Ambisonics had it all in the 1970s - a genuine "fix in the mix" capability where you could listen in any row you desired by a simple control. More modern and complicated, multi-miked recordings do not preserve the information recorded via coincident microphone systems, and instead essentially smear the necessary information or attenuate it it altogether. My earlier posts on early reflections and the like offer a high level explanation of why.
I see in audio there's an analogy to "Apple" - I want something that just works and "Microsoft" - I like to fiddle with the computer.
I blame the Industry Bstrds who killed quad all those years ago. Could be 16 discrete channels in every recording file as a standard by now. Mixer built right into the player; reads an initial "master" setup then you're free to alter levels, EQ, Delay any way you'd like. Send / mix any channel to any speaker - say, up to 8; ignore them all except the mono mic if that's what floats you.
I blame the Industry Bstrds who killed quad all those years ago. Could be 16 discrete channels in every recording file as a standard by now. Mixer built right into the player; reads an initial "master" setup then you're free to alter levels, EQ, Delay any way you'd like. Send / mix any channel to any speaker - say, up to 8; ignore them all except the mono mic if that's what floats you.
Two channels can only encode one spatial dimension of information (in part). Normally with two channels that is the lateral information represented by the difference or side (S) channel. But that is also a compromise, as to represent horizontal information fully, we actually need three channels of information. The result is that stereophonic reproduction cannot encode information well from the sides of the listener.
This misunderstanding is why we ended up with the nonsense of quadrophonics - and the fall-out of that meant the system that did work (ambisonics) never gained sufficient traction. That is a huge pity.
So even if we could add height reliably to stereo material, there is no spare information capacity available to place it without compromising the information that was recorded. This is different to energy preserving transformations discussed previously in this thread that are easily reversed. I would advocate for keeping the recording information at maximum fidelity and adding HRTFs at home as per your preference. They will always be an effect in two channel recording and reproduction.
I would suggest as I tried to previously, that early recordings - especially thos in mono - offer the correct reproduction of early reflections (pressure). And it is from these cues that we establish depth perception. Just don't listen via two loudspeakers, as that will lessen the fidelity considerably!
I think here you are confusing stereo width (which changes the mid to side balance) with stero image breadth, which is a term normally used to describe the perceived width of sound sources (why a piccolo is perceived to be small, and why a bus is perceived to be large). Again early reflections aid our perception here, but image breadth manipulators are generally based of frequency spreaders like "pseudo-stereo". You won't like them
This too is a misconception. Quadrophonics failed precisely because it was based on a false assumption. 16 channels, for example, would be far better given over to higher order ambisonics, but just four would have been a substantial improvement. In fact if we are talking of only horizontal information, three channels actually outperfoms the four of quadrophonics!
I think i see your point but as i understand it you either use shadowing/reflectors at takes ( recording on mics) or playback ( on speakers) to allow the creation of (early) reflections which will then create an immense envelopping soundstage ( envelopping is a hint to me it use reflections to create the effect).
I often used Blumlein couple at takes in close proximity to real stone wall ( irregularly shaped with random reflections) to 'thicken' acoustic guitars ( playing in direction of the wall). It's an interesting effect but do not change couple SRA which stays 90°. Iow if reproduction playback system is free of most ER it stays 90°. If acoustic is not well treated it gives a mishmash of the SRA and seems immersive but at the price of the image i created, there is a loss of imaging and recorded angle isn't reproduced anymore. In my view it isn't high fidelity but an effect.
That said each one it's own and if this is what you like it's fine.
I agree with your two last sentences. That said we are not always trying to produce something homothetic to reality when producing music. The best exemple is reverbs ( or in general treatments) on lead voices. It's rarely natural in any way in 90% of releases. But it doesn't matter as long as it please the artist and the audience.
There would be a stop from most artists. In pop music what is delivered is not only a score but the whole production's choices are part of what is delivered to customers.
I already told but i've seen multiple days of session dedicated to a snare which didn't fit in the mix of an album. Do you really think artists would let consummers take care of what they spent month ( sometimes years) to release in final shape who pleased them?
I'm sure it won't happen except for a few 'experiences' from time to time ( in fact i only know one French artist who did that and it was a way to make buzz and discover new artists for his own label...).
It's not this different wrt bootleg's live recording: artist are against this not because of money only, but they don't have the final artistic choice on what is released...
Last edited:
As I said, audiophiles tend to believe what they choose to believe. Yeah, that’s real scientific.
But I’m referring to the stereo recordings. If only you could hear what I’ve heard with my ears.
That is undoubtedly true and the basis of several commercial systems that have come and gone. However, I would argue a better solution is to use a "purist" coincident microphone recording which can then be shuffled for optimal results. If shadowing is part of the recording process, there is no going back to recover information that was never recorded.
It is exactly that, or rather the lack of smearing of information in time. Where it gets complicated is how the x and z information is perceived when only the y information can be produced accurately.
But better still is the Soundfield microphone (or like set-ups) that offer W (pressure), X, Y and Z (velocities), and ideally the means to reproduce them thereafter too.
I should have been more specific. Pseudo-HRTF recordings don't necessarily present an amorphous 'sense of space' beyond the speaker boundaries. It can place solid images there in a manner similar to well executed QSound.
I forgot you do this as a profession. It's unlikely, save for maybe J-Disc, that the field has much overlap with the nature recording alchemists constantly creating and experimenting with novel mic arrays. In one 2-mic sample coyotes start yipping far field nearly 90 degrees right of my seat. After a minute nearby coyotes start barking much closer and a little further towards the back wall. They're in a canyon. The reverberation sweeps from far right-of-right to far left-of-left constantly getting taller until my entire room left-of-left and floor to ceiling is emanating. It has a recognizable and credible structure, not the phony envelopment of phase rotations, reversals and delays. Even so it's baby steps which ignores left/right speaker crosstalk, center phantom images with the the imprinted HRTF of left/right speakers, etc..
Agreed for much of the recording world this is impractical but practicality doesn't change how humans hear. Cracking this nut may be the next great transformative step in audio reproduction.
Such perceptions of structure in all that we hear, whether it be rooms, canyons, people or loudspeakers, can be found in the bispectrum or cross-bispectrum that relate how different frequencies relate to each other (and if they are actually related).
The brain appears to be geared for just such processes too, so we appear to have embedded a natural hierarchy that goes from basic magnitude via the spectrum to the bispectrum, and then to labelling/naming/identifying of the objects we perceive from these measures.
How to record and reproduce this hierarchy of information with optimal fidelity was established in the 1970s. The first-order Soundfield microphone was the commercial result. Since that time, higher-order microphone systems have been devised, but never appear to gain much traction. One suspects artistry will always trump science in audio engineering.
The microphone system is actually the simplest problem to solve, however, since incorporating the requisite number of loudspeakers in a domestic living environment will always provide a significant stumbling block. Thus we are often left instead with a few loudspeakers and flawed recording/reproduction systems to make use of them to best effect.
But shuffling loudspeaker feeds still appears the most viable means of optimising loudspeakers. Maybe one day in the future we might all give time to calibrating individual HRTFs, and sit with motion detection devices compensating for our natural and continuous head movements? I know already which engineering route I would advocate.
You might’ve missed my edit to my comment where I added time t to the physical coordinates x, y and z. So, there are four (count ‘em) coordinates and all 4 dimensions are captured in the recording. In a perfect world you should be able to reproduce a large, holographic, high, deep and wide soundstage but in reality that is rather difficult because, I am contending, too many things can and do go wrong during playback - especially for CDs. That’s why CDs frequently sound thin, congealed, bass shy, generic, compressed, electronic, synthetic, threadbare, two-dimensional, closed in, airless, bland and like paper mache.
I did miss it! How ever you choose to define them, four information channels are required to characterize the sound field around a listener up to first order (which is about 700Hz). Time is not one of those channels, however, rather it is the basis of all of them.
I find thinking of magnitude (pressure) and three components of velocity in the x, y and z directions (aka B-format) is the most useful way of referring to them. Higher orders of sound field characterization require more information channels, although as expressed in this forum already, shuffling can help with any inherent compromises here.
I would contend that the often-cited term holographic is misplaced. With loudspeakers, there exists the attempted reconstruction of the recorded sound field around one or more listeners. Holography works via a different principle.
This statement I disagree with completely, and for a number of reasons, not least its sweeping and unjust generalization...
In the early days of digital recording, the new-fangled technology was not well-understood. Analogue recordings from the 1980s are often superior to their digital counterparts for this reason. Dithering, for example, is well-known to remove problematic artefacts of digital recording, yet it was not properly established in recording equipment until the 1990s (and then often appearing as a hard-to-find switchable option and left switched off!). This period of CD releases cast an unfortunate shadow (audible glare?) on digital technology, but one that no longer has any validity.
Much so-called hi-fi equipment is still not suited to the dynamic range of digital recording. Valve power amps with single figure power dissipations are a good example. Here we would expect soft-clipping of peak SPLs to be the norm if we want an average SPL to establish reasonable fidelity. And music devoid of its considerable peak SPLs sound subjectively dull at best. I would suggest this is a genuine reason for further discussion if we wish to reveal some of the "imaginary".
All else in your list you could likely ascribe to how the material was recorded or mastered - and even what the producer/engineer intended. Not many recordings are intended for someone sat at home attempting to recreate the experience of being at a recording. And for good reason too, as often listening to what is recorded can resolve much greater information than you would have experienced sat in the recording environment.
I could go on and on here, but I feel we are straying off-topic... CD's, however, are not the "guilty party", and there is nothing imaginary in binary code.