[gregorio]

Our spatial hearing isn't idiotic. It knows how to "get the picture" as long as the cues make sense.

No, our brain will try to make sense of whatever it's given! Whether or not the "sense" it ends up creating is likeable/preferable, depends on the individual.

[1] True, only binaural can reach very high level of real world feel, but crossfeed modified the sound to be if not "real world" at least more natural and pleasant. [2] The problem with binaural is that you need your own or at least very similar HRTF-signature for it to work really well. [3] Crossfeed simulates HRTF at so coarse level that the result works for anybody. [4] "Real world" is even today a challenge in sound reproduction.

1. No, binaural is not good at creating a real world feel, it's only good at capturing and reproducing the real world of the sound which enters your ears!
2. No, that's only one of the problems with binaural.
3. It's a course simulation of a simulation, so pretty much it doesn't "work" for anybody! That doesn't mean to say it can't be preferred by some though.
4. It's not a particular challenge but that's irrelevant anyway because almost no one is trying to create a "real world" sound recording or reproduction because in the real world no one perceives the actual sound entering their ears!

You just have to keep coming back to referencing the "real world", a real world which usually does not exist and even when it does, that's not what we're trying to record or reproduce anyway!

2a. Of course, but that doesn't change how spatial hearing works.

Absolutely it does! It wouldn't make a difference if your hugely oversimplified concept of how "spatial hearing works" were correct but as it's incorrect, processing absolutely can and does change our spatial perception!

1. Translates well when listened to with speaker or headphones? [1a] Decca Tree produces huge ITD values totally incompatible with our spatial hearing without crossfeed. ... ITDs of even 5 ms make no sense at all. [1b] Maybe elephants would consider such ITD ok. [1c] I am not a fan of Decca Tree ...
[2] How?
[3] The less "binaural" it sounds the more evident it is it is for speakers. If the recording has been "modified" enough for HPs then I listen to it without crossfeed of course, but that happens rarely.

1. Potentially either, depending.
1a. Well, that depends. A Decca Tree would produce fairly large ITD values, which *might* be incompatible with our spatial hearing if it were just a two mic stereo pattern but it's not, it's a 3 mic array!! So, what ITDs of 5ms are you talking about? Firstly, a Decca Tree is never placed on the same horizontal plane as the instruments so at most we're talking about around 4ms timing difference between mics B and C but if we take the violins as an example, between mics A and B the difference would be variable. Some of the violins might have a timing difference up to about 4ms between mic A and B, other violins would have 0ms and most anywhere between 0 and about 3ms. Mic A is panned to the centre of the stereo image, in this example acting effectively as a crossfeed. Being a triangle, we've always got this type of crossfeed interaction between all the mics! Additionally, the mics are not just recording the instruments but also the reflections of those instruments from all the surfaces in the recording venue. So we've now got a bunch of signals, all over the place with varying timing differentials from 0ms up to about 2 secs or so. And furthermore, we've generally got a relatively small ILD difference between A, B and C, although that varies with frequency with a Decca Tree. But heck, let's just crossfeed it anyway!
1b. Yes, although according to your "simplified to the point of nonsense" it would have to be an elephant with 3 ears sitting 10 feet above the conductor!
1c. Yes, agreed. It can produce a rather dry and narrow/overly focused image, which is why we almost never use a Decca Tree on it's own, even quite early on it was commonly paired with outrigger mics. But oh no, now our elephant is going to need 5 ears and two of them would need to be about 50 feet apart! Then there's room mics and spot mics and an elephant with up to around 40 ears placed in the most ridiculous positions! Questions: A. How do you know when you're listening to a recording which used only a Decca Tree? B. If the vast majority of the recordings you're listening to have a Decca Tree in combination with various other mics, how do you isolate the Decca Tree mics from all the others to conclude that you're not a fan of Decca Tree?
2. Already explained! We have a complex interaction of level differences, reflections and timing differences, plus frequency factors such as masking, absorption, cancellations and summing and you're reducing all of that down to a simplistic crossfeed with 250ms delay.
3. And how much is "enough" modification? Regardless of the amount of modification, you are damaging or destroying it with crossfeed. Effectively then, it's just your subjective preference for improving some aspects of spatial positioning while damaging/destroying others!

[1] With speakers it does serve purpose because acoustic crossfeed makes sure the ITDs at our ears are scaled to "allowed" levels. [2] With headphones without crossfeed things go horribly wrong.
[3] If the intent was to have that bass line close to my left ear then I think the producer is an idiot. Who in their right mind would have such intents?

1. A. There are no "allowed" or disallowed levels, only levels we might or might not find subjectively more pleasing. B. Despite being told numerous times, you ignore the obvious fact that "acoustic crossfeed" has the word "acoustic" in it and therefore includes acoustics! You then repeatedly ask "how it's different", because obviously if you're going to ignore acoustics, the interaction and perception of acoustics, then there is no difference!
2. Things can go somewhat wrong, other things can go somewhat wrong caused by crossfeed. Which wrong things do you prefer? You've made clear your preference but that's not a preference shared by all and we don't have to be idiots or delusional not to share your preference!!
3. Again, your judgement and preference. Maybe I would think that producer was an idiot too, maybe the producer really is an idiot but maybe I've just missed the point. Almost all musical developments throughout the history of music (and it's production) have been negatively received by some/many. Regardless, I want to hear what the producer/artists intended, even if they were all idiots!

No, not more "pleasant". More precise because room acoustics is not convoluted into the music.

That's obviously not true, although maybe it is to you because you're ignoring room acoustics but then if you're going to ignore room acoustics how come you're making statements about it?

gregorio and pinnahertz in my opion are a) too puristic and b) don't place various things in proper perspective.

Make up your mind, which is it, are we too puristic or spatially ignorant? We can't be both! In actual fact we are neither. You are fixated on some "real world" which does not exist and which we're not trying to recreate anyway and therefore effectively talking about eliminating spatial distortion from recordings which are deliberately massively spatially distorted! But many of your assertions to support your preference are simply not true and/or based on incorrect conclusions/assumptions. I'm not even going to get into your 1kHz crossfeed crossover, which would sometimes result in some really bizarre artefacts but if you like it that way, that's your choice.

G

 

[71 dB]

1. Of course, but the material is mixed on speakers, that's the target.
2. I could easily make the same argument in reverse: the image with cross-feed without ITD is broken.

3. Again...I could make the same remark in reverse. I'm afraid you'll either have to accept my history with it or reject it. I really don't care. It doesn't change my opinions, and it certainly won't change yours.

1. Yes, that has been my point all the time. The material is mixed on/for speakers.
2. Without ITD? Crossfeed doesn't nullify ITD.
3. I accept your fine history. You must be talented and hard working according to what you have accomplished. Most of what you write on this board I agree totally with or learn from, but for some reason crossfeed is something we disagree a lot about. Maybe it's because I listen to modern classical recordings a lot and those recordings have great spatiality that makes crossfeed shine. I did not listen to 70's rock in my youth. My father is a jazz guy who thinks all rock music is for idiots. So instead of hearing The Who, my childhood was filled with Max Roach's drum solos and Clifford Brown's virtuosity on speakers. My music listening was passive up until about 1988 when I was 17. I got into acid house and the following trends of modern electronic dance music. I got into classical music in 1997. I didn't listen to any rock music until 2001 when I was 30. I had though pretty much all music from the 70's sucked, but then I discoved Tangerine Dream and King Crimson in 2008, both almost unknown in Finland. Spotify is good for exploring. The point is I discovered the 70's music with excessive ILD at an older age so I don't have childhood nostalgia for that like you seem to have. My concept of the 70's is more from TV, which of course was monophonic, but hey, we got a color-tv in 1975 I think. I was 4.

 

[WoodyLuvr]

So to get back to the question crossfeed or not? What is the consensus?

 

[71 dB]

Already explained! We have a complex interaction of level differences, reflections and timing differences, plus frequency factors such as masking, absorption, cancellations and summing and you're reducing all of that down to a simplistic crossfeed with 250ms delay.

You call a hard panned ping pong stereo recording something that contains "complex interaction of level differences, reflections and timing differences"? Sorry, but it doesn't. Ping Pong is spatial nonsense and it has to be modified into something that makes sense to our spatial hearing. Crossfeed is one method to do that. Crossfeed is simple, but at least it's something. No crossfeed is nothing. 250 µs (not 250 ms!) delays are much better than nothing. It amazes me how you keep thinking excessive ILD/ITD are ok. Their are spatial nonsense for elephants.

 

[71 dB]

So to get back to the question crossfeed or not? What is the consensus?

Crossfeed is the most controversal topic in the history of humankind. There is no hope of consensus if you ask me. Peace between Palestinians and Israel is easier than consensus on crossfeed.

 

[WoodyLuvr]

"finally" ^_^. it was one of the first xfeed VST I used on foobar.



modo talk: please refrain from making personal attacks like you did. Katz book is a classic for sure. but you might want to also have a look at headfi's rules. not a best seller by any mean, but fairly relevant to being in the forum.

You would recommend this VST (xfeed) over the Case Meier version?

 

[ironmine]

"finally" ^_^. it was one of the first xfeed VST I used on foobar.

This crossfeed would benefit greatly if it had a mid/side level adjustment. To my ears, it makes the center too hollow, so I like putting in front of this crossfeed some other plugin (e.g., Fabfilter Pro-Q2) that can reduce the level of the side channel.

Anyway, so far, 112 dB Redline Monitor remains my default crossfeed. It just preserves more details and the bass sounds fabulously natural in it. Also, it has a very user-friendly interface with the settings which are understandable and do their job nicely. I can bring "speakers" closer for more details, or I can move them away for a more relaxed sound. I can also adjust the volume of the center to make the soundstage either flat or curved forward (with the center closer to me than the sides) or curved outwards (with the center distanced further from me than the sides).

 

[WoodyLuvr]

This crossfeed would benefit greatly if it had a mid/side level adjustment. To my ears, it makes the center too hollow, so I like putting in front of this crossfeed some other plugin (e.g., Fabfilter Pro-Q2) that can reduce the level of the side channel.

Anyway, so far, 112 dB Redline Monitor remains my default crossfeed. It just preserves more details and the bass sounds fabulously natural in it. Also, it has a very user-friendly interface with the settings which are understandable and do their job nicely. I can bring closer "speakers" for more details, or I can move them away for a more relaxed sound. I can also adjust the volume of the center to make the soundstage either flat or curved forward (with the center closer to me than the sides) or curved outwards (with the center distanced further from me than the sides).

Okay I will consider buying that then. How about VST-chainer called Console ART-Teknika - http://www.console.jp/en/ is that Windows 10 friendly as well?

 

[71 dB]

That's obviously not true, although maybe it is to you because you're ignoring room acoustics but then if you're going to ignore room acoustics how come you're making statements about it?

Typical room has more reverberation and worse acoustics than a studio. Studio is kind of in the middle of the no acoustics headphones and too much acoustics living rooms. Botg headphones and speakers are "wrong", just in opposite directions. In studio we have acoustic crossfeed, with speakers we have acoustic crossfeed. With headphones we have either nothing or something such as crossfeed. If this opinions makes me ignorant then...

Make up your mind, which is it, are we too puristic or spatially ignorant? We can't be both! In actual fact we are neither. You are fixated on some "real world" which does not exist and which we're not trying to recreate anyway and therefore effectively talking about eliminating spatial distortion from recordings which are deliberately massively spatially distorted! But many of your assertions to support your preference are simply not true and/or based on incorrect conclusions/assumptions. I'm not even going to get into your 1kHz crossfeed crossover, which would sometimes result in some really bizarre artefacts but if you like it that way, that's your choice.

G

Purism and igronance aren't mutually exclusive, in fact lack of understanding can lead to dogmatic purism. For example a purist may assume that downsampling hi-res audio to 16/44.1 reduces audible resolution because he/she doesn't understand what happens sonicly and how 16/44.1 is all you need in consumer audio. Similarly someone may think crossfeed messes up spatial information. The opposite is true, messed up spatiality is scaled to be less messed up. Elephant audio becomes human audio such like downsampling 192 kHz to 44.1 kHz makes bat audio human audio.

I have listened to my crossfeeders half a decade for countless hours and never have I encountered "really bizarre artefacts" unless you mean natural pleasant sound is a bizarre artefact. To me excessive ILD/ITD is the source of spatial artefacts. So yes, crossfeed is my choice.

 

[71 dB]

1. A. There are no "allowed" or disallowed levels, only levels we might or might not find subjectively more pleasing. B. Despite being told numerous times, you ignore the obvious fact that "acoustic crossfeed" has the word "acoustic" in it and therefore includes acoustics! You then repeatedly ask "how it's different", because obviously if you're going to ignore acoustics, the interaction and perception of acoustics, then there is no difference!
2. Things can go somewhat wrong, other things can go somewhat wrong caused by crossfeed. Which wrong things do you prefer? You've made clear your preference but that's not a preference shared by all and we don't have to be idiots or delusional not to share your preference!!
3. Again, your judgement and preference. Maybe I would think that producer was an idiot too, maybe the producer really is an idiot but maybe I've just missed the point. Almost all musical developments throughout the history of music (and it's production) have been negatively received by some/many. Regardless, I want to hear what the producer/artists intended, even if they were all idiots!

1. The limit (set by physical reasons) for ITD if about 640 µs for sounds originating not very close and 700-800 µs for close sounds. To me mixing music to be very close to your head doesn't make much sense, so I'd limit ITD to 640 µs. If very close sound is the intent, then such intents aren't respected by speakers. Acoustic crossfeed is of course different from electronic crossfeed, but both reduce ILD/ITD information following the same principle: More at low frequencies and less at higher frequencies. Crossfeed is much closer to acoustic crossfeed than no crossfeed, so why would anyone choose no crossfeed?
2. Which wrong do I prefer? Crossfeed wrong of course because I don't even recognize it being wrong whereas spatial distortion feels very wrong. People who like crossfeed a lot exists in numbers, but even if I was the only one I would believe in it. How often are masses right? The scientific facts of human hearing are what they are despite of the weird preferences people hold about spatiality in music for technical, cultural, nostalgic and commercial reasons.
3. Well, make a call to those producers and artists and ask whether they want their music listened with or without crossfeed. The answer might surprise you. I don't need to call anyone, because I use my own head to figure out the proper crossfeed setting.

That's obviously not true, although maybe it is to you because you're ignoring room acoustics but then if you're going to ignore room acoustics how come you're making statements about it?

I don't ignore room acoustics, but headphones don't have room acoustics! Do you want your no room acoustics with or without spatial distortion? That's the question here?

 

[jgazal]

4. It's not a particular challenge but that's irrelevant anyway because almost no one is trying to create a "real world" sound recording or reproduction because in the real world no one perceives the actual sound entering their ears!

You just have to keep coming back to referencing the "real world", a real world which usually does not exist and even when it does, that's not what we're trying to record or reproduce anyway!

I am glad you used the word almost.

It is always a matter of reference or perspective, isn’t it?

But “almost no one” may become in the short term or may be already a wrong degree.

Let’s see a few examples:

1. Sennheiser Ambeo (first order ambisonics)

2. YouTube VR

YouTube VR recommends content With first order ambisonics, probably downmixed to binaural with a generic HRTF. Potentially better with a Realiser crossfeed free PRIR.​

Attention: use YouTube app to rotate your visual point of view in the monoscopic 360 degree video!​

2.A. Cindy Crawford closet - Vogue

Important demonstration for wives!

Cindy voice need to derotate in order to enhance the immersion.

Mobile devices and tables feed the tracking. Probably better with crosstalk cancellation with speakers or crossfeed free externalization with headphones.

You may want to follow these instructions:

Try to listen with loudspeakers in the very near field at more and less +10 and -10 degrees apart and with two pillows one in front of your nose in the median plane and the other at the top of your head to avoid ceiling reflections (or get an IPad Air with stereo speakers, touch your nose in the screen and direct the loudspeakers sound towards your ears with the palm of your hands; your own head will shadow the crosstalk).

2.B. Bewitched love - Manuel de Falla - Orchestre national d'Île-de-France

Audio content that derotates according to the visual point of view! This is an spot microphones mixing probably with binaural synthesis.

2.C. Showcase/Showdown Eigenbeams/Ambisonics - mh acoustics

Audio content that derotates according to the visual point of view! This is an eigenmike probably downmixed to first order ambisonics and then downmixed again to binaural with a generic HRTF and finally streamed through YouTube.

With a HOA streaming, a personalized HRTF an Realiser A16 this would be the lowest distortion path available (what Professor Choueiri says “being fooled by audio”).

Can you imagine that with an stereoscopic video?

Ping pong - compare with Ricoh below...

3. Netflix VR

Probably atmos bed and objects downmixed to binaural with generic HRTF. Potentially better with a Realiser crossfeed free PRIR.

4. Google Daydream “Fantastic Beasts” VR

Maybe a mix of first order ambisonics and objects? Or maybe full binaural synthesis (similar to BACCH-dSP)? IDK.

5. Ricoh Theta V with “spatial audio”

Some frustrating content done with 360 monoscopic cameras in which the audio does not derotate.

Probably not Ambisonics.

Just showing to demonstrate the potential of home-made music oriented video distribution with spatial accuracy:

Ping pong - compare with eigenmike above...

You may see that those are not minor players in the entertainment market...

(...)
3. Again, (...). Maybe I would think that producer was an idiot too, maybe the producer really is an idiot but maybe I've just missed the point. Almost all musical developments throughout the history of music (and it's production) have been negatively received by some/many. Regardless, I want to hear what the producer/artists intended, even if they were all idiots!
(...)
Make up your mind, which is it, are we too puristic or spatially ignorant? We can't be both! In actual fact we are neither. (...)
G

If you think about Google Daydream VR, Netflix VR, Samsung VR, Facebook virtual hangouts etc, you will realize that they are whole highways in which immersive sound content formats can reach consumer’s virtual environments.

You just need to use them.

I guess consumers don’t care about the bit depth or the the sample rate of distributed content, but I believe they going to click in sponsored pages and happily pay for streaming/downloads that render audio spatial immersion that is accurately correlated with visual cues.

That is a huge incentive for mastering engineers to persue spatial accuracy with reference to the original sound-field or virtual designed visual cues, don’t you think?

I know the number of streamings with Spotify, Deezer, Pandora, Tunein Radio etc is higher than music streamed with video formats, but is that an intrinsic barrier?

Is there any other objection to distribute music content through such distribution channels and consumers virtual environments?

There is a reason why Google has chosen - ambisonics - to distribute vr content in YouTube VR and Daydream VR.

Perhaps you may financially benefit from reducing your negativity about shifting your reference to the real sound-field or virtual designed visual cues.

 

[gregorio]

[1] You call a hard panned ping pong stereo recording something that contains "complex interaction of level differences, reflections and timing differences"?
[2] Crossfeed wrong of course because I don't even recognize it being wrong whereas spatial distortion feels very wrong.

1. Where did I call it that? In fact I called it something quite different but you've chosen to misrepresent what I said, why I wonder?

2. That is EXACTLY my point, thanks for admitting it! It's all about what YOU are able to "recognise" and of course conversely, what YOU are not able to recognise. I am able to recognise that it's wrong (and even more so if I only crossfed below 1kHz), while you are not able to recognise it's wrong. That's not the issue, the issue is that you're calling me spatially ignorant when you are the one incapable of recognising it and you're calling me and everyone else who disagrees with you idiotic and delusional based on YOUR INABILITY!! This is probably a good place to leave this line of discussion!

It is always a matter of reference or perspective, isn’t it?

Particularly in this case, as I've been talking about commercial stereo music releases!

G

 

[gregorio]

(A) Gregorio, I advise you to read Bob Katz' book "Mastering Audio: The Art and the Science".
(B) Modern high-quality VST plugins do not operate at fixed internal sample rates. At least, I never heard about them (and I read the manuals for all the plugins which I use). But I do have a number of the state-of-the-art plugins which, as an option, perform internal oversampling to achieve a higher precision.
(C) 1. Variant A = Taking the signal from 44/16 to 176/24, processing it with plugins and outputting the result to the DAC in 176/24.
2. Variant B = Leaving the signal in 44/16 format, processing it with plugins and outputting the result to the DAC in 44/16.
A will give a better, more accurate result than B.
(D) 2. No, dithering must not be the last step.

A. Why would you advise me to read a book I've already read and discussed with the author?

B. Some do have fixed sample rates. Most have variable rates, some have oversampling options and some upsample/downsample without mentioning it in the manual. Upsampling does NOT increase precision, how can there be higher precision from processing frequencies which do not exist! Why isn't that obvious to you?

C. What have those two variants got to do with your original statement or my response? An at least equally and possibly more accurate variant (which you haven't mentioned?) than your variant A would be: 44/16 (without upsampling), processing it with plugins and outputting the result to the DAC in 44/24. Although more technically more accurate, it almost certainly wouldn't be audibly better than outputting at 44/16. Again, how can processing 16 bits + 8 additional zeros be higher precision than processing 16 bits + no additional zeros, when both are calculated at 64bit float precision?

D. I suggest YOU read Bob Katz' book "Mastering Audio: The Art and the Science"!!!

G

 

[jgazal]

Particularly in this case, as I've been talking about commercial stereo music releases!
G

I liked that!

Added a new video, now with music oriented content.

 

[71 dB]

That is EXACTLY my point, thanks for admitting it! It's all about what YOU are able to "recognise" and of course conversely, what YOU are not able to recognise. I am able to recognise that it's wrong (and even more so if I only crossfed below 1kHz), while you are not able to recognise it's wrong. That's not the issue, the issue is that you're calling me spatially ignorant when you are the one incapable of recognising it and you're calling me and everyone else who disagrees with you idiotic and delusional based on YOUR INABILITY!! This is probably a good place to leave this line of discussion!

I don't care about errors I can't hear/notice, and even if I did, I'd have to compare pros and cons. The pros of crossfeed are massive. The cons are theoretical. So, I wasn't admitting that much. Maybe your ears are superior to my ears and somehow welcome excessive ILD/ITD information beyond the theories of spatial hearing, but I can only use my own ears to listen to music. My ears seem to obey the theories of human spatial hearing. So, I apologize for calling you ignorant while not realizing you are an überhuman.