CD Quality Is Not High-Res Audio

[Phelonious Ponk]

follow up:

You did have to be way more careful in the days before non-destructive editing and a gazillion virtual tracks.

You did. The grunge build-up could be significant, to the point that I marvel at how clean some of the heavily-overdubbed recordings from the 70s sound (Steely Dan comes to mind).

Tim

 

[LL21]

...Ever notice how compressed Born To Run is? ...Tim

+1...when will we hear great remasters of Bruce Springsteen? Are there any?

 

[LL21]

... it is the mix and master engineer who have to get it right and then the format (for me) does not matter.

in my limited experience, totally agree. Great CDs can sound great. FIM, MA Recordings, Analogue Productions, Verve Masters, Blue Note Remasters, Channel, Harmonia Mundi...some truly beautiful mastering jobs out there. On more modern stuff, I have enjoyed the superb remasters by Bob Ludwig on CD of the Police and Sting albums.

 

[FrantzM]

in my limited experience, totally agree. Great CDs can sound great. FIM, MA Recordings, Analogue Productions, Verve Masters, Blue Note Remasters, Channel, Harmonia Mundi...some truly beautiful mastering jobs out there. On more modern stuff, I have enjoyed the superb remasters by Bob Ludwig on CD of the Police and Sting albums.

Indeed! Seems to be that in the end .. great recordings , regardless of medium remain ... Great Recordings! One can marvel at the very famous Audio Writer, TAS founder stating that to his ears the Mercury CD could be said to equal or at least surpass the Vinyl ... the famous shaded dogs ... The person who mastered those LPs and CDs (and if I not mistaken, the SACD) the late and great Irma Cozart hared the same sentiment ... so does your humble servant .

 

[Phelonious Ponk]

+1...when will we hear great remasters of Bruce Springsteen? Are there any?

None that I"m aware of. The compression on Born to Run can't be escaped. It's analog compression; it's on the tape. Landau was going for that wall of sound thing and he got it. That doesn't mean it couldn't be remastered and improved, though. Born to Run really doesn't bother me, though. The River? Such great songs, pretty badly recorded. I'm not sure you could fix it in mastering. And then theres most of his stuff from the digital age; I find Magic almost unlistenable. Sonically, the best Springsteen, to my ears, is Tunnel of Love. It was good on vinyl, it's good on CD. None of that keeps me from listening, though. Great sound is a bonus. Great music is a necessity.

Tim

 

[LL21]

None that I"m aware of. The compression on Born to Run can't be escaped. It's analog compression; it's on the tape. That doesn't mean it couldn't be remastered and improved, though. Born to Run really doesn't bother me, though. The River? Such great songs, pretty badly recorded. I'm not sure you could fix it in mastering. And then theres most of his stuff from the digital age; I find Magic almost unlistenable. Sonically, the best Springsteen, to my ears, is Tunnel of Love. It was good on vinyl, it's good on CD. None of that keeps me from listening, though. Great sound is a bonus. Great music is a necessity.

Tim

interesting...I have had the MFSL tunnel of love remaster in my amazon inbox for a while...is this one to go for you think? Pure guesswork...just asking for your own guesstimate.

 

[LL21]

interesting...I have had the MFSL tunnel of love remaster in my amazon inbox for a while...is this one to go for you think? Pure guesswork...just asking for your own guesstimate.

actually these are supposedly Japanese imports based on the "LP version"...supposedly better? or just all foreign import hype.

 

[Phelonious Ponk]

interesting...I have had the MFSL tunnel of love remaster in my amazon inbox for a while...is this one to go for you think? Pure guesswork...just asking for your own guesstimate.

I haven't a clue about the MFSL version, but the vinyl and CD are both pretty good; very good relative to much of Bruce's work. I think it's just better recording, so I'd guess the MFSL will be good as well.

Tim

 

[LL21]

I haven't a clue about the MFSL version, but the vinyl and CD are both pretty good; very good relative to much of Bruce's work. I think it's just better recording, so I'd guess the MFSL will be good as well.

Tim

thanks!

 

[rbbert]

On the face of it that sounds right. What I am not so sure of however is if the equivalent voltage increments is the same or not. My plebe mind tells me that if the increments are the same, you are throwing away headroom (for listening) if they are smaller then you really are getting more density and resolution

The voltage steps are the same, but because of the logarithmic nature of auditory perception and the decibel scale, there are more voltage steps per decibel, so the quantization errors should be smaller. Whether that makes a difference depends on who you ask.

 

[rbbert]

The output of a DAC is also a smooth analog waveform ...

Well strictly speaking the output of the actual DAC is steps, but after the reconstruction filter (which is a part of virtually all DAC "chips") it is a smooth analog waveform. I have proposed the analogy (as has tomelex) to the discrete nature of analog tape and the analog LP as well, it's all a question of scale.

 

[LL21]

I haven't a clue about the MFSL version, but the vinyl and CD are both pretty good; very good relative to much of Bruce's work. I think it's just better recording, so I'd guess the MFSL will be good as well.

Tim

Ok...did a little surfing. rumor has it in 1995 the Japanese did a nice job remastering born in the usa and tunnel of love...more than a few people did comparisons, and a few ordered and were pleasantly surprised. So I ordered both since I don't particularly enjoy listening to my current 0.01 order-in-the-mail-from-the-90s Born in the USA...which says a lot given that I enjoy listening to nearly every single one of my 1800 albums...and that one usually sits on the shelf despite liking Bruce Springsteen's music...and I don't own tunnel of love at all. both used...we shall see.

 

[Phelonious Ponk]

Ok...did a little surfing. rumor has it in 1995 the Japanese did a nice job remastering born in the usa and tunnel of love...more than a few people did comparisons, and a few ordered and were pleasantly surprised. So I ordered both since I don't particularly enjoy listening to my current 0.01 order-in-the-mail-from-the-90s Born in the USA...which says a lot given that I enjoy listening to nearly every single one of my 1800 albums...and that one usually sits on the shelf despite liking Bruce Springsteen's music...and I don't own tunnel of love at all. both used...we shall see.

Are these Japanese remasters CDs or vinyl. And if they're CDs, where did you order them?

Tim

 

[LL21]

Are these Japanese remasters CDs or vinyl. And if they're CDs, where did you order them?

Tim

CDs...from Amazon. I looked up the album numbers which related to the comments about a set of 1995 remasters from japan which apparently are good, and found them on amazon...a bunch second hand actually. not cheap, but with the time, sweat and tears we put into music...and I like bruce, I figured lets give it a shot. I have been pleasantly surprised by some finds.

 

[LL21]

http://www.amazon.co.uk/gp/product/B0009J8GXK/ref=oh_details_o00_s00_i00?ie=UTF8&psc=1

http://www.amazon.co.uk/gp/product/B0009J8GY4/ref=oh_details_o01_s00_i00?ie=UTF8&psc=1

Quite sure you can find these on Amazon.com as well.

 

[Phelonious Ponk]

Thanks. Let us know how they sound.

P

 

[LL21]

Thanks. Let us know how they sound.

P

For sure. Should be a week or 2. While I have not always found remasters to be dramatically better than the original...I would say I have found 90% of them to be much better than the original CDs they replaced (FIM, Analogue Productions, MFSL, Bob Ludwig/Gateway, and even the sometimes-not-so-well liked RVG Remasters)...this is replacing either the original Police or U2 CDs...or replacing much older jazz recordings. The difference between original Red Garland Trio stuff and remastered is a huge difference imho. Maybe about 8% are a smidge better but not that I would necessarily go thru the brain damage again to track down a second hand copy of the remaster if I had to do it all over again. Maybe 2% is simply a marketing ploy...I returned those.

 

[Julf]

Well strictly speaking the output of the actual DAC is steps, but after the reconstruction filter (which is a part of virtually all DAC "chips") it is a smooth analog waveform.

Indeed. The "stair-step waveform" is one of the most persistent audiophile myths. A reconstruction filter is an essential part of the DAC (apart from some NOS dacs that are designed "wrong" on purpose), and are also an essential part of the process. As a result of the reconstruction filter, the output waveform does not have steps, but actually reproduces the original analog waveform. It is surprising how some people just don't seem to get the Nyquist–Shannon theorem.

Here is a good write-up: [url='http://onetwothreeaudio.wordpress.com/2013/08/07/digital-audio-basics-stair-steps-and-sample-rates/"]123audio: Digital Audio Basics: Stair Steps and Sample Rates[/url]

 

[Groucho]

Indeed. The "stair-step waveform" is one of the most persistent audiophile myths. A reconstruction filter is an essential part of the DAC (apart from some NOS dacs that are designed "wrong" on purpose), and are also an essential part of the process. As a result of the reconstruction filter, the output waveform does not have steps, but actually reproduces the original analog waveform.

I can see where this explanation would fail to satisfy. Clearly the input to the reconstruction filter is a quantized stair step and the deviations from the original signal could be repetitive and linked to the sample rate vs. signal characteristics, the results of which would be audible as 'tones' and which the filter could do nothing about. Reducing the bit depth from 24 to 2 or 1 would sound especially appalling. But the explanation misses out the other essential element, which is dither. Dither effectively scrambles the 'rounding up' or 'rounding down' of quantization so that the non-random effects of quantization are removed completely, substituting pure random noise - or at least as pure as 32 or 64 bit floating point mathematics will allow, which is very very pure indeed.

So what this idea says is that when going from 24 bit audio (as recorded in the studio) to 16 for release on CD, say, the addition of appropriate dither will result in *no difference* in audio quality except for an increase in noise level from a very very very low level, to a very very low level - orders of magnitude below the ordinary analogue noise inherent to every recording, and way below the acoustic noise in the listening room. This particular scenario is obviously the most common in recording studios, and the one where the proof is actually mathematical, as the 24 and 16 bit recordings have been made with identical hardware, and the only difference is made by software whose efficacy can be proved mathematically. I presume that the process is completely automated and the technician doesn't even need to know what dither is when the software generates the 16 bit file from the studio master. You may protest that this is the Achilles Heel and that we just don't know if the process has been followed correctly, but then you are simply putting your faith in the idea that the 24 bit format may have been produced less badly by incompetents than the 16 (what else have they got wrong?), rather than the 24 bit somehow being inherently, audibly better than the 16.

 

[Julf]

Reducing the bit depth from 24 to 2 or 1 would sound especially appalling.

Indeed - the signal-to-noise ratio would be only 6 or 12 dB. That is why DSD needs noise shaping to shift that noise to inaudible frequencies.

Dither effectively scrambles the 'rounding up' or 'rounding down' of quantization so that the non-random effects of quantization are removed completely, substituting pure random noise - or at least as pure as 32 or 64 bit floating point mathematics will allow, which is very very pure indeed.

Yes. Without dither, the quantisation noise might have components that are related to the input signal, while dither turns it into random noise.

So what this idea says is that when going from 24 bit audio (as recorded in the studio) to 16 for release on CD, say, the addition of appropriate dither will result in *no difference* in audio quality except for an increase in noise level from a very very very low level, to a very very low level - orders of magnitude below the ordinary analogue noise inherent to every recording, and way below the acoustic noise in the listening room.

Exactly. There is a difference between the 24-bit and 16-bit version - the former has a theoretical SNR (without dither) of 144 dB, the latter "only" 96 dB - both way below the actual source.

This particular scenario is obviously the most common in recording studios, and the one where the proof is actually mathematical, as the 24 and 16 bit recordings have been made with identical hardware, and the only difference is made by software whose efficacy can be proved mathematically. I presume that the process is completely automated and the technician doesn't even need to know what dither is when the software generates the 16 bit file from the studio master. You may protest that this is the Achilles Heel and that we just don't know if the process has been followed correctly, but then you are simply putting your faith in the idea that the 24 bit format may have been produced less badly by incompetents than the 16 (what else have they got wrong?), rather than the 24 bit somehow being inherently better than the 16.

Well put.