Comment by nerd_stuff | Hacker News Reader

Comment by nerd_stuff | Hacker News Reader

Parent Full thread

Dithering basically raises the noise-floor by trading harmonic distortions for random noise at the cost of the last bit of information (for the uninitiated dithering means you randomly flip the last bit to create a soft truncation.) It's "noise" which isn't technically harmonic distortion but that's kind of splitting hairs isn't it?

I have no patience for articles like this because instead of setting the record straight it just adds more pseudoscience to the mix. Their "experiment" doesn't prove what they claim it proves.

Is 24-bit useful for playback? Yes it is. Do you need it for properly mastered audio? Absolutely not. But there are 10,000 home studios across the country who have music files at 24-bits and a large chunk of those engineers don't know how to master audio. Is a raw 24-bit track better than a poorly mastered 16-bit track? It is to me.

Is 192kHz useful for anything? Probably not. Does it hurt anything? Maybe a little. I would expect a quality music player to go up to 96 anyhow so 192 isn't so bad. Similar reasons as before.

Does this reflect badly on the Pono player? Not really. Over-engineering isn't a bad thing in my book.

> It's "noise" which isn't technically harmonic distortion but that's kind of splitting hairs isn't it?

That's not splitting hairs. Harmonic distortion and uncorrelated noise are completely different from a mathematical stand point. They also sound audibly different to my ears.

Merely saying that "24-bit is useful" or "it is better to you" does not mean that you can actually hear the difference, and you include no empirical justifications for why you believe this to be true. Perhaps I could help you conduct a scientific experiment at home?

24 bits has more information. It's useful for mixing and in cases where things haven't been properly gain-staged, an engineer had an off day, a track hasn't been mastered, etc. Seems pretty basic to me.

It probably sounds basic to you because you still aren't doing the math, and you still aren't doing the experiments.

The whole point of 24-bit is to give you flexibility when the singer decides that he's going to whisper one verse from three feet away and then shove the mic down his throat while he screams the chorus. Or you can be sloppy when you set the preamp gain—just give yourself enough headroom, and you'll adjust the levels later. Notice that I said "later", not "never". Even the most amateurish home engineer with a hangover the size of Texas is going to adjust the levels of different tracks in the mix. Once you've done the mixdown, the levels are "reasonable", and you can listen at home on 16-bit system. You won't hear the difference.

No mastering, no compression. 16-bit audio is still enough, once you've mixed a song.

What math exactly would you like me to do? 16 bits gives a theoretical maximum range of 96dB while the human ear can hear over 130dB. Good headphones can handle over 100dB signal to noise.

Maybe you should save your grandiose lectures for somebody who doesn't know enough to see through it. There's no reason to assume levels will be at a "reasonable" level and there's no reason to think a high-quality portable music player should only play back 16 bits. That's absurd. Should consumer level music players stop at 16? Sure. I want better than that but if you want to listen to mp3s on your phone nobody's stopping you.

I'm left wondering how much of this anti-Pono talk is shilling on behalf of smartphone manufacturers. I need 24 bit playback in roughly the same way I need a car that does 0-60 in under 3.5 seconds - I like things that are over-engineered so I can geek out on how awesome they are. But thanks to all the killjoys we can't even get excited about the first high quality portable music player because some amateur sound engineers want to show off they once read a blog post on Shannon-Nyquist theory.

130 dB dynamic range is unreasonable. It's not like having a car that can go 0-60 in under 3.5 seconds, it's more like having a furnace that you can set to keep your house at 50°C. 130 dB SPL is downright bad for you. It will result in pain and permanent hearing loss. 130 dB SPL is very close to OSHA limits for instantaneous (not sustained) volumes, and it is physically painful to experience.

Can you elaborate? What makes you think that people are shilling for smartphone companies?

Reply on news.ycombinator.com