I think this should be the right place for this link. It's one of the most well-written articles I've read on audio quality in general and digital audio specifically. The references in this article provide good and detailed read on the subjects as well. Well worth it.
24/192 Music Downloads are Very Silly Indeed
24/192 Music Downloads are Very Silly Indeed
This paper by Julian Dunn provides some of the reasons why higher sample rates may help especially given that not all processing is done using the best available techniques. He also discusses some of the ways processing can be improved. Whilst this paper covers 96k rather than 192k it is basically the same concept.
http://www.nanophon.com/audio/antialia.pdf
I am unclear why the author suggests 16 bits is enough as whilst you can hear signals well below the noise floor of a 16 bit signal so signal linearity can be encoded, You can also hear the hiss of the noise floor and this is a distraction to the music. Shaped dither that uses the behaviour of the minimum audible field to mask this hiss is quite processor intensive, prone to exposure by post processing and because it pushes alot of energy into the HF region can heat up the tweeter causing thermal distortion.
This paper has many good points but I think it may be supporting the 44k 16 bit argument rather to strongly. I think this standard was developed as a good compromise between high quality and storage capacity. I am not convinced it is a perfect match for the human auditory system, as seems to be proposed in this paper.
http://www.nanophon.com/audio/antialia.pdf
I am unclear why the author suggests 16 bits is enough as whilst you can hear signals well below the noise floor of a 16 bit signal so signal linearity can be encoded, You can also hear the hiss of the noise floor and this is a distraction to the music. Shaped dither that uses the behaviour of the minimum audible field to mask this hiss is quite processor intensive, prone to exposure by post processing and because it pushes alot of energy into the HF region can heat up the tweeter causing thermal distortion.
This paper has many good points but I think it may be supporting the 44k 16 bit argument rather to strongly. I think this standard was developed as a good compromise between high quality and storage capacity. I am not convinced it is a perfect match for the human auditory system, as seems to be proposed in this paper.
I did find his point on 16/24bit a bit confusing, to be fair. He does argue against his case here in two places, the noise floor is mentioned. Certainly, in modern times the storage wasted by using 24 bit won't really scare anyone.
But as a whole I like the way he presents it and I thought it'd be nice to see what people here have to say.
But as a whole I like the way he presents it and I thought it'd be nice to see what people here have to say.
- Status
- Not open for further replies.