"Rock/Pop" MultiChannel shining in Stereo

[STC]

37 minutes ago, Miska said:

I would much rather have 192/16 recording than 48/24 one.

 

+1. However, my reason is for the use of DSP which I find preserve the SQ when the sampling rate is higher. 

[Kal Rubinson]

1 hour ago, mansr said:

The title was changed this afternoon.

Is there a reason?  All of my comments are referenced to classical music.

[mansr]

Just now, Kal Rubinson said:

Is there a reason?  All of my comments are referenced to classical music.

You'll have to ask the OP.

[Daccord]

57 minutes ago, Kal Rubinson said:

Is there a reason?  All of my comments are referenced to classical music.

All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy.

[fas42]

On 12/2/2018 at 12:53 AM, Le Concombre Masqué said:

Edited on December the 5th : I have not experimented with Classical or Acoustic Jazz but dig that when MCH aims at recreating natural acoustics, reverb, etc, down mixing to Stereo is a non sense. However, when reproducing (better, those who know say) the natural acoustics of the concert hall  etc is not the goal, ie with Talking Heads, Steely Dan or Bjork, MCH to Stereo offers better bass delineation, smoother and more natural vocals and more noticeable soundstage vs the standard Stereo. 

 

 

Again, the smartest solution if you want Rock/Pop to work some 'magic', say, better sense of bass, and completely natural vocals, very rich sound field - is to optimise playback of stereo. Most playback rigs are not clean enough to present ordinary stereo recordings with the necessary lack of distortion and colouration to do this well; meaning that hacks like MCH can be used to add some more heft.

 

Well done Rock/Pop playback can be vastly superior to highly conventional classical recordings, as an immersive experience; because there are "no rules". It only takes competence of the system overall to throw up a powerful illusion with stereo - but currently only highly focused tweaking is likely to allow this to manifest.

[gmgraves]

3 hours ago, mansr said:

The title was changed this afternoon.

Thank you. I was beginning to doubt my sanity (I still am, just not about this subject any more)!

[gmgraves]

2 hours ago, Miska said:

I'm not joking at all. That's why I also use loudspeakers and headphones that can reproduce at least up to 50 kHz.

 

That is certainly part of the hires, and there are many examples where that is the case. As well as many where that is not the case. But for me, hires means wide-band recording.

Tell me Miska, what breed are you? Cocker Spaniel, Lhasa Maltese, Labrador Retriever? ? Cause humans can't hear 25 KHz and if you're a male, much over 30, you probably can't hear 20 KHz any more and that's assuming that you haven't been listening to rock-n-roll at 115 dBm most of your life or been working in a boiler factory!

 

2 hours ago, Miska said:

That is part of it, but to large extent can be dealt with suitable upsampling filters if the content doesn't really reach Nyquist anyway. You have filter artifacts only when your content actually hits the filter.

 

Very true. So that's even less justification for high bandwidth and outrageous sampling rates. 

 

2 hours ago, Miska said:

Lot of recordings have dynamic range much less than 24/32-bit. You are going to have hard time finding few recordings that involves real microphones and preamps and reaches 20-bit dynamic range. If not for anything else than recording venue background noise. Of course completely digital productions with software synthesizers can reach full 32-bit dynamic range because there's never anything analog involved in the production chain.

 

I would much rather have 192/16 recording than 48/24 one.

That's not the point. The point is that 24 (or 32)-bit allows for far more recording headroom than 16-bit. It allows one to use an overall lower record level, stay out of the "mud" at the bottom of the range, while avoiding over-modulation at all costs. This is especially useful if you are doing location recording, which is the only kind I do, so it's sort of a big deal for me where you don't really know how loud the musicians are going to get at any given time. I still don't understand why you feel you need all that top end. Studies I have seen show that there's no music (or harmonics) up there. Even a trumpet, one of the most harmonic-rich instruments has nothing above about 35 KHz, not that most of us audio types could hear it. But that's your business.   

[gmgraves]

2 hours ago, STC said:

 

This is due to the lack of HRTF. Smyth realizer with head tracking is required. QSound binaural too can do it pretty well. Even the frontal stage is not as good as stereo loudspeakers. However,  depth, height, envelopment and spatial information are far superior to loudspeakers playback. The ambiance is more relevant for realism than FR.

 

 

 

Most of this is probably true 20 or 30 years ago. Head tracking binaural is becoming a reality.

 

Once you know how to externalize the sound outside the head, then binaural will be real enough although the FR may not be precise. IIRC, they have taken over 100 different pinnae and created a filter based on the average FR of the different pinnae. This can work , but those at the two extreme ends will never get it right unless the filter takes into consideration your pinna shape. IIRC, that's what SR is doing.

 

Here is the video of a binaural recording where you can hear the rear sound. 

 

 

 

Thank you for that. Basically it's like everything else. You solve all the problems and anything will work well from self-driving cars to binaural sound.

[gmgraves]

1 hour ago, Daccord said:

All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy.

Well, that's one way of looking at it, but here's another: Anyone who's musical tastes allows him or her to be satisfied listening to rock/pop is already a very dull boy (or girl). Just sayin' 

[Kal Rubinson]

1 hour ago, Daccord said:

All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy. All classical and no rock/pop makes Jack a dull boy.

Weren't you supposed to write that 100 times?

FWIW, I do listen to rock/pop/etc. but they are of only subsidiary interest to me.  

[Daccord]

5 minutes ago, gmgraves said:

Well, that's one way of looking at it, but here's another: Anyone who's musical tastes allows him or her to be satisfied listening to rock/pop is already a very dull boy (or girl). Just sayin' 

 

Obviously The Shining never made it to PBS

[Miska]

6 hours ago, gmgraves said:

Tell me Miska, what breed are you? Cocker Spaniel, Lhasa Maltese, Labrador Retriever? ? Cause humans can't hear 25 KHz and if you're a male, much over 30, you probably can't hear 20 KHz any more and that's assuming that you haven't been listening to rock-n-roll at 115 dBm most of your life or been working in a boiler factory!

 

I don't want to go into this again, but I'll just say as I've said before that IMO it is needed for accurate transient reproduction. So that you capture and reproduce all harmonics that exist without limiting bandwidth at any point in the channel. That includes microphones too.

 

6 hours ago, gmgraves said:

Very true. So that's even less justification for high bandwidth and outrageous sampling rates.

 

What is outrageous? I don't see anything outrageous in currently used distribution formats. Some of my ADCs can spit out 768/32 PCM and that would be maybe something I would consider not necessarily worthy distribution format.

 

6 hours ago, gmgraves said:

That's not the point. The point is that 24 (or 32)-bit allows for far more recording headroom than 16-bit. It allows one to use an overall lower record level, stay out of the "mud" at the bottom of the range, while avoiding over-modulation at all costs. This is especially useful if you are doing location recording, which is the only kind I do, so it's sort of a big deal for me where you don't really know how loud the musicians are going to get at any given time.

 

For recording yes! But for final distribution format? When you put something out for distribution you very well know what you are putting out. And remember that you are no limited to just 16-, 24-, or 32-bit. I have demonstrated that you could very nicely distribute for example 120/18 format to optimize file size without losing anything in terms of content. When you output for example to FLAC and have dithered just 18 or 20-bit worth of data and LSBs zeroed, FLAC encoder understands this and compresses much more efficiently. You can inspect the content to see how many bits it actually uses before noise floor.

 

6 hours ago, gmgraves said:

I still don't understand why you feel you need all that top end. Studies I have seen show that there's no music (or harmonics) up there. Even a trumpet, one of the most harmonic-rich instruments has nothing above about 35 KHz, not that most of us audio types could hear it. But that's your business.

 

Studies have show that yes there is certainly content above 20 kHz! Many percussive instruments put out lot of energy above 20 kHz. And I have tested this myself and examined lot of hires recordings too. I run all hires content I buy through analysis so I know what I have.

 

https://www.cco.caltech.edu/~boyk/spectra/spectra.htm

 

From this study the best transient example is claves:

https://www.cco.caltech.edu/~boyk/spectra/11.htm#a

I have tested many of these same things too. Castanets also produce similar spectrum. Maracas produces spectrum closer to crash cymbal:

https://www.cco.caltech.edu/~boyk/spectra/13.htm

 

One of my personal favorites is also soprano  glockenspiel, it puts out distinct spectrum lines and very sharp transient attack.

 

In addition things like electric guitar with nice tube or semiconductor distortion puts out lot of harmonics. And so do synthesizers (because for example modular synths can run square and sawtooth waves).

 

[Ralf11]

11 hours ago, Miska said:

... accurate transient reproduction. So that you capture and reproduce all harmonics that exist without limiting bandwidth at any point in the channel.

....

 

 

I have often wondered about this.  The studies re human HF hearing limits are based on sine wave testing.  A more complex waveform could sound different.

 

As to whether it does... who knows?  It is a testable hypothesis however.

[fas42]

Worrying about those above human hearing frequencies is a waste of time, if the goal is realistic presentation. It's the percussive instruments that do that type of thing, typically, and their contribution is only a tiny, tiny part of the musical message, normally.

 

Synthesizers can do anything, of course - and often sound somewhat boring, and "what's the point of that sound in the mix?!". IME synthesizers need optimal playback far more than conventional, say classical, instruments - in part because we have no memory of what "it's supposed to sound like" - get it right, without worrying about the ultrasonic areas, and this type of sound production can be absolutely "magical".
 

[esldude]

12 hours ago, Miska said:

 

I don't want to go into this again, but I'll just say as I've said before that IMO it is needed for accurate transient reproduction. So that you capture and reproduce all harmonics that exist without limiting bandwidth at any point in the channel. That includes microphones too.

 

snip........................

Of course your ears will limit bandwidth.  So if you have an adequate amount of bandwidth beyond the ears it wouldn't matter if you capture less than the full bandwidth of the source as the ears could not be effected.  Now how much more is needed is an issue that could bear some examination, but I'm pretty sure we don't need 100 khz through put in the entire reproduction chain. 

 

96 khz sample rates with 40 khz bandwidth seems plenty considering few practical microphones for music recording can do more.  Few speakers can manage 40 khz on the other end.  

[esldude]

1 minute ago, esldude said:

 

oops double post. 

[gmgraves]

18 hours ago, Daccord said:

 

Obviously The Shining never made it to PBS

I saw 'The Shining' in the theatre when it first came out. When my then girlfriend and I exited the theatre, there was a line for the next showing. Both of us ran up and down that line telling people to save their money and avoid the movie like the plague! I honestly believe that it was one of the worse films I have ever had the misfortune to see! Luckily I don't remember much about it (Other than "Here's Johnny". Put it out of my mind, altogether, I guess). So, if your little rock-'n-roll taunt was based on something in that film, you'll forgive me for not picking up on it.  

[Miska]

7 minutes ago, esldude said:

Of course your ears will limit bandwidth.  So if you have an adequate amount of bandwidth beyond the ears it wouldn't matter if you capture less than the full bandwidth of the source as the ears could not be effected.  Now how much more is needed is an issue that could bear some examination, but I'm pretty sure we don't need 100 khz through put in the entire reproduction chain. 

 

96 khz sample rates with 40 khz bandwidth seems plenty considering few practical microphones for music recording can do more.  Few speakers can manage 40 khz on the other end.  

 

Better safe than sorry, so make productions record everything that exists in real world. And at least as much any speaker or headphone can capture. So about 100 kHz is enough.

 

About 90% of my listening is with headphones, others may have different weightings. With headphones I can hear more into details what DSP algorithms are doing. For speakers you can get super tweeters to help out if the normal tweeters roll off too early.

 

[esldude]

8 minutes ago, gmgraves said:

I saw 'The Shining' in the theatre when it first came out. When my then girlfriend and I exited the theatre, there was a line for the next showing. Both of us ran up and down that line telling people to save their money and avoid the movie like the plague! I honestly believe that it was one of the worse films I have ever had the misfortune to see! Luckily I don't remember much about it (Other than "Here's Johnny". Put it out of my mind, altogether, I guess). So, if your little rock-'n-roll taunt was based on something in that film, you'll forgive me for not picking up on it.  

I didn't run up and down the line telling people don't go, but had the same experience.  My girl at the time insisted we go see it. She agreed it was horrible.  We saw some people we knew for the next showing and did tell them see something else.  I've never understood the accolades for that film.  It was a waste of some fine actors and actresses. 

[esldude]

10 minutes ago, Miska said:

 

Better safe than sorry, so make productions record everything that exists in real world. And at least as much any speaker or headphone can capture. So about 100 kHz is enough.

 

About 90% of my listening is with headphones, others may have different weightings. With headphones I can hear more into details what DSP algorithms are doing. For speakers you can get super tweeters to help out if the normal tweeters roll off too early.

 

I only listen to headphones when I have no other choice.  

 

I find it hard to believe the difference in an end to end recording with 40 khz bandwidth would be heard by humans as even 1% different than one with 100 khz bandwidth. 

 

 

[Miska]

2 minutes ago, esldude said:

I only listen to headphones when I have no other choice.  

 

I find it hard to believe the difference in an end to end recording with 40 khz bandwidth would be heard by humans as even 1% different than one with 100 khz bandwidth. 

 

Do you have any particular reason to artificially limit the bandwidth? I don't find any good reason to do so.

 

I remember lot of people saying that nobody can see difference between SD and 1080p video. And then that nobody can see difference between 1080p and 4K video. And that nobody can hear difference between 128 kbps MP3 and RedBook. Oh yeah.

 

[Ralf11]

Let's all do a square wave decomposition and see what we get with limits at 30 kHz, 60 kHz, and 100 kHz

 

?

[esldude]

Just now, Miska said:

 

Do you have any particular reason to artificially limit the bandwidth? I don't find any good reason to do so.

 

Seriously?

 

The expense and difficulty of getting speakers to do 100 khz.  The data increase for those bandwidths.  The chance of other distortions and problems getting into the system by expanding bandwidth.  The limitations and expense of microphones that will do 100 khz bandwidth.  

 

At 100 khz air absorbs sound at about 100 db/ 100 ft.  So recording end and speaker end many/most recordings would have rolled off 100 khz by 30 db or more.  Our ears are put together to respond fairly well to around 15 khz.  The perception onto 20 khz is diminished and simply a side effect because it isn't cut off sharply.  Like many microphones are good to about 20 khz, but have a diminished response into the mid 30 khz range.  

[esldude]

11 minutes ago, Miska said:

 

Do you have any particular reason to artificially limit the bandwidth? I don't find any good reason to do so.

 

I remember lot of people saying that nobody can see difference between SD and 1080p video. And then that nobody can see difference between 1080p and 4K video. And that nobody can hear difference between 128 kbps MP3 and RedBook. Oh yeah.

 

I've never, ever heard anyone say there is no visible difference in SD or 1080.  Nor 1080 and 4k.  The difference is more obvious than 128 kbps MP3 and redbook. 

 

The people who developed MP3 didn't claim it was transparent.  There are those who claim highrate MP3 is close to transparent.  

[Miska]

3 minutes ago, esldude said:

Seriously?

 

The expense and difficulty of getting speakers to do 100 khz.  The data increase for those bandwidths.  The chance of other distortions and problems getting into the system by expanding bandwidth.  The limitations and expense of microphones that will do 100 khz bandwidth.  

 

Yes, seriously. You don't need to worry about speakers, just worry about the recording chain. Recording chain needs to be at least as good as any playback chain that exists out there.

 

Developing such systems is just work, the kind of stuff I do every day. 100 kHz is not that much in the end.

 

6 minutes ago, esldude said:

At 100 khz air absorbs sound at about 100 db/ 100 ft.  So recording end and speaker end many/most recordings would have rolled off 100 khz by 30 db or more.  Our ears are put together to respond fairly well to around 15 khz.  The perception onto 20 khz is diminished and simply a side effect because it isn't cut off sharply.  Like many microphones are good to about 20 khz, but have a diminished response into the mid 30 khz range.

 

When I play instrument, my hand-to-ear distance is about 0.5m. There's not that much attenuation in that distance. I put microphone on same distance as my ears and then listen the signal from microphone through headphones. It must sound identical to the direct sound. It is pretty simple.