> CDs added additional frequencies into the high range, which people thought sounded crisp and new, but it never had the warmth of cassettes and vinyl.
Sure, there is an audible difference between digital audio and vintage analog equipment — but it has nothing to do with quality problems in digital audio. In fact, it’s exactly the opposite: vinyl and cassettes deliver low-quality, heavily distorted audio (with a dynamic range equivalent to that of 5 or 6 bit digital audio). This does create a distinctive “warm” tone, and a lot of people really like the sound (or the nostalgia) created by this distortion.
The article you linked does not demonstrate quality problems with digital audio; it is a bunch of nonsense intended to sell a product.
First of all, it compares a “digital square wave” with an “analog sine wave”, and remarks that “the digital signal does not follow the smooth flow of its analog compliment.” Of course they look different, they’re not the same wave! A square wave is “supposed” to look blocky, it’s not a digital vs. analog thing.
At least it mentions the Shannon-Nyquist theorem and gets one thing right: “there is no difference whatsoever between an analog sine wave and a digital one”
But then it totally blows it again with its comparison of a “digital” square wave to a “natural” one, observing that the “natural” square wave is more curved.
Again, this is nonsense because you’re not comparing the same wave! The “digital” image is of an ideal, mathematically pure square wave, whereas the “natural” image is bandlimited. The first mathematically pure wave cannot exist in the real world, because it would require infinite bandwidth — the voltage would have to change instantly.
In fact, a digital signal is bandlimited; it’s just a bad illustration. The digital square wave, when converted to analog, will not include any frequencies above half the sampling rate, making it match whatever our bandlimited analog signal is supposed to look like.
Then the article goes into full-on snake oil mode, trying to convince you that “technology that resolves that problem, by creating infinite phase, is currently available” if only you buy their “Sensorium™ LSV III Function Generators and Altitudinal Oscillators [that] utilize a polynomial transition region algorithm”. This is very much nonsense; by the Nyquist-Shannon sampling theorem, you can reproduce a bandlimited waveform exactly from a sampled signal, no matter how “complex” your waveform is.
> Then the article goes into full-on snake oil mode, trying to convince you that “technology that resolves that problem, by creating infinite phase, is currently available” if only you buy their “Sensorium™ LSV III Function Generators and Altitudinal Oscillators [that] utilize a polynomial transition region algorithm”.
Definitely my favorite part. It's complete gibberish, but because it uses audio nerd words (phase, oscillator, function generator) it sounds like it might not be.
Sure, there is an audible difference between digital audio and vintage analog equipment — but it has nothing to do with quality problems in digital audio. In fact, it’s exactly the opposite: vinyl and cassettes deliver low-quality, heavily distorted audio (with a dynamic range equivalent to that of 5 or 6 bit digital audio). This does create a distinctive “warm” tone, and a lot of people really like the sound (or the nostalgia) created by this distortion.
The article you linked does not demonstrate quality problems with digital audio; it is a bunch of nonsense intended to sell a product.
First of all, it compares a “digital square wave” with an “analog sine wave”, and remarks that “the digital signal does not follow the smooth flow of its analog compliment.” Of course they look different, they’re not the same wave! A square wave is “supposed” to look blocky, it’s not a digital vs. analog thing.
At least it mentions the Shannon-Nyquist theorem and gets one thing right: “there is no difference whatsoever between an analog sine wave and a digital one”
But then it totally blows it again with its comparison of a “digital” square wave to a “natural” one, observing that the “natural” square wave is more curved.
Again, this is nonsense because you’re not comparing the same wave! The “digital” image is of an ideal, mathematically pure square wave, whereas the “natural” image is bandlimited. The first mathematically pure wave cannot exist in the real world, because it would require infinite bandwidth — the voltage would have to change instantly.
In fact, a digital signal is bandlimited; it’s just a bad illustration. The digital square wave, when converted to analog, will not include any frequencies above half the sampling rate, making it match whatever our bandlimited analog signal is supposed to look like.
Then the article goes into full-on snake oil mode, trying to convince you that “technology that resolves that problem, by creating infinite phase, is currently available” if only you buy their “Sensorium™ LSV III Function Generators and Altitudinal Oscillators [that] utilize a polynomial transition region algorithm”. This is very much nonsense; by the Nyquist-Shannon sampling theorem, you can reproduce a bandlimited waveform exactly from a sampled signal, no matter how “complex” your waveform is.
For a more detailed overview about how and why digital audio works, I recommend this video: https://xiph.org/video/vid2.shtml