Good one about formaldehyde! But wait, what happen when i have to make overdub on a Studer 827? What is the dust accumulating at feet of the playback head? Or the stylus grinding the end of the groove? Isn't it some part of the music being pulled out from the media?!
Isn't analog like putrefaction: first you kill the sound ( as you fix a live act) and it slowly degrade to a point it's just... noise?! LOL.
Isn't analog like putrefaction: first you kill the sound ( as you fix a live act) and it slowly degrade to a point it's just... noise?! LOL.
To the OP:
Your idea is not totally nuts but it have been applied for different target:
http://www.josephson.com/c700a.html
The differents capsule help modifying directivity when you are in doubt...
That said if you want to encode the x/y/z axis you'll need an array. There is no escape. The issue i see is: when in music do you need more than width and depth? I mean there is few instruments which move on vertical axis when played.
It is true for moving things around you: when have you had a musician performing circles around you during a live act?
Maybe at a Pink Floyd concert where a plane crashed on the main stage ( from other side of stadium) as a spectacular 'mise en scene'...
For cinema, videogames there is a real need of 'immersive audio', it can bring something. For music i still have to be convinced to invest 5000euros to listen to the rototoms behind me on 'the wall' ( or was it Dsotm?) in 5.1... the rest being only ambience mics... it doesn't brings anything to me justifying the expense.
Your idea is not totally nuts but it have been applied for different target:
http://www.josephson.com/c700a.html
The differents capsule help modifying directivity when you are in doubt...
That said if you want to encode the x/y/z axis you'll need an array. There is no escape. The issue i see is: when in music do you need more than width and depth? I mean there is few instruments which move on vertical axis when played.
It is true for moving things around you: when have you had a musician performing circles around you during a live act?
Maybe at a Pink Floyd concert where a plane crashed on the main stage ( from other side of stadium) as a spectacular 'mise en scene'...
For cinema, videogames there is a real need of 'immersive audio', it can bring something. For music i still have to be convinced to invest 5000euros to listen to the rototoms behind me on 'the wall' ( or was it Dsotm?) in 5.1... the rest being only ambience mics... it doesn't brings anything to me justifying the expense.
It would be very interesting and informative to find the reasons behind the decision to use a sampling rate of 44.1kHz. This is slightly more than 2x20kHz. The only reason I can think of using such a low sampling rate, is assuming that corrupting the amplitude of upper limit frequencies is acceptable or not noticeable for most people.
It originally comes from 1970's digital studio recorders.
According to the sampling theorem, you need twice the highest signal frequency of interest, and 20 kHz was considered to be the highest frequency of interest. As practical anti-aliasing and reconstruction filters have transition bands, in practice you need a bit extra, say 44 kHz.
Professional video recorders were used to do the recording, so the sample rate needed to be such that you could simply convert the digital signal into something that looks like a video signal. I don't know the details, but 44.1 kHz was very suitable for the recorders used in countries with the PAL system and 44.056 kHz for NTSC.
One of these two was later chosen for CD to ensure compatibility with the older video-recorder-based digital tape recorders. The difference in sample rate between 44.1 kHz and 44.056 kHz was small enough to also just copy 44.056 kHz data to CD.
According to the sampling theorem, you need twice the highest signal frequency of interest, and 20 kHz was considered to be the highest frequency of interest. As practical anti-aliasing and reconstruction filters have transition bands, in practice you need a bit extra, say 44 kHz.
Professional video recorders were used to do the recording, so the sample rate needed to be such that you could simply convert the digital signal into something that looks like a video signal. I don't know the details, but 44.1 kHz was very suitable for the recorders used in countries with the PAL system and 44.056 kHz for NTSC.
One of these two was later chosen for CD to ensure compatibility with the older video-recorder-based digital tape recorders. The difference in sample rate between 44.1 kHz and 44.056 kHz was small enough to also just copy 44.056 kHz data to CD.
Digital process limited by highest frequency analog signal and the dynamic range of the analog signal. If the frequency and the dynamic range of analog signal is unlimited, so we can not use digital process. Human hearing capability is limited. The frequency is limited and the dynamic range is limited. Analog process and digital process create distortion, phase shift, etc that change the original signal.
Some prefer certain distortion because it sound more pleasing, like filter application in camera. Some prefer higher noise because he think the noise is the original signal. But many of them failed in the blind listening test.
Some prefer certain distortion because it sound more pleasing, like filter application in camera. Some prefer higher noise because he think the noise is the original signal. But many of them failed in the blind listening test.
This Wikipedia article has some more information on recording digital audio on a video recorder:
https://en.m.wikipedia.org/wiki/PCM_adaptor
https://en.m.wikipedia.org/wiki/PCM_adaptor
I used the early Sony digital audio recorders that output to U-Matic video. Odd that I don’t remember the sample rate but highly likely it was 44.1.
Some early digital recorders ran at a sample rate of 50kHz, but that rate didn’t last long.
Some early digital recorders ran at a sample rate of 50kHz, but that rate didn’t last long.
Hi guys thanks the input I found the correct key word of the research is
So yeah if we view sound in term of colour there is definietly more than one way to record sound and do ADC and DAC. But I guess the first real step to do this is how do we mechanically transfer it to the something like reel to reel first before we can digitalize it "correctly"
And Enlish is not my first language hopefully my train of thought from the topic of Are our recording technology outdated to the idea chromesthesia is no confusing.
chromesthesia.
https://en.wikipedia.org/wiki/Chromesthesia. and more interesting if we look into the composer section in the wiki.So yeah if we view sound in term of colour there is definietly more than one way to record sound and do ADC and DAC. But I guess the first real step to do this is how do we mechanically transfer it to the something like reel to reel first before we can digitalize it "correctly"
And Enlish is not my first language hopefully my train of thought from the topic of Are our recording technology outdated to the idea chromesthesia is no confusing.
Last edited:
A colleague of mine has that with music. He's a good amateur pianist. I don't quite get the point, though.
For what it's worth, I have heard him complain about the way violins sound on CD. I don't remember whether that applied to all classical CDs or just most of them.
For what it's worth, I have heard him complain about the way violins sound on CD. I don't remember whether that applied to all classical CDs or just most of them.
Suppose an audio signal is fed into a frequency separator and every channel from the latter is sampled digitally. The outputs of the separator would be pure sinusoid bursts. Each channel digital output is also stored separately.
What happens if we assume a perfect synchronisation between the many frequencies when they are converted back into analogue and the many signals are fed into an adder? Does the combined/recreated signal resemble the original, or sampling individual frequencies breaks it? I should think, the original signal should be the result at the output of the adder.
I will try to analyse this mathematically.
What happens if we assume a perfect synchronisation between the many frequencies when they are converted back into analogue and the many signals are fed into an adder? Does the combined/recreated signal resemble the original, or sampling individual frequencies breaks it? I should think, the original signal should be the result at the output of the adder.
I will try to analyse this mathematically.
Schemes like that are often used, but normally with the filter bank on the digital side (keywords discrete wavelet transforms, subband coders, perfect reconstruction filter banks). With a high enough sample rate and correctly chosen filter transfers, you can in principle get perfect reconstruction. In fact there are schemes where the filter outputs get subsampled and any aliasing cancels again at reconstruction. I don't know the details of any of those methods.
Last edited:
As far as I know, it's used for lossy compression methods like MP3 (give the largest part of the bit budget to the subbands that are most audible) and for time-spectrum analyses (see how the spectrum changes over time).
Here’s a good intro to Discrete Wavelet Transforms: https://towardsdatascience.com/the-wavelet-transform-e9cfa85d7b34
It’s not just pure frequencies that get extracted(although you can do that) but also different set patterns that you might be interested in.
They‘re useful in audio classification and machine learning, audio compression, and audio effects processing, like source separation (separating voice from background noise, for example, to make mobile phones sound clearer)
It’s not just pure frequencies that get extracted(although you can do that) but also different set patterns that you might be interested in.
They‘re useful in audio classification and machine learning, audio compression, and audio effects processing, like source separation (separating voice from background noise, for example, to make mobile phones sound clearer)
I believe that this is not true and that DAC that costs a 1/10 of the stated figures now exists that do what is claimed above. One need to understand that there are listeners that are fixed on purchase price and just can belive or accept that this is now a reality. They cling to the "high-end" dogma and for some reason, try to keep that dogma alive. So beware of this if you feel discouraged by the above and feel that you need to spend a lot of money to create a top notch system in terms of sound - no you don't. Just build it wisely - speakers is what one need to focus on - electronics (DAC, amplifiers) is a "done deal" taking it's influence into a account on a system level view.
//
Which dac would you guys recommend for Ed?
Maybe he would like the sound of Holo May? Chord Hugo TT2?
Maybe he would like the sound of Holo May? Chord Hugo TT2?
IIRC, Sony on U-Matic was 44.056kHz in North America. This created some issues in transferring to redbook. The Soundstream digital recorder, used somewhat famously by Telarc in their earlier recordings, had a 50kHz sample rate.