[theaudiologist]

my question is when artists release records in vinyl do they always record in 24 bit and release it in 24 bit quality?

a couple of other questions:

1. do 16 and 24 bit make a difference in electronic music?
2.if an album is released in both cd and vinyl formats at the same time are they going to sound the same?

 

[kraken2109]

I'm not sure I understand your question. Professional recording is usually done in at least 24bit. Vinyls are analogue and therefore don't have a bit depth or sampling rate.
 
1. I suggest you give the thread on 24bit on this board a read.
2. They won't sound the same because vinyl is analogue and therefore will never be a 'perfect' copy of the original master. Two vinyls technically won't sound the same either.

 

[Speedskater]

Most vinyl has the equivalent resolution  of a 14 or 15 bit digital system.

 

[Steve Eddy]

Most vinyl has the equivalent resolution  of a 14 or 15 bit digital system.

Where are you getting that from? I'm not aware of any vinyl system that's capable of delivering a dynamic range near 90dB.

se

 

[Speedskater]

I was being generous. How about 12 or 13 bit?

 

[Steve Eddy]

I was being generous. How about 12 or 13 bit?

That's a bit closer.

se

 

[jcx]

cutting masters in vinyl production are deliberately different, tweaked for the media, limitations - won't sound the same as used to make the CD if you could get the 24/96 used to cut the stamping master
 
 
vinyl noise varies strongly with frequency, RIAA EQ is an attempt to better match our hearing thresholds to the record playback noise - exploits our poor low frequency sensitivity, highest at few kHz
 
80+ dB may be possible at mid frequencies with vinyl
 
vinyl has other requirements too - LP are mixed to mono at low bass frequencies to pack the grooves closer
 
there is cartridge tracking limit,  giving a signal slew rate limit as low as 5 kHz for most records - yes you can have higher frequencies - but the peak value you can cut, highest playback level drops proportionately above there
 
tone arm resonance, record warp, centering error and room noise give very visible low end noise to vinyl playback
 
 
most CD with modern close micing are mixed/produced too hot for the Red Book pre-emphasis - you  never see it used today
 
modern shaped dithers give CD better than 96 dB perceptual weighted S/N

 

[Don Hills]

... vinyl has other requirements too - LP are mixed to mono at low bass frequencies to pack the grooves closer ...

 
Your list is excellent, there's just one small quibble that I quoted above.
The mix to mono is to prevent excessive vertical excursion, not excessive horizontal excursion. Too much out of phase bass may result in the groove becoming too shallow on peaks, allowing the stylus to slip out of the groove or even be catapulted out. I suspect you know that, I'm just mentioning it in case someone else didn't.

 

[dclaz]

On point 2... 
 
The thing is that vinyls can sound better than the CD counterpart due to better mastering. CD's are often mastered in such a way that they lose fidelty, see http://en.wikipedia.org/wiki/Loudness_war 
 
I guess most vinyl rips are 96kHz/24bit to ensure little is lost in the analogue to digital conversion.

 

[wakibaki]

I believe Celemony's Capstan either works at or outputs 2496 by default. While it is primarily targeted at tape, I believe it can be used to process vinyl rips.

Again, it is a commercial application, intended for use professional production.

Like so much software, however, now that the cat's out of the bag, the functionality can't be hard to mimic.

Capstan is €3790.

I would expect to see post production software like this to be used increasingly in 'domestic' systems for the elimination of wow and flutter. You could probably do this direct off a turntable with a short delay. A drastic improvement over unprocessed vinyl sound (at least in technical terms) is probably possible, but such a system would almost certainly be rejected by vinyl enthusiasts.

w

 

[jcx]

Capstan may be good with input that matches its algorithm's built in models, assumptions about the time errors and the music tonal content, instruments note's pitch - but I bet there are instances where it is going to fail - like the "killer samples" found for every perceptual encoder
 
they take a swipe a Plangent Processes in their ad copy - but having a tape bias oscillator frequency to use a time reference does make it more robust when it can be used