OT: Quote from CNN describing digital audio

[ronwasserman]

This was from an article talking about how vinyl record sales are increasing. But there was this little comment about digital audio - "Digital recordings capture samples of sound and place them very close together as a complete package that sounds nearly identical to continuous sound to many people."

Perhaps it's the lack of sleep, but I have no idea what the reporter was trying to convey here. Anyone care to decode this statement?

Hope you're all doing well.

Ron

[Drew Mazurek]

Sounds like they are trying to describe sampling as it contrasts to the continuous nature of analog, that's all.

[froyo]

Hello. I agree with dmaruzek, it sounds to me like they were trying to describe the Nyquist theorem in layman's terms and failed miserably.

[PTUser NYC]

Yeah, they're describing discrete slices of audio like they are trigonometry or something - the area under a curve being poorly represented by a series of narrow rectangles. They don't consider the reconstruction filter at all.

[yelhahc]

..."sounds nearly identical to continuous sound to many people."

Somewhere there must be a few people swearing at their CDs for not making a continuous sound.....LOL

[Eric Lambert]

...it sounds to me like they were trying to describe the Nyquist theorem in layman's terms and failed miserably.

They weren't trying to get anywhere near something like Nyquist, just trying to describe to common folks how an acoustic wave is sliced up, documented, then put back together again. For someone like my father, the description was good enough. Keep the audience in mind. For the inquiring minds of engineers, yeah, it's probably an insult.

[PTUser NYC]

They weren't trying to get anywhere near something like Nyquist, just trying to describe to common folks how an acoustic wave is sliced up, documented, then put back together again. For someone like my father, the description was good enough.

I don't think so, because it leaves them with the idea that there is something missing between the slices, and there just isn't - the only thing missing is in the bandwidth, but I don't think you can hear that anyway.

I think its plain wrong, irresponsible, and perpetuates poor audio mythology.

[Postman]

I'd love to hear the report's take on mp3.

[Barnabas]

I don't think so, because it leaves them with the idea that there is something missing between the slices, and there just isn't - the only thing missing is in the bandwidth, but I don't think you can hear that anyway.

If there was not something missing between the samples, then why do we have higher sampling rates of 96 and 192? Aren't those sampling rates designed to fill in the blanks between the 44.1K samples?

[Razorfish]

If there was not something missing between the samples, then why do we have higher sampling rates of 96 and 192? Aren't those sampling rates designed to fill in the blanks between the 44.1K samples?

That ain't the way it works at all. It isn't like a movie, with discreet 'pictures'. The DA process, if I am correct, is a smooth, continuous stream. Like somebody said, the issue is bandwidth. Most mics down't even capture sound about 20k (if you're lucky--most dynamics and ribbons top out between 14k and 18k I think). There was a good article in Sound on Sound by Hugh Robjohns about this a few months back. Basically, in order to capture a sound of 20,000 Hz, you have to double the sample rate (has to be at least 40KHz), otherwise you get artifacts--it works like a mirrow. That is the Nyquist theorem. If you try to sample 30 KHz, you will get artifacts that sound at 30KHz - 20KHz = 10 KHz. That is audible, and is why all modern digital converters use hard low pass filters of some sort. Dan Lavry I think was commenting the other day on Gearslutz that mics that pick up sound up to 100k are actually more of a work hazard than a benefit. The gist is, if you are sampling sine waves of 100 Hz, you only need to work at a 200 Hz sample rate. Does that make sense?