Finally, music we can buy in 768 khz sampling rates.

[BDWoody]

The first time we made a profit, we, after having paid the bands their royalties, used the money to buy a stronger computer in order to record in 352khz.

I guess my question would be why you felt that was what you wanted to do with your hard earned money?

What benefit do you feel you get from such a high sample rate? This isn't meant to be obnoxious, I'm simply interested in the thought process.

Thank you for generously sharing these files!

 

[BDWoody]

The reason for high resolution is that when we increase the sample rates the smoother the samples get in

Monty's video is a regular around here.

Does this match your understanding of sampling theory?

 

[voodooless]

The reason for high resolution is that when we increase the sample rates the smoother the samples get in the audible spectrum, much closer to analog.
View attachment 176900

Really ? If you bring us images with staircases, you've already lost the plot! That is not how digital audio (or digital sampling in general) works

Try making your staircase on an 18 kHz sine wave. Good luck! Strange enough, a DAC can perfectly reproduce it with only a 44.1 kHz sampling rate. Shannon really was a brilliant man!

@BDWoody beat me to it with the video I bet that video get's most views from ASR references by now

 

[sergeauckland]

It really does astound me that people in this industry, doing good work in recordings, should nevertheless have such a poor understanding of the technologies they're using. Staircases indeed!

S.

 

[BDWoody]

Strange enough, a DAC can perfectly reproduce it with only a 44.1 kHz sampling rate. Shannon really was a brilliant man!

To me, if there is any magic in audio, it is in sampling theory. As a math guy, it's some mind blowingly cool shit...

 

[danadam]

What benefit do you feel you get from such a high sample rate? This isn't meant to be obnoxious, I'm simply interested in the thought process.

I'm not sure you'll get any different answer than already given:

I have the impression that it's a bit rounder.

 

[Blumlein 88]

Oh my not the stair case mis-understanding.

Look you make really good recordings. We really do appreciate you taking part in the forum in good faith. But that is not how sampling works. It is an intuitively easy way to look at it, but incorrect. Most of us here likely think there is absolutely nothing gained beyond 96 khz, and some would say 48 khz is enough. Whatever may be gained by higher rates, smoother waveforms isn't what is happening.

Seriously watch the Monty video carefully. If you don't believe it, tell us why not.

 

[BDWoody]

I'm not sure you'll get any different answer than already given:

Well, they are here at ASR posting bravely...so I thought throwing a little theory at them might fill in some missing gaps. It isn't just end users who are confused...

 

[xaviescacs]

To me, if there is any magic in audio, it is in sampling theory. As a math guy, it's some mind blowingly cool shit...

I'm more or less a math guy too but without signal processing experience. I had to watch it twice and read a bit to convince myself. After this, all makes perfect sense and you can stop loosing your time.

 

[voodooless]

So, is nobody brave enough to ABX the samples I made?

 

[voodooless]

Try making your staircase on an 18 kHz sine wave. Good luck!

Oh wait.. is this supposed to be the argument FOR higher sample rates? Because if you can't staircase it, you need more steps?... Woops

 

[danadam]

One could argue that with all that high frequency noise, the resulting waveform is actually less smooth than it would be from 44.1 kHz file

 

[mansr]

To me, if there is any magic in audio, it is in sampling theory. As a math guy, it's some mind blowingly cool shit...

And yet the proof is so simple.

 

[mhardy6647]

... and, to extend the in-apt analogy, just think of how hard it would be to ascend a staircase with a very large number of steps which were all of itty-bitty height and depth.

 

[KSTR]

So, is nobody brave enough to ABX the samples I made?

Only few people will have 768k-capable DACs.
Foobar's ABX is not ideal when dealing with different sample rates as that may give cues (from the additional sample rate switching).
In a quick run I could not detect any differences.

I think it would have been best to filter the original only (with a proper sinc to 24kHz) and reduce bit depth to 16 (with dither) but save in 24bits so that the file format is exactly identical and no resampling is involved etc.

Maybe a compare between 192/24 and 44.1/16 (or 48/16) would be more attractive (but again with the above constraints wrt file format)?

 

[antcollinet]

Hi David
If we were in this business for the money, we would be better of making a different type of music.
Jazz is not very popular.
And giving away free samples does cost us.
Amazon are charging us per download.
Our "goal is to build a bridge (liaison) between the studio (engineer and musicians) and the people who love to listen to beautiful music using high quality audio equipment. Next to standard studio recording sessions they organize special recording sessions with a live audience" see About us
This is the reason we dare be on this forum, "building a bridge."
The first time we made a profit, we, after having paid the bands their royalties, used the money to buy a stronger computer in order to record in 352khz.
The 2nd time we invested the money in our first Josephson microphone.
3rd time we bought our Merging.
Last year we bought one more Josephson in order to experiment with surround and "one mic +"
recordings.
It is a passion for sound and jazz that drives us and unfortunately very much a
"don't give up your day job" situation.
I work as a musician and teach bass and jazz ensemble at MSA here in Amsterdam.
Frans works fulltime as an audio technician, partly for the Dutch radio making radio plays but also freelance recording all kinds of music. From classical to pop and music for children.
The reason for high resolution is that when we increase the sample rates the smoother the samples get in the audible spectrum, much closer to analog.
View attachment 176900
Greetings Peter

Ouch - I'm *really* surprised a recording studio doesn't have a better idea of how digital sampling works.**

The only thing higher sampling rates give you is the ability to up the bandwidth beyond 20KHz - so inaudible.

And reduce the quantisation noise floor from inaudible to really really inaudible. I see somone has already posted the requisite video upthread: you should watch it.


**(I'll admit, I didn't until I watched the video recently..... but then, I'm not a recording studio )

 

[Killingbeans]

The reason for high resolution is that when we increase the sample rates the smoother the samples get in the audible spectrum, much closer to analog.

Sorry, but no. Whoever told you that is full of s¤¤t. The only thing you get from higher sample rates is higher frequency content.

Those three pictures do not represent the analog output of a DAC. You need to go back and study the sampling theorem one more time.

 

[sarumbear]

The reason for high resolution is that when we increase the sample rates the smoother the samples get in the audible spectrum, much closer to analog.

This is as good an audiophool statement as it gets

Little knowledge is dangerous...

 

[Robin L]

If I owned a recording studio [I used to freelance, had all my own gear, recorded a lot of classical music concerts], I'd put the money into microphones and record at 24/96. Having wider bandwidth isn't going to do a bit of good unless the microphones capture those frequencies. And they don't, and we don't hear those frequencies anyway. What we hear are the various sound signatures of different microphones. That can [and does] have a profound effect on sound quality. Very few microphones get as high as 40khz and a lot of the most frequently used microphones roll the top off well before that. It's the sound below those ultrasonic tones that matters anyway, and there's lots of variation of sound among microphones.

These recordings sound great. I'm sure Sound Liaison chooses the best microphones they've got. The studio certainly looks appropriate for the kinds of music recorded there. But I'm pretty sure the 768khz sampling rate doesn't do anything but increase file size.

 

[xaviescacs]

The reason for high resolution is that when we increase the sample rates the smoother the samples get in the audible spectrum, much closer to analog.

There is this relation between the bandwidth of the analog signal and the sample rate necessary to perfectly reconstruct it. The sampling theorem applies to a bandwidth limited signal, which makes perfect sense because our hearing is bandwidth limited, that's incontrovertible. Reading the theorem in the opposite direction, if more samples are introduced then an analog signal with a wider bandwidth can be reconstructed, which means that we can construct an analog signal with frequencies above 20 kHz, but the resulting analog signal between 20 Hz and 20 kHz will remain the same, because it's already perfect, by the sampling theorem. Exactly what @Killingbeans is saying.

The only thing you get from higher sample rates is higher frequency content.

Intuitively, the Fourier transform of an arbitrary continuous function in R has an infinite number of terms, but how many harmonics or components are necessary to reconstruct a function that is bounded? A finite number of them. How many exactly? That's what the sampling theorem tells us. It's nothing mysterious really.

Note: I'm explaining this because it's the best way to learn something, please correct me if I'm wrong.