Hacker News new | past | comments | ask | show | jobs | submit login

Yes. He's complaining against dB with no reference, not against dB(A) for example. (Apart from the naming of some of them being silly)





But dB without reference makes sense in many many occasions. Either because the reference is implicit (not ideal, but we have many implicit assumptions in communication), or because it's genuinely a ratio. Attenuation, gain.

If you every find an "official" written document that uses dB not as attenuation/gain and is not specifying the reference (at least in a footnote), it's written either by idiots or for idiots, or both.


No it doesn't. It's always bad for the actual unit to be implicit.

The unit of a gain/attenuation is [1]. There is no implicit unit in that case.

dB(A) is a weighting. It’s not a reference and it’s not units. I think some of the confusion here comes from people not actually understanding units.

A-weighting describes how different frequencies are summed up. It’s like saying “RMS”. RMS is not units, A-weighting is not units. You can apply A weighting to voltage, digital signals, or audio. They all have different units but can all be A-weighted.

You could invent a new unit for A-weighted audio, but you would need several.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: