I think this should be a sticky.Avoid the hype. https://people.xiph.org/~xiphmont/demo/neil-young.html#toc_intro
Unfortunately this article is a wild mix of fact an fiction.
While one might agree that 192kHz sampling rate is overkill,
the assumptions about the human hearing are result of the typical engineers
view of the ear brain combination as a simple microphone.
The author should have consulted someone with a medical or biological
background before drawing oversimplified conclusions.
Just a few points (there are much more):
20 - 20kHz: This is for pure sine static tones, there is for example
- bone conduction which exceeds 20kHz
- Intermodulation of sound sources (with the IM products in the audible range)
which varies with distance and add spatial clues.
(If you record two sound sources close miked, bandwidth limited to 20 kHz and
then mix them in your stereo mix you completely loose that)
20 kHz brickwall filter: you never ever would use such a thing deliberately
in audio - pre ringing, post ringing, phase shifts ... why not move this out
of the audio band now that we can ? Any reason to live with the limitations
of 40 years ago ?
16bit is enough: If we want records which ignore the "loudness war" we probably
need a bit more. (maybe not 24 but 20 or so). A real life example:
Classical symphonies quite often have rather soft as well as loud movements,
the soft parts easily will be down -20 to -30 dB compared to full scale.
The loud parts might be too loud at home (imagine a real symphony orchestra
in your living room) with the side effect that you will have to turn the
volume up for the soft parts. Now these are recorded at a much lower level
and you are listening to 10 to 12bit ... not so good, right ?
For the typical pop production even 16bit are more than enough, though.
For a more balanced view I would recommend this paper:
http://www.lavryengineering.com/pdfs/lavry-white-paper-the_optimal_sample_rate_for_quality_audio.pdf
While one might agree that 192kHz sampling rate is overkill,
the assumptions about the human hearing are result of the typical engineers
view of the ear brain combination as a simple microphone.
The author should have consulted someone with a medical or biological
background before drawing oversimplified conclusions.
Just a few points (there are much more):
20 - 20kHz: This is for pure sine static tones, there is for example
- bone conduction which exceeds 20kHz
- Intermodulation of sound sources (with the IM products in the audible range)
which varies with distance and add spatial clues.
(If you record two sound sources close miked, bandwidth limited to 20 kHz and
then mix them in your stereo mix you completely loose that)
20 kHz brickwall filter: you never ever would use such a thing deliberately
in audio - pre ringing, post ringing, phase shifts ... why not move this out
of the audio band now that we can ? Any reason to live with the limitations
of 40 years ago ?
16bit is enough: If we want records which ignore the "loudness war" we probably
need a bit more. (maybe not 24 but 20 or so). A real life example:
Classical symphonies quite often have rather soft as well as loud movements,
the soft parts easily will be down -20 to -30 dB compared to full scale.
The loud parts might be too loud at home (imagine a real symphony orchestra
in your living room) with the side effect that you will have to turn the
volume up for the soft parts. Now these are recorded at a much lower level
and you are listening to 10 to 12bit ... not so good, right ?
For the typical pop production even 16bit are more than enough, though.
For a more balanced view I would recommend this paper:
http://www.lavryengineering.com/pdfs/lavry-white-paper-the_optimal_sample_rate_for_quality_audio.pdf
Yes. It is called "evidence-based".
Yes - has it ever been shown that that contributes to how we hear things?
Can you explain that one?
No reason to live with the limitations of 40 years ago - there are better filter designs these days that make 20 kHz "brickwall" filters well-behaved.
For recording, yes. Monty acknowledges that. Once you record and normalize the sound, 16 bits is more than enough for storage, distribution and reproduction.
Ah, a "balanced" paper from someone with a commercial interest, as opposed to Monty that isn't actually trying to sell anything.
You can't. The best we can do is trying to present facts to counteract the pseudoscience.
The most dramatic possible fidelity improvement for the cost comes from a good pair of headphones.
This is a fact? If so, where's the evidence? Looks like a claim to me.
This is a fact? If so, where's the evidence? Looks like a claim to me.
- Status
- Not open for further replies.