardioids will do it. That is preposterous.

 

Yes it is preposterous! That's one reason why I never mentioned Head Related Transfer Function. Microphones don't capture that (and aren't "trying" to capture it and don't need to capture it) and binaural heads aren't successful at capturing it (that's why binaural is so poor at localizing sounds behind the listener). Microphones, if used correctly, pick up the cues from the differences in sound intensity and phase that are created by the musical ensemble. These differences in perspective are capture by the recording medium and played back over speakers. The HRTF effects occur at the listener's head, in his own listening environment not at the recording microphone. I'd have thought that you would have realized that.

[gmgraves]

3 hours ago, semente said:

 

If we use dynamic compression and reduce gain wouldn't it be possible to "move" an instrument back in a recording?

Offhand, I don't see how. Perhaps if you were to explain your theory a bit more completely, I could see where you're coming from. 

[Audiophile Neuroscience]

30 minutes ago, semente said:

 

In pure two channel orchestral recordings you can hear the tympani in the back, and this lead to the unpleasant practice of spot-mic'ing them. You know it's in the back because there's a lot of diffuse sound coming through, sometimes with a bit of delay, and the sound is lower in level than that of the string sections.

 

I can hear similar things quite clearly and also proposed similar/same mechanisms for the perceptual cues underpinning the illusion.Research confirmed this, at least the one I previously referred to.

 

Quote

I have just turned off my laptop but tomorrow I will try to find a good example, probably from Dorian.

 

Thank you but I have plenty of good examples . Edit, but of course for others that would be great.

 

Quote

As for height, I can't figure out what kind of mechanism would produce an impression of vertical position without using another pair of speakers place vertically above the regular pair and recording an extra pair of channels...

I can't either but others have proposed mechanisms.I can't remember what that paper said now about height but will look again. George has good ideas on the topic. All I know is I can hear height illusions, not as precise as lateral info, but there, sometimes uncanny.

[gmgraves]

1 hour ago, fas42 said:

 

Recordings sound enormously different from each other - 'quality' recordings are boring, generally - for the same reason that a super smooth, straight highway is far less interesting that a winding country road, for an enthusiastic driver - they are both means for people to progress quickly between two points, but I know which I prefer ...

 

Which means the cues that provide depth information are not "standard" in their quality, and level. They are always there, but vary in their audibility - between every recording. The higher the "resolving power" of the playback rig, the more likely a particular recording will show obvious depth - ultimately, it becomes impossible to find a recording that doesn't present a 3D picture in some manner.

That's where you and I part company. In my experience, a recording either has decent imaging or it doesn't. Realistic imaging is what I do, what I got into recording for. Most orchestral recordings don't image at all because they use a forrest of microphones all mixed together electronically. Most Jazz recordings don't image at all because they are recorded using the traditional three-channel-mono arrangement. Some people don't care about such things, but it is MY THING! If the imaging is not on the recording, (and I've said this before) then the finest stereo system in the world can't make the recording image, I can pick-out a studio recording a mile away. To me the two-dimensional aspects of it stick-out like a sore thumb.  

[adamdea]

1 hour ago, Audiophile Neuroscience said:

 

Would it be a fair compromise to say that in 2 channel recordings there are accurate lateral distance cues (for objective reaons you explained) and that for depth and height, far less accurate, less precise and ambiguous cues varying with the skill of the recordist?

I reckon delete height

“and that for depth far less accurate, less precise and ambiguous cues which will be subject to further confusion in reproduction and will be interpreted differently by different subjects depending on unpredictable characteristics of the room, reproduction system, hrtf and other non auditory modes,  both direct and indirect so far as they contribute to the subject’s model of the listening space.”

[fas42]

51 minutes ago, gmgraves said:

 No, you can move an ENTIRE ensemble nearer or further by the ratio of direct sound to reverb, but you can't move individual instruments in an ensemble using reverb. 

 

You can't ?? ... if the instruments are recorded separately, as in my post of the jazz trio - and different levels, settings of reverb are added to the individual tracks, then the result is a layering of the depth positions - I hear this behaviour in nearly all pop recordings; it's clearly defined to my ears.
 

Quote

 

Simply poppycock. if the imaging cues are in the stereo recording, the cheapest, simplest stereo system will replay them in all their dimensions! My desktop system images beautifully and it consists of a $400 hybrid tube/SS amplifier and a $200 dollar pair of bass reflex "mini-monitors"!

 

 

If the cues are extemely obvious then, yes, the lowest order playback will make the presentation very clear. But if the cues are more subtle, then only better playback will reveal that detail - the cues in the recordings are not on/off switches for our hearing, either present or completely absent - again, everything in audio is a continuum; there are always shades of grey.

 

Quote

Three channel mono.  A group of instruments in the left channel, a group of instruments in the right channel, and a group of instruments and/or the vocalist in the phantom center channel! In you're rather extreme example, the center and right channels will have artificial reverb added to match that of the musician(s) in the left channel. But this is totally irrelevant to my missive. I'm not talking about ambience our reverb, I'm talking about "on-stage" imaging.

 

No, on a good system you will hear the performers overlaid upon each other, in the correct lateral position - with each 'operating' in the space as recorded - there will be a combined sound, and also the sound of each instrument echoing in the space as set up. In the hearing, you will be listening to the mix, and then soloing a particular player - exactly as if you had the controls of the desk under your hands.

 

The imaging is a function of everything you can hear; the amount of "on-stage" presentation will be a function of where and how it was recorded.

[mansr]

40 minutes ago, gmgraves said:

The HRTF effects occur at the listener's head, in his own listening environment not at the recording microphone. I'd have thought that you would have realized that.

Correct, and that is exactly why a recording can't convey the necessary height cues. On playback, the HRTF responds to the location of the speakers, not the location of the instruments during recording.

[fas42]

24 minutes ago, gmgraves said:

That's where you and I part company. In my experience, a recording either has decent imaging or it doesn't. Realistic imaging is what I do, what I got into recording for. Most orchestral recordings don't image at all because they use a forrest of microphones all mixed together electronically. Most Jazz recordings don't image at all because they are recorded using the traditional three-channel-mono arrangement. Some people don't care about such things, but it is MY THING! If the imaging is not on the recording, (and I've said this before) then the finest stereo system in the world can't make the recording image, I can pick-out a studio recording a mile away. To me the two-dimensional aspects of it stick-out like a sore thumb.  

 

Live, studio, and manipulated recordings all have a different sense about them - but they can all sound like the "real thing" - not necessarily real in the sense that all the musicians appear to be on a particular stage in front of me, in a particular venue; but real as in that a group of musicans have informally, or formally, got together in some fashion where everyone could be heard, positioning possibly to suit themselves and what they wanted to project - that they are not rigidly in the "correct" position for how one oftens hears that live is neither here nor there, for me; the sound they produce is everything, that they are in the "right place" is far less important.

[STC]

27 minutes ago, fas42 said:

You can't ?? ... if the instruments are recorded separately, as in my post of the jazz trio - and different levels, settings of reverb are added to the individual tracks, then the result is a layering of the depth positions - I hear this behaviour in nearly all pop recordings; it's clearly defined to my ears.

 

+1 I agree to this.

[STC]

19 minutes ago, mansr said:

Correct, and that is exactly why a recording can't convey the necessary height cues. On playback, the HRTF responds to the location of the speakers, not the location of the instruments during recording.

 

I believe we are still talking about stereo playback with two loudspeakers. In a good proper stereo recordings like what you find with Linn or Chesky (and hopefully no spot miking is involved in the recordings), the sound from the speakers represents the actual soundstange of the ensemble size.

 

In good stereo recordings, sounds comes from anywhere between the two speakers including sound behind the back of the speakers. Unlike in nature, if a sax is located is at about 55 degrees on the left on stage to a listener, you need two speakers to create the phantom location with stereo recordings. A left speakers placed at 60 degrees cannot create the 55 degree placement of the sax without the right speakers. That's the illusion which we get with stereo and therefore also answers why stereo without crosstalk cancellation will never ever sound real.

 

HRTF, is not responding to the locations of the speakers but the sum of two sounds from both speakers. There will always be error with stereophonics and that is another reason why we can easily identify a recorded sound and live performance.

 

I don't know why this is even an issue as there are already tools that can measure stereo width and depth of the recordings. I have used them to measure which setup got better depth and soundstage.

 

One such tool that I used was MusicScope from Xeviro.

 

 

The stereo meter was a great tool for us to compare various recordings soundstage and depth quality. 

 

 

 

[mansr]

27 minutes ago, STC said:

HRTF, is not responding to the locations of the speakers but the sum of two sounds from both speakers.

Same thing.

[STC]

39 minutes ago, mansr said:

Same thing.

 

So if I have a guitar at 30 degrees on the left and a mouth organ at 30 degrees on the right captured with best stereo microphone and played back with the best system and MBL 101 placed at 60 degrees apart , they are not going to to sound alike due to HRTF?

[dalethorn]

3 hours ago, fas42 said:

 

Recordings sound enormously different from each other - 'quality' recordings are boring, generally - for the same reason that a super smooth, straight highway is far less interesting that a winding country road, for an enthusiastic driver - they are both means for people to progress quickly between two points, but I know which I prefer ...

 

Which means the cues that provide depth information are not "standard" in their quality, and level. They are always there, but vary in their audibility - between every recording. The higher the "resolving power" of the playback rig, the more likely a particular recording will show obvious depth - ultimately, it becomes impossible to find a recording that doesn't present a 3D picture in some manner.

 

Not what I tried to say.  What I am saying is that real recordings that companies sell have a lot of work go into them that takes time and skill to fix the various issues, and one person with a binaural headset cannot get anywhere near that.  Nobody would pay money to hear the binaural crud I capture, simply because I'm not doing the work to make the sound pleasant, agreeable, less irritating, etc.  And those things are far, far more important than piffle like HRTF.

[dalethorn]

3 hours ago, sandyk said:

 Dale

 Have you tried playing those binaural recordings via speakers as well, and if so, did you obtain any Height Cues from them ?

Many people when first playing around in this area, not just with Binaural recordings , often record a thunderstorm as well, did you also try this ?

 

Alex

 

I didn't make a representative effort.  Certainly the speaker presentation would be more pleasant, which I suppose is an odd thing to say about a binaural recording.  I think what it really means is when you do a lousy job on a recording, piping it right into your ears with a headphone isn't all that pleasant.

[dalethorn]

8 hours ago, Sonicularity said:

What would happen if someone swapped the left and right channels on their headphones?   If someone is perceiving depth and positioning in front of them, then shouldn't everything appear behind the listener if they flipped their headphones around?  No?  Maybe they need to switch the headphones around and face the opposite direction they normally sit? 

 

For me at least, my ears hear differently, and while I've adjusted psychoacoustically to hear a natural sound that way, switching left and right on familiar recordings sounds really weird to me.  It's nearly as bad as having a channel out of phase.

[dalethorn]

5 hours ago, semente said:

 

When you listen to a piano recital in a music hall there's direct sound coming from the piano and reflected sound coming from every non-absorbing surface. The latter arrives from every direction.

The most significant shortcoming of 2- or multi-channel stereo is it's inability to recreate this reflected sound. It's undoubtedly better than mono but still far from the real thing.

 

Ray Kimber in his IsoMike technique claims to correct some of that, which he says is certain left-right interference rather than a simple failure.  His recordings do sound pretty good.