Making 24-bit/96Khz DVD-A Recordings from Vinyl
Dec 20, 2007 at 6:41 PM Post #31 of 45

bigshot

Headphoneus Supremus
Joined
Nov 16, 2004
Posts
21,955
Likes
3,924
Location
Hollywood USA
I'd like to comment on the charts posted at the beginning of this thread, but I'd like to get an answer on their context. ADD would you please see post 27?

Thanks
Steve
 
Dec 20, 2007 at 6:45 PM Post #32 of 45

slwiser

Headphoneus Supremus
Joined
Jun 23, 2001
Posts
6,317
Likes
24
Quote:

Originally Posted by bigshot /img/forum/go_quote.gif
I'd like to comment on the charts posted at the beginning of this thread, but I'd like to get an answer on their context. ADD would you please see post 27?

Thanks
Steve



To me the chart makes no sense but as you said the context is all important. My interpretation would be that it has to be a slice in the time domain, maybe 10 ms or so reflection (sampling) on the music as it was playing in front of him and he captured that image; but I really don't know.
 
Dec 20, 2007 at 9:12 PM Post #33 of 45

Simon Sez

New Head-Fier
Joined
Nov 27, 2007
Posts
35
Likes
0
If I took a record (vinyl) and recorded it @ 24bit 48khz in pro tools would it sound better than the CD of the same name. I realize that higher conversion = better sound, but I'm not really giving it a better sample rate am I? Also if it did sound better, how much better would it sound?
 
Dec 20, 2007 at 10:58 PM Post #34 of 45

hciman77

Headphoneus Supremus
Joined
Feb 25, 2004
Posts
2,890
Likes
12
Quote:

Originally Posted by Simon Sez /img/forum/go_quote.gif
If I took a record (vinyl) and recorded it @ 24bit 48khz in pro tools would it sound better than the CD of the same name. I realize that higher conversion = better sound, but I'm not really giving it a better sample rate am I? Also if it did sound better, how much better would it sound?


No. It would probably sound different because most likely the mastering would not be the same. I recently compared a set of wav files, one was from a rip of an LP, the other was from a rip of the SACD - same work/same artist. The two files were quite different. Neither had a dynamic range that was challenging - the SACD rip had a DR of 80db and the LP rip had a DR of 64db, also the SACD rip was on average 1db louder, however more telling was that the energy levels at different frequencies , could be very different. The SACD was much quieter in some parts and much louder in some parts and vice versa. In general over most of the frequency range the quiet parts were quieter and the loud parts louder on the SACD, but the higher frequencies which were at a low level (-65db or so) were louder on the SACD by between 4 and 6db than on the LP. Here are the curves from 13K to 22K. Do they look even similar ?

Fig1.gif


Fig2.gif



Unless you have identical source material, comparisons are largely uninformative.
 
Dec 21, 2007 at 1:57 AM Post #35 of 45

james__bean

1000+ Head-Fier
Joined
Feb 13, 2004
Posts
1,029
Likes
22
I make my own vinyl rips all the time. People always talk about how you can't hear above 20kHz, but I've directly compared a rip done in 24/48 to the same rip converted to 16/44 with foobar2000's ABX test. I picked out the difference like 95% of the time, which leaves the probability that I can't hear the difference like .01%. Its easy when you listen to the symbols.
 
Dec 21, 2007 at 9:01 AM Post #36 of 45

ADD

500+ Head-Fier
Joined
Jun 29, 2007
Posts
922
Likes
26
Quote:

Originally Posted by slwiser /img/forum/go_quote.gif
To me the chart makes no sense but as you said the context is all important. My interpretation would be that it has to be a slice in the time domain, maybe 10 ms or so reflection


The charts are actually a fair chunk of time. I don't have the snippet anymore (I copy pasted it out of the entire side), but it is around the last 30 seconds or so of the LP side. You can see the time increments running across the very top.

The LP was transcribed using my X-Fi card in ASIO mode at 24-96. The X axis is basically time going by. The Y axis represents the frequencies in hertz contained in the recording at the point in time on the corresponding X axis.

To answer regarding the actual recording, it is a 180 gram LP Speaker's Corner reissue of the Mercury Living Presence recording SR90329 (Liszt Piano Concerto No. 1, recorded in 1962 on 3-track 35mm full-coat film with a modified Westrex film recorder and Telefunken / Schoeps 201 mic capsules).

Each time you see the graph go to 30 - 40 Khz, that is the sound of the cymbals, which as you can see are crashed quite regularly at the end of piece. The redbook CD series of Mercury Living Presence were reissued back in the 90s.

The actua colour of the graph data represents dB output relative to the 0 dB reference level. So as one expects, the higher the harmonic frequencies, the lower the output in general. Of course, being classical music, the RMS level of the entire side is pretty low to begin with - around -25 dB below reference, otherwise there is of course no headroom for the peaks.

Although this is a specific example, it is a pretty typical result I am getting from these reissues with my Project Debut Phono and Ortofon 2M Red. I have a few Classic Records reissues too, and they seem the show similar headroom, though the bulk of the material on all my recordings sits in the up-to-25 khz range.
 
Dec 21, 2007 at 6:32 PM Post #37 of 45

bigshot

Headphoneus Supremus
Joined
Nov 16, 2004
Posts
21,955
Likes
3,924
Location
Hollywood USA
OK, that explains why it doesn't look typical... It's an audiophile recording on an audiophile pressing compared to a CD release that was notorious for being particularly bad. This is a one in ten thousand sort of example. Most recordings wouldn't even have those frequencies.

However, even with that upper frequency extension in the master, I would bet that no one could hear a difference if it was properly dithered down to 16/44.1. Those frequencies are inaudible, especially when there are so many other frequencies clashing about in a cymbal crash.

Janis is good in this, but Richter is the one who owns this.

See ya
Steve
 
Dec 21, 2007 at 6:34 PM Post #38 of 45

jp11801

aka JP-nums or JP-numbers
Lead Organizer for Can Jam '09
Joined
Nov 11, 2004
Posts
4,286
Likes
248
Location
SoFlo
one thing that rarely gets talked about during the needle drop discussion is the quality of essentially consumer grade A>D convertors. I would have to imagine that the A>D convertors in the M-audio device or alesis masterlink that I owned and loved is not even close to up to par with the SOTA convertors a studio might use??
A general rule for me when buying software is if it is new unless there is a compelling reason to buy the vinyl like a complete different master done by an engineer I trust I go cd. A good example would be RHCP last LP was done by Steve Hoffman and KEvin Gray at RTI this is a no brainer vinyl purchase but a straight digital transfer to analog tape then cut to vinyl I would probably pass.
I always go vinyl for classic rock and most jazz 1950-1975ish when I can.

Finally to answer the OPs question I did drop the vinyl to 24/96 for the simple reason that I could then dropped it down the 16/44 for redbook burning. This not make sense to the science minded but it was what was suggested and the time factor did not make a difference to me.
 
Jan 5, 2008 at 2:00 AM Post #39 of 45

See Why Audio

New Head-Fier
Joined
Jan 4, 2008
Posts
1
Likes
0
Hi all,

Fascinating to see what people really think about hi-res digital V CD. So many facts and even more fiction :p
Sorry, it’s a big, long post and I hope you’ll stick with me. Maybe save it for later. I speak from experience 'cos this is my living so I hope you'll agree with at least some of it!

CD resolution Vs. 96/24?
There is no comparison. 96/24 is better – indeed I couldn’t do my work to the high standard my clients demand without it - and there is no sonic reason why it isn’t better for the amateur or hobbyist. There are plenty of cost and efficiency factors so that probably explains why people are still willing to defend CD resolution as if it were something worthy of defending in its own right, yet they still won't recognize the clear benefits that 96/24 delivers. It’s all too easy to dismiss 96/24 on the grounds that ‘you can’t hear a difference’ and that annoys the h*ll out of me. I’m here to tell you that you can hear a difference. We ALL can. Even if you’re deaf I hope to prove this:

Of course 96/24 uses more resources and processing time but in the world of real, living sound you don’t get anything for free and limitations come at a huge cost to accuracy. Strictly speaking, it is impossible to ‘capture’ sound - this cannot be overstated. There will always be some link in the recording/reproducing chain that gives the game away. Any increase in sampling rate cannot in itself be a bad thing!

Put simply: 44.1/16 is NOT enough for the job I do. My job is to take audio from vinyl sources and create a remaster for CD re-issue (or in some cases back to vinyl). If I was asked to work at 44.1/16 as a base sample/bit rate, I would be at a disadvantage. I would be losing a wealth of valuable information that I would never be able to find afterwards. Ever. For those of you who noted that fixing defects and unwanted sonic artefacts is easier in 96/24, I know this to be true from experience. It doesn’t mean ‘easier’ in terms of resources, it means ‘possible’ purely in sonic terms.

Not strictly related but I’ll give you an example of how more info is better: I often receive vinyl material that is originally recorded in mono. However I always make a stereo recording of the mono vinyl for restoration purposes. If you fold a stereo recording of a mono source down to mono, it will produce less noise however it will hinder restoration work because you no longer have the option of taking final samples from the ‘good’ channel, say, if a click is only on the left. Stretch this analogy to almost any region of sound and you soon realize that more information is always going to be more useful to an audio restoration project. Strictly speaking the only limiting factors are cost and speed.

This area of audio ‘salvage’ is becoming increasingly more important because original source material is fast disappearing. Master tapes do not last forever and even if well kept they will degrade through unavoidable processes such as chemical breakdown of polymers and tape strata, demagnetization and cosmic ray bombardment. A large number of well known recordings have already gone forever - it's just not public knowledge yet. Nobody really knows how this is going to affect the re-issue side of the industry because it's still not clear whether anyone cares... but I do! What is perfectly clear to me, if not to the music industry is that the sound quality locked up in those vinyl grooves is more than good enough for a CD re-issue or even better. And vinyl doesn’t degrade in sound quality like tape does. Look after it properly and it’ll be as good as new in hundreds of years time. Not so with tape. Digital? Stays the same forever of course but if it’s not good enough to begin with...

OK so I do this for a living – what of the hobbyist? Here sadly we must leave a lot of vinyl replay behind us because you do need a very good vinyl set-up to make a difference. We’re talking moving coil, Linn, Rega, low oxygen-copper leads, outboard A/D, yada yada, etc.
Sorry.

A note on CD: It’s important to remember that 44.1/16 (hereafter ‘CD’) was chosen because it satisfied the requirements of the audio industry at the time (early 1980's) and represented the agreed maximum bandwidth for the then new CD format, after arguments over size of the disk, maximum recording time, sound quality etc. The decision was taken with great care and no little discussion. It’s also very important to remember that CD was introduced more than quarter of a century ago. Things have moved on.

CD sound quality over analogue:
So many arguments... my bugbear is this: the argument still persists that because the average listener cannot hear up to, never mind OVER 22.05 KHZ (the sample rate per channel of CD) then it follows that overhead is built into the CD system. This thinking was valid at the time but is now known to be flawed. The assumption was that the only thing required was to reproduce all hearable frequencies at a basically accurate volume over a reasonably accurate time. Unfortunately it's that simplicity that got them caught out.
We now know that there are three fundamental values to any sound format; 1) frequency, 2) quantization (volume), 3) location (add echoes as you like ;p).
Location is the hard one, and it hardly ever gets mentioned because it has no strict measurable ‘rate’ but is itself a factor of the combination of the sample rate (frequency mostly) AND the quantization amount plus other factors such as reproducer (speaker). For the purposes of this argument I shall call it ‘timing’.

For frequency response 22.05 khz is enough for humans. However for timing and quantization (volume), 22.05 at 16 bit is not and it is audibly not. The problem is that you hardly ever get a chance to make a comparison. Go home and listen to a CD, then go to a hi-fi shop and listen to the same music on a SACD or DVD-A and you’ve already lost too much time in memory. Do it side-by-side and it’s a lot easier but still flawed – the equipment usually doesn’t sound the same, never mind the mastering.

OK so think about it purely mentally: Try and imagine two trains running on parallel tracks, both running at the same speed but one infinitesimally ahead of the other. What does that sound like? Optically it would be ‘a photo finish’. Sonically? The smaller the difference between the two trains location, the more numbers you need to throw at the measurement to make a measurement at all. In other words, more resolution. The human ear/brain combination is an astonishingly capable tool for detecting sound but perhaps not in the way that you think: bats can hear higher frequencies but they are only doing better, what we can ALSO do, and that is: locate sounds 'in space' and ‘in time’. For a bat it's a question of survival. Darwinism has seen to it that a bat can detect incredibly high frequencies but what is usually lost in the description is that bats further process those frequencies against an environment of time and space - and this we humans can also do (although not as well!)

Another way of saying this is: two sound events (from a stereo for example) can be very slightly ahead or behind each other and we can tell the difference with great accuracy, time after time. Think of a drum kit. You’ve never heard a drum kit recorded so it really, but really sounds like it’s in front of you, have you? (You’re very lucky if you have) but you know when it’s ‘live’ because it’s those very small timing differences that we can distinguish (plus all the surround info from real life). It’s those timing differences that CD cannot reproduce, and, given a good enough system, a vinyl LP can go some way towards reproducing. Fact: Analogue doesn’t have that timing problem. Digital always will. The only way to battle it is to throw more numbers into the digital mix.

Here’s a real life example: I shall use Bob Marley Live at the Lyceum for example for three important reasons: 1) I just worked on it, 2) it’s a live recording that didn’t suffer too much at the hands of studio engineers, 3) it has a top frequency response of around 17Khz, well below CD’s capabilites.
I’ve had the CD issue for years but listened to it maybe once - it sounds like a cardboard cut-out of the recording I was used to on vinyl, no atmosphere. The 96/24 DVD-A I just made, taken from vinyl (itself mastered from the original analogue tapes), walks all over the CD and takes names. Island had the master tapes as a source for their CD, I had a vinyl LP. Therefore and with no real argument, LP is clearly capable of producing something nicer sounding, not on a basis of frequency response but on something else. Something I now know to be accuracy in timing and quantization. Timing.

OK so it’s possible to argue that none of this matters – the average listener makes their choices on matters less lofty than such esoteric sound quality arguments. Good vinyl replay costs a lot. But for professionals it remains a constant factor in our choices because we know it works better at 96/24 and above. Fact. No argument.

All I can say in summation is that there IS a difference and it’s there to be heard. You CAN hear it.

Colin AKA See Why Audio
 
Jan 5, 2008 at 2:26 AM Post #40 of 45

grawk

Itinerant Miscreant
Joined
Mar 14, 2005
Posts
8,739
Likes
34
Quote:

Originally Posted by james__bean /img/forum/go_quote.gif
I make my own vinyl rips all the time. People always talk about how you can't hear above 20kHz, but I've directly compared a rip done in 24/48 to the same rip converted to 16/44 with foobar2000's ABX test. I picked out the difference like 95% of the time, which leaves the probability that I can't hear the difference like .01%. Its easy when you listen to the symbols.



the 24/16 difference is irrellevant. Why you can hear 48 vs 44 is simple. That isn't an easy conversion to do well. Most sample rate converters doing 44 from 48 do a lousy job.
 
May 10, 2010 at 11:20 AM Post #42 of 45

nick_charles

Headphoneus Supremus
Joined
Feb 26, 2008
Posts
3,180
Likes
334


Quote:
Hi all,

Fascinating to see what people really think about hi-res digital V CD. So many facts and even more fiction :p
Sorry, it’s a big, long post and I hope you’ll stick with me. Maybe save it for later. I speak from experience 'cos this is my living so I hope you'll agree with at least some of it!

CD resolution Vs. 96/24?
There is no comparison. 96/24 is better – indeed I couldn’t do my work to the high standard my clients demand without it - and there is no sonic reason why it isn’t better for the amateur or hobbyist. There are plenty of cost and efficiency factors so that probably explains why people are still willing to defend CD resolution as if it were something worthy of defending in its own right, yet they still won't recognize the clear benefits that 96/24 delivers. It’s all too easy to dismiss 96/24 on the grounds that ‘you can’t hear a difference’ and that annoys the h*ll out of me. I’m here to tell you that you can hear a difference. We ALL can. Even if you’re deaf I hope to prove this:

Of course 96/24 uses more resources and processing time but in the world of real, living sound you don’t get anything for free and limitations come at a huge cost to accuracy. Strictly speaking, it is impossible to ‘capture’ sound - this cannot be overstated. There will always be some link in the recording/reproducing chain that gives the game away. Any increase in sampling rate cannot in itself be a bad thing!

Put simply: 44.1/16 is NOT enough for the job I do. My job is to take audio from vinyl sources and create a remaster for CD re-issue (or in some cases back to vinyl). If I was asked to work at 44.1/16 as a base sample/bit rate, I would be at a disadvantage. I would be losing a wealth of valuable information that I would never be able to find afterwards. Ever. For those of you who noted that fixing defects and unwanted sonic artefacts is easier in 96/24, I know this to be true from experience. It doesn’t mean ‘easier’ in terms of resources, it means ‘possible’ purely in sonic terms.

Not strictly related but I’ll give you an example of how more info is better: I often receive vinyl material that is originally recorded in mono. However I always make a stereo recording of the mono vinyl for restoration purposes. If you fold a stereo recording of a mono source down to mono, it will produce less noise however it will hinder restoration work because you no longer have the option of taking final samples from the ‘good’ channel, say, if a click is only on the left. Stretch this analogy to almost any region of sound and you soon realize that more information is always going to be more useful to an audio restoration project. Strictly speaking the only limiting factors are cost and speed.

This area of audio ‘salvage’ is becoming increasingly more important because original source material is fast disappearing. Master tapes do not last forever and even if well kept they will degrade through unavoidable processes such as chemical breakdown of polymers and tape strata, demagnetization and cosmic ray bombardment. A large number of well known recordings have already gone forever - it's just not public knowledge yet. Nobody really knows how this is going to affect the re-issue side of the industry because it's still not clear whether anyone cares... but I do! What is perfectly clear to me, if not to the music industry is that the sound quality locked up in those vinyl grooves is more than good enough for a CD re-issue or even better. And vinyl doesn’t degrade in sound quality like tape does. Look after it properly and it’ll be as good as new in hundreds of years time. Not so with tape. Digital? Stays the same forever of course but if it’s not good enough to begin with...

OK so I do this for a living – what of the hobbyist? Here sadly we must leave a lot of vinyl replay behind us because you do need a very good vinyl set-up to make a difference. We’re talking moving coil, Linn, Rega, low oxygen-copper leads, outboard A/D, yada yada, etc.
Sorry.

A note on CD: It’s important to remember that 44.1/16 (hereafter ‘CD’) was chosen because it satisfied the requirements of the audio industry at the time (early 1980's) and represented the agreed maximum bandwidth for the then new CD format, after arguments over size of the disk, maximum recording time, sound quality etc. The decision was taken with great care and no little discussion. It’s also very important to remember that CD was introduced more than quarter of a century ago. Things have moved on.

CD sound quality over analogue:
So many arguments... my bugbear is this: the argument still persists that because the average listener cannot hear up to, never mind OVER 22.05 KHZ (the sample rate per channel of CD) then it follows that overhead is built into the CD system. This thinking was valid at the time but is now known to be flawed. The assumption was that the only thing required was to reproduce all hearable frequencies at a basically accurate volume over a reasonably accurate time. Unfortunately it's that simplicity that got them caught out.
We now know that there are three fundamental values to any sound format; 1) frequency, 2) quantization (volume), 3) location (add echoes as you like ;p).
Location is the hard one, and it hardly ever gets mentioned because it has no strict measurable ‘rate’ but is itself a factor of the combination of the sample rate (frequency mostly) AND the quantization amount plus other factors such as reproducer (speaker). For the purposes of this argument I shall call it ‘timing’.

For frequency response 22.05 khz is enough for humans. However for timing and quantization (volume), 22.05 at 16 bit is not and it is audibly not. The problem is that you hardly ever get a chance to make a comparison. Go home and listen to a CD, then go to a hi-fi shop and listen to the same music on a SACD or DVD-A and you’ve already lost too much time in memory. Do it side-by-side and it’s a lot easier but still flawed – the equipment usually doesn’t sound the same, never mind the mastering.

OK so think about it purely mentally: Try and imagine two trains running on parallel tracks, both running at the same speed but one infinitesimally ahead of the other. What does that sound like? Optically it would be ‘a photo finish’. Sonically? The smaller the difference between the two trains location, the more numbers you need to throw at the measurement to make a measurement at all. In other words, more resolution. The human ear/brain combination is an astonishingly capable tool for detecting sound but perhaps not in the way that you think: bats can hear higher frequencies but they are only doing better, what we can ALSO do, and that is: locate sounds 'in space' and ‘in time’. For a bat it's a question of survival. Darwinism has seen to it that a bat can detect incredibly high frequencies but what is usually lost in the description is that bats further process those frequencies against an environment of time and space - and this we humans can also do (although not as well!)

Another way of saying this is: two sound events (from a stereo for example) can be very slightly ahead or behind each other and we can tell the difference with great accuracy, time after time. Think of a drum kit. You’ve never heard a drum kit recorded so it really, but really sounds like it’s in front of you, have you? (You’re very lucky if you have) but you know when it’s ‘live’ because it’s those very small timing differences that we can distinguish (plus all the surround info from real life). It’s those timing differences that CD cannot reproduce, and, given a good enough system, a vinyl LP can go some way towards reproducing. Fact: Analogue doesn’t have that timing problem. Digital always will. The only way to battle it is to throw more numbers into the digital mix.

Here’s a real life example: I shall use Bob Marley Live at the Lyceum for example for three important reasons: 1) I just worked on it, 2) it’s a live recording that didn’t suffer too much at the hands of studio engineers, 3) it has a top frequency response of around 17Khz, well below CD’s capabilites.
I’ve had the CD issue for years but listened to it maybe once - it sounds like a cardboard cut-out of the recording I was used to on vinyl, no atmosphere. The 96/24 DVD-A I just made, taken from vinyl (itself mastered from the original analogue tapes), walks all over the CD and takes names. Island had the master tapes as a source for their CD, I had a vinyl LP. Therefore and with no real argument, LP is clearly capable of producing something nicer sounding, not on a basis of frequency response but on something else. Something I now know to be accuracy in timing and quantization. Timing.

OK so it’s possible to argue that none of this matters – the average listener makes their choices on matters less lofty than such esoteric sound quality arguments. Good vinyl replay costs a lot. But for professionals it remains a constant factor in our choices because we know it works better at 96/24 and above. Fact. No argument.

All I can say in summation is that there IS a difference and it’s there to be heard. You CAN hear it.

Colin AKA See Why Audio


Not at the consumer end according to the peer-reviewed AES article by Meyer and Moran, 20007. In which a 16/44.1 A/D/A stage was inserted after the analog outputs of several high resolution sources, not one listener was able to reliably detect the difference in DBT, they had over 50 trained listeners and over 500 trials. This paper is available from the AES.
 
May 14, 2010 at 12:53 AM Post #44 of 45

techenvy

500+ Head-Fier
Joined
Sep 23, 2009
Posts
802
Likes
10


Quote:
There's no advantage to capturing in 24/96 if the source is an LP. The only thing that will benefit from the added resolution is the bed of surface noise down at a low level where you can't hear it anyway. Capturing this high makes applying filters slow and increases the opportunity for hard drive underruns.

Capture to 16/44.1 and burn to redbook, and it will sound exactly the same.

See ya
Steve


really,, nice,   i bought one of those audio technica's tt with the usb output for ripping vinyl, but worried about the dac,  so i didn't bother, but i very much like to play the lp like it should be, more enjoyable and pleasing to my ear as well as therapeutic,  but ill bet the cheap cart o the AT may affect the recorded  sq?  should i not bother ripping vinyl with this tt?
http://www.google.com/products/catalog?hl=en&client=firefox-a&q=audio+technica+lp&cid=3798842594874120957&ei=5NXsS43xJ4O0wwWe4dXmCg&sa=title&ved=0CAcQ8wIwADgA#p
 
May 14, 2010 at 6:14 AM Post #45 of 45

regal

Headphoneus Supremus
Joined
May 27, 2006
Posts
3,645
Likes
17
I've been searching for a Lp to digital transfer shop that uses really good ADC like the PM2 or other ladder/R2R based ADC.  I think the loudness war is part of the reason CD's started sounding bad after 2000 but also the ADC's went S-D about the same time.
 

Users who are viewing this thread

Top