Digitally-recorded albums on vinyl?

[infinitesymphony]

If an album is produced entirely in the digital realm, is there a reason to purchase it on vinyl?

For example, Radiohead's albums. The last handful have been produced for the most part using digital technology. Some albums may have been mastered to tape for a warmer sound (and I know that OK Computer and earlier albums were mostly recorded to tape), but obviously most of the effects and programming on an album like Kid A were digitally-created.

 

[adhoc]

Quote:

Originally Posted by infinitesymphony /img/forum/go_quote.gif If an album is produced entirely in the digital realm, is there a reason to purchase it on vinyl? For example, Radiohead's albums. The last handful have been produced for the most part using digital technology. Some albums may have been mastered to tape for a warmer sound (and I know that OK Computer and earlier albums were mostly recorded to tape), but obviously most of the effects and programming on an album like Kid A were digitally-created.

Almost none at all. However, sometimes the LP can exhibit less compression than the CD and thus be objectively "better". The CDs were cut so hot that if transferred to LPs they would almost certainly skip.

This was the case for say, Depeche Mode's Playing The Angel (certainly) or any of the Coldplay albums (I... think).

 

[memepool]

Quote:

Originally Posted by infinitesymphony /img/forum/go_quote.gif If an album is produced entirely in the digital realm, is there a reason to purchase it on vinyl? For example, Radiohead's albums. The last handful have been produced for the most part using digital technology. Some albums may have been mastered to tape for a warmer sound (and I know that OK Computer and earlier albums were mostly recorded to tape), but obviously most of the effects and programming on an album like Kid A were digitally-created.

It's usually much more complicated than everything being analogue or digital these days though isn't it? When you have rock bands like U2 and even Country singers using digital delays even if they record acoustically to tape the sound will have been originated digitally or been sampled and played back somewhere in the chain. It depends how far do you want to take it.

Recording to analogue tape adds compression which will never be able to be replicated inside a plugin. There are just too many variables in a mechanical electromagnetic chemical process like that which varies from one machine to the next, one stock to the next.

Unless the record is pressed from a CD originated entirely inside Cubase or something with VST instruments and no samples, it's going to have gone through some analogue processing and this will alter the sound, for better or worse depending on the skill of the producer/engineer/musician/artist and quality and type of equipment used. Even duplicating the same file via different audio codecs seems to degrade it somewhat after a while so that it comes to lack sparkle somehow.

As close as digital editing allows us to come it's never going to be an exact science no matter what the theory may say and obviously the skills of the people involved will ultimately shape the results.

 

[infinitesymphony]

Good points... There are still many analog and tube-based stages in recording (ex. guitar amps, tube compressors, tape machines, tube mic preamps). But, if the album isn't mastered to tape, it would seem like a digital format would be the closest representation to how the album sounded in the studio.

But I suppose that even if the vinyl is cut using a close analog approximation of the digital original, the fact that vinyl playback itself is a different medium will mean that the music sounds different--different elements of the mix and frequency spectrum will be emphasized.

Someone in another thread brought up an interesting point. Most digitally-recorded albums are currently recorded at 24-bit/48kHz or higher, but very few are released in a high-res format. Vinyl might closer-approximate the original signal in this case.

 

[Jigglybootch]

Quote:

Originally Posted by infinitesymphony /img/forum/go_quote.gif Someone in another thread brought up an interesting point. Most digitally-recorded albums are currently recorded at 24-bit/48kHz or higher, but very few are released in a high-res format. Vinyl might closer-approximate the original signal in this case.

This is really the only case I see for releasing digital recordings on vinyl. One could cut the vinyl from the original 24/192 master, rather than converting it to 16/44.1, which would sound better in theory.

 

[hciman77]

Quote:

Originally Posted by Jigglybootch /img/forum/go_quote.gif This is really the only case I see for releasing digital recordings on vinyl. One could cut the vinyl from the original 24/192 master, rather than converting it to 16/44.1, which would sound better in theory.

That only works if you can transparently capture the full range of the master onto Vinyl and in the case of 24/192 you cant do this, you cannot get a dynamic range of 120db onto vinyl, you cant get the 96db that you get with 16/44.1. You can get 75db - 80db which is good enough for all practical purposes, and apart from some (very rare) classical recordings as much as is ever on a CD but if you are using the full potential of a 24/192 recording medium then you will lose some of it going to vinyl. Given that unless you live in an anechoic chamber or Zabriskie Point your background noise level is going to be 35 - 40db a 75db DR lets you go from barely audible (above noise) to deafeningly loud anyway and a 120db DR looks a lot like overkill. 120db + 35db = hearing damage

http://www.dangerousdecibels.org/hearingloss.cfm

 

[bordins]

You may check out a thread at Steve Hoffman's forum below.

Does vinyl suffer from low signal-to-noise ratio ?

 

[infinitesymphony]

I don't think that a 24/192 signal needs to be captured "transparently" onto vinyl to sound better than a CD.

High-res recordings offer better signal accuracy, extended frequency response, and improved filtering during the D/A process. Yes, 24-bit allows for greater dynamic range, but that doesn't mean you need a song with 144 dB relative peaks to hear the difference. Having a higher bit-depth means that there are more values from which to choose. At 16-bit, there are 65,536 possible amplitudes, while at 24-bit, there are 16,777,216 possible amplitudes. Similarly, there are 192,000 samples/sec in a 192 kHz recording, which means that it will be more signal accurate with less guess-work than with a 44.1 kHz recording.

All of these attributes combine to make high-res formats more accurate to the original source without requiring unrealistic dynamic range or extended frequency response that's outside the range of human hearing. Still, some people say that vinyl sounds better than high resolution formats--maybe it's because maximum dynamic range isn't as important as microdynamics or filterless frequency response.

 

[hciman77]

Quote:

Originally Posted by infinitesymphony /img/forum/go_quote.gif Having a higher bit-depth means that there are more values from which to choose. At 16-bit, there are 65,536 possible amplitudes, while at 24-bit, there are 16,777,216 possible amplitudes.

Yes, but the finite limits of vinyl mean that this extra resolution in the recording isnt carried over to the pressing, the cutting head just isnt capable of that level of resolution, neither is the physical medium. A system with a DR of 80db is capable of about 8k unambiguous levels, that is it. DR and resolution are inextricably linked, if vinyl was capable of better than 80db then it would have higher resolution than it does, you cannot cram more information onto it just because you start with a higher resolution source. When you transfer from 192/24 (144db) to vinyl (~80db on a good day) you are by necessity compressing the range of the signal , otherwise how else do you get 144db into 80db. When it is compressed down all the extra resolution is lost and the "levels" are effectively aggregated.

Quote:

Similarly, there are 192,000 samples/sec in a 192 kHz recording, which means that it will be more signal accurate with less guess-work than with a 44.1 kHz recording.

If a signal is confined to a bandwidth of 20 kHz and the dynamic
range between the abient noise floor and the loudest signal is
80 dB, say, a 192 kHz 24 bit sampling will not capture
more information about that signal than a 44.1 kHz 16 bit system because there isnt any more information to capture. The higher sampling gets you higher bandwidth but over 20 - 20k it isnt relevant in information terms.


I am not knocking vinyl btw, I have been actively researching TT rigs for myself but within the realm of the physically achievable 192/24 onto LP doesnt buy you any advantages.

 

[hciman77]

Quote:

Originally Posted by bordins /img/forum/go_quote.gif You may check out a thread at Steve Hoffman's forum below. Does vinyl suffer from low signal-to-noise ratio ?

Thanks, it was an interesting read.

 

[infinitesymphony]

Quote:

Originally Posted by hciman77 /img/forum/go_quote.gif Yes, but the finite limits of vinyl mean that this extra resolution in the recording isnt carried over to the pressing, the cutting head just isnt capable of that level of resolution, neither is the physical medium. A system with a DR of 80db is capable of about 8k unambiguous levels, that is it.

Hmm... That's a good point. I hadn't thought of dynamic range and resolution basically being the same thing. I figured that since most non-classical recordings only use about 20 dB of dynamic range, the resolution of those 20 dB would be the most important factor.

Quote:

Originally Posted by hciman77 If a signal is confined to a bandwidth of 20 kHz and the dynamic range between the abient noise floor and the loudest signal is 80 dB, say, a 192 kHz 24 bit sampling will not capture more information about that signal than a 44.1 kHz 16 bit system because there isnt any more information to capture. The higher sampling gets you higher bandwidth but over 20 - 20k it isnt relevant in information terms.

Why would a signal be confined to a 20 kHz bandwidth? Also, a higher sampling rate will at least allow the low-pass filter to be moved away from the audible range of hearing, which should be a noticeable improvement. It's true that humans can't hear well beyond 20 kHz, but there are some studies that suggest frequencies above the range of hearing can still affect aural perception (though I haven't seen any of those studies firsthand

). Plus, it's just nice to know that they're there.

 

[hciman77]

Quote:

Originally Posted by infinitesymphony /img/forum/go_quote.gif Why would a signal be confined to a 20 kHz bandwidth? Also, a higher sampling rate will at least allow the low-pass filter to be moved away from the audible range of hearing, which should be a noticeable improvement. It's true that humans can't hear well beyond 20 kHz, but there are some studies that suggest frequencies above the range of hearing can still affect aural perception (though I haven't seen any of those studies firsthand ). Plus, it's just nice to know that they're there.

Fair points. As for 20K+, sigh, maybe 20 years ago I would have cared, these days my hearing certainly doesnt go above 16K

There were some interesting AES studies about the effect of Low pass filters back in 1983, sadly I cant recall the references exactly, I will try and hunt them out, however I think that they used audio professionals who may have been in their 30s thus making their perceptions of 20k filters er moot