Common, people!... A simple low band pass filter makes the same without your marketing bs. The difference we can expect only near the highest frequencies. When we take a linear approximation, it brings a lot of higher harmonics. And when we cut them, a nice smooth upsampled signal remains.
From a signal processing view,. according to Nyquist, the ideal sampled sequence is a series of Dirac pulses, so the most accurate upsampling strategy is zero-padding and then pass through a digital low-pass filter (or equivalently convolution with a sinc pulse).
Yes and that is what EVERYONE does and it's NOT straight line approximation but creates a "curved" signal as it should. Nothing special... but since this is a High-End audio marketing video it obviously has to be bullshit ;-)
I actually understood this explanation of up-sampling. I will have to listen to it a few more times however , before I can give my own version of what it means ha ha! Thank you.
Hi Hi, it is the same method that is used from end of 90'-is from Philips and it is the best way how to screw original signal even more. 🤣🤣 There are giant problems by transition functions which nobody is measuring so everything seems fantastic with 120bd signal to ratio which from 20 to 20khz. That is why audiophile hates everything approximated by sine functions whit this does like MID sound. So now you get from original sound record a MIDI sound where everything is nice and smooth like sinus wave. Yes let's replace everything by sine waves. 😁
I have just purchased the CXN v2, and will be using Airplay to stream Apple Music... I am told that Airplay restricts the quality to 24 / 44.1. Is the then upsampled to 24/384 as your video suggests? Or would I achieve a better rate if I connected the Mac to the CXN by USB, rather than Airplay???
A good test would be to feed the atf algorithm with 1 & 2 bit samples and or samples wich were watered down due agressive noise filtering,, if see if it could magically could restore & improve it, samples recorded with those old nokia phones would be a great test for atf !,eventrough i don’t expect muc!!
the ATF is combine with ADC (AD1955 32Bits Analog Devices), well...But it`s not very clear about Upsampling ; Because the AD1955 maximum rate is 192Khz..your Upsampling is about twice of that, it's ok for me, but) You say is not "linear interpolation", it's a curve...Ok but twice more samples is not a curve for a 192Khz source signal, it's a straigtht line too (two points)
48KhZ is not use in studio for Audio: is primary use for TV (and DAT?). is the same think for 192KHz ok. But your answere is unconnected with the question about non linear interpolation ATF. For my music i use redbook 44.1Khz for the listening. and use 24Bits 96Khz +The DeltaSigma/ Dithering noise shaping technique for Signal processing ONLY . The filtering is about 22 KHz max on the output chain, so artefact is not audible for me! (because of natural bandwith of my hear!) Best regards
The explanation is incorrect. Trying to explain this in terms of diagrams with "smooth" curves doesn't make sense, unless those smooth curves happen to coincide very closely with the analogue input signal. There is no obvious reason why they should based on the hand waving approach shown here. The method used in the hardware might be OK - but must surely have to depend on filtering. As an explanation this is poor.
Common, people!... A simple low band pass filter makes the same without your marketing bs. The difference we can expect only near the highest frequencies. When we take a linear approximation, it brings a lot of higher harmonics. And when we cut them, a nice smooth upsampled signal remains.
From a signal processing view,. according to Nyquist, the ideal sampled sequence is a series of Dirac pulses, so the most accurate upsampling strategy is zero-padding and then pass through a digital low-pass filter (or equivalently convolution with a sinc pulse).
Yes and that is what EVERYONE does and it's NOT straight line approximation but creates a "curved" signal as it should. Nothing special... but since this is a High-End audio marketing video it obviously has to be bullshit ;-)
Does anyone else find it ironic that this video on high-fidelity audio has muffled audio? I had to crank the volume up to understand it.
This seems to be a wide spread issue with audio companies, their instructional videos all use 50c mics for some reason
I actually understood this explanation of up-sampling. I will have to listen to it a few more times however , before I can give my own version of what it means ha ha! Thank you.
When filming with this camera switching setup, have your actor briefly glance at the side camera now and then - otherwise it gets uncanny.
Hi Hi, it is the same method that is used from end of 90'-is from Philips and it is the best way how to screw original signal even more. 🤣🤣
There are giant problems by transition functions which nobody is measuring so everything seems fantastic with 120bd signal to ratio which from 20 to 20khz.
That is why audiophile hates everything approximated by sine functions whit this does like MID sound.
So now you get from original sound record a MIDI sound where everything is nice and smooth like sinus wave.
Yes let's replace everything by sine waves. 😁
I have just purchased the CXN v2, and will be using Airplay to stream Apple Music... I am told that Airplay restricts the quality to 24 / 44.1. Is the then upsampled to 24/384 as your video suggests? Or would I achieve a better rate if I connected the Mac to the CXN by USB, rather than Airplay???
A good test would be to feed the atf algorithm with 1 & 2 bit samples and or samples wich were watered down due agressive noise filtering,, if see if it could magically could restore & improve it, samples recorded with those old nokia phones would be a great test for atf !,eventrough i don’t expect muc!!
Blind testing>?
Is there any way to disable it?
the ATF is combine with ADC (AD1955 32Bits Analog Devices), well...But it`s not very clear about Upsampling ; Because the AD1955 maximum rate is 192Khz..your Upsampling is about twice of that, it's ok for me, but)
You say is not "linear interpolation", it's a curve...Ok but twice more samples is not a curve for a 192Khz source signal, it's a straigtht line too (two points)
48KhZ is not use in studio for Audio: is primary use for TV (and DAT?). is the same think for 192KHz ok. But your answere is unconnected with the question about non linear interpolation ATF.
For my music i use redbook 44.1Khz for the listening. and use 24Bits 96Khz +The DeltaSigma/ Dithering noise shaping technique for Signal processing ONLY .
The filtering is about 22 KHz max on the output chain, so artefact is not audible for me! (because of natural bandwith of my hear!)
Best regards
The explanation is incorrect. Trying to explain this in terms of diagrams with "smooth" curves doesn't make sense, unless those smooth curves happen to coincide very closely with the analogue input signal. There is no obvious reason why they should based on the hand waving approach shown here. The method used in the hardware might be OK - but must surely have to depend on filtering. As an explanation this is poor.
Uhm, this is bullshit, no one in their right mind do any kind of straight line upsampling. It would just sound really shitty!