I was aware of mp3 being lossy and flac being lossless, but I had thought that CDs used a lossy format of some kind. Which was apparently wrong. I’ll look more into it.
Alec at Technology Connections did a good series of videos on Youtube about it, and he’s always a hoot.
There is an argument made mostly by the “vinyl is best” audiophile crowd that ALL digital audio is lossy because of an upshot of the Nyquist-Shannon sampling theorem, a key principle in how digital audio works. In order to avoid distortion due to aliasing, the sample rate of a digital recording must be at least twice the bandwidth of the signal. There’s a problem where if the frequency of the audio is way higher than the sample rate, it could produce the same series of samples that a lower frequency sound might, so it might play back incorrectly.
In practice, this means any audio that is digitized must be band-pass filtered before sampling. We tend to recognize the range of human hearing to be from about 20 Hz to about 20,000 Hz, so digital audio is low pass filtered to remove any sound above 20,000 Hz before being recorded. A sample rate of 44,100 Hz is a little more than twice that maximum bandwidth to allow for some wiggle room. There are people who claim they can hear sounds above 20,000 Hz, but they had to draw the line somewhere.
A practical upshot of this is you can’t record an ultrasonic dog whistle to a CD; even if the microphone could pick it up it would be filtered out before sampling. But, when taking practical constraints such as the biological limits of human hearing and the inertia of the speaker cones into consideration, CD audio can perfectly replicate any sound humans can hear in 2-channel stereo, which is why it hasn’t really been improved on in terms of sound quality in 40 years.
Compression, whether lossless or lossy, is a whole other can of worms that red book audio has nothing to do with. It’s amazing they were able to produce consumer-grade electronics that could decode that amount of raw data in real time in 1980, especially given the robust error correction it has built in. The processing power required to decompress audio before decoding it just wasn’t available while the Red Book was being written, so they didn’t include compression in the standard.
We treat CDs as obsolete here in the Spotify era but they’re still technological marvels.
Thanks for the info and detailed response!
I was aware of mp3 being lossy and flac being lossless, but I had thought that CDs used a lossy format of some kind. Which was apparently wrong. I’ll look more into it.
Alec at Technology Connections did a good series of videos on Youtube about it, and he’s always a hoot.
There is an argument made mostly by the “vinyl is best” audiophile crowd that ALL digital audio is lossy because of an upshot of the Nyquist-Shannon sampling theorem, a key principle in how digital audio works. In order to avoid distortion due to aliasing, the sample rate of a digital recording must be at least twice the bandwidth of the signal. There’s a problem where if the frequency of the audio is way higher than the sample rate, it could produce the same series of samples that a lower frequency sound might, so it might play back incorrectly.
In practice, this means any audio that is digitized must be band-pass filtered before sampling. We tend to recognize the range of human hearing to be from about 20 Hz to about 20,000 Hz, so digital audio is low pass filtered to remove any sound above 20,000 Hz before being recorded. A sample rate of 44,100 Hz is a little more than twice that maximum bandwidth to allow for some wiggle room. There are people who claim they can hear sounds above 20,000 Hz, but they had to draw the line somewhere.
A practical upshot of this is you can’t record an ultrasonic dog whistle to a CD; even if the microphone could pick it up it would be filtered out before sampling. But, when taking practical constraints such as the biological limits of human hearing and the inertia of the speaker cones into consideration, CD audio can perfectly replicate any sound humans can hear in 2-channel stereo, which is why it hasn’t really been improved on in terms of sound quality in 40 years.
Compression, whether lossless or lossy, is a whole other can of worms that red book audio has nothing to do with. It’s amazing they were able to produce consumer-grade electronics that could decode that amount of raw data in real time in 1980, especially given the robust error correction it has built in. The processing power required to decompress audio before decoding it just wasn’t available while the Red Book was being written, so they didn’t include compression in the standard.
We treat CDs as obsolete here in the Spotify era but they’re still technological marvels.
Commentary: I would expect someone familiar with Technology Connections to ideally call it a Hoot for sure.