Why audiophiles still use CDs?

Crystal Mastering » CD mastering overview

Le Concombre Masqué · January 22, 2015

which player to you use for computer playback ? differences are huge...

Hi alfe,

Exactly how would you suggest mastering engineers adapt their masters to CD?

Also, what would make you assume mastering engineers "adapt their masters" to "their hard drives"?

And exactly what do you imagine they're doing to effect this?

When I master an album, I master it for the *recording* and not for the delivery format.

I have found this to be true of all the recordings I admire - all were mastered for the recording. The delivery format will do what it will do - which is determine just how much of the recording we get to hear. In my opinion, attempts to "compensate" for the format invariably take the recording *away* from where it should be, not toward it.

All that said, even before any mastering changes, I find the results exactly as Roch describes. Just taking the original recording and transferring it to a CD, then comparing the disc (in any transport or player I've ever heard) and playback from the computer with the original recording will reveal (at least to my ears) that the computer playback sounds identical to the original and the CD never does. With luck, it can get *very* close. But in my view, there is a huge difference between very close and spot on. And most of the time, it isn't that close.

All just my perspective of course.

Best regards,

Barry

Soundkeeper Recordings

www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

alfe · January 22, 2015

Hi alfe,

Exactly how would you suggest mastering engineers adapt their masters to CD?

Also, what would make you assume mastering engineers "adapt their masters" to "their hard drives"?

And exactly what do you imagine they're doing to effect this?

When I master an album, I master it for the *recording* and not for the delivery format.

I have found this to be true of all the recordings I admire - all were mastered for the recording. The delivery format will do what it will do - which is determine just how much of the recording we get to hear. In my opinion, attempts to "compensate" for the format invariably take the recording *away* from where it should be, not toward it.

All that said, even before any mastering changes, I find the results exactly as Roch describes. Just taking the original recording and transferring it to a CD, then comparing the disc (in any transport or player I've ever heard) and playback from the computer with the original recording will reveal (at least to my ears) that the computer playback sounds identical to the original and the CD never does. With luck, it can get *very* close. But in my view, there is a huge difference between very close and spot on. And most of the time, it isn't that close.

All just my perspective of course.

Best regards,

Barry

Soundkeeper Recordings

www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

Hi Barry,

bdiament · January 22, 2015

which player to you use for computer playback ? differences are huge...

Hi Le Concombre Masqué,

I have several players, ranging from server type applications to professional audio software.

These include (in no particular order) iTunes, Amarra, Fidelia, DSP Quattro, Peak, Wave Editor, Triumph, soundBlade, and a bunch of others. I find that despite the differences in the software, what I described above has been consistently true -- for my ears, *all* beat CD playback from any transport or player I've ever heard.

Which player do you use?

Best regards,

Barry

bdiament · January 22, 2015

Hi Barry,

Crystal Mastering » CD mastering overview

Hi alfe,

I see the link.

I don't see the answers to my questions though. Perhaps I missed them.

Are you affiliated with Crystal?

Best regards,

Barry

Le Concombre Masqué · January 22, 2015

I recently became a fan of HQplayer (see my signature for the context, which matters I guess, notably the use of a delta sigma dac). i used to mighty appreciate Amarra 3 for PCM and to consider A+ (never been a big fan though) the best for dsd

I turned to Computer audio when my old Studer A (but it might have become half D with all the repairs) 730 fell me. What I get from a quite cheap DAC and my Mac is way above what I got from the Studer spinning CDs. And what I get today is so much better, using the same DAC, just tweaking and upgrading players, than what I got 1 1/2 year ago that I don't feel in a hurry to change my DAC.

best regards

Hi Le Concombre Masqué,

I have several players, ranging from server type applications to professional audio software.

These include (in no particular order) iTunes, Amarra, Fidelia, DSP Quattro, Peak, Wave Editor, Triumph, soundBlade, and a bunch of others. I find that despite the differences in the software, what I described above has been consistently true -- for my ears, *all* beat CD playback from any transport or player I've ever heard.

Which player do you use?

Best regards,

Barry

Soundkeeper Recordings

www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

esldude · January 22, 2015

You shouldn't hear much (if any) difference between so-called high-res and 16/44 from the same master, IF that master is analog, or originally recorded at 16/44 (or less). The reason for the former is because there is nothing on those masters that exceeds the capability of 16/44 to capture it. Professional, analog tape recorders from Ampex, Studer, etc. and used in pro-recording all over the world before the advent of digital were simply not maintained to frequencies above 15 KHz. That doesn't mean that they didn't have anything on them above that frequency, they may have had significantly useful info up into the low 20KHz region. But it would be the rare recording indeed that kept that kind of high-frequency content beyond the first couple of playings due to self-erasure and print-through. If they are not fresh masters, they just don't have much of their +15KHz material left. As for a performance that was originally captured as a 16/44 master, that should be self explanatory.

The useful analogy here is this one: If you watch an HD broadcast of an old NTSC network videotape on a new OLED high-definition TV screen, that old videotape is not going to magically become high definition. It's still going to be a maximum of 525 X ~800 lines of video, because that's all that was recorded on the tape.

On the other hand, a carefully made 16/44 CD of an analog master can sound superb, better than that performance ever sounded on vinyl and in many cases better than some SACD or so-called high-res re-issues of same.

Being an Audacity cowboy, it is instructive to look at hires recordings. For examples, I used Eric Clapton's 461 Ocean Blvd from 1974. The hires version I had is 96/24 from HDtracks. It does have ultrasonic content. I used a steep filter to remove everything below 20 khz. Not much left after that. I then slowed down the file to 1/4 normal speed so I could clearly hear the ultrasonic content. A bit of low level tape hiss, and occasionally what must have been cymbals on the drumset. Not very loud either.

Did the same with Tea for the Tillerman by Cat Stevens from 1970. HDTracks 24/192. Don't know if they used some noise reduction or just a better machine. The ADC was an MSB. It was nearly devoid of the tape hiss. It also had the tinkling of cymbals every few seconds at low level.

The same thing from some of Barry's samples showed a fairly clean sound with again some low level tinkling sounds. Just not much up there. And Barry uses microphones good to 40 khz.

A similar procedure of upsampling a CD, and carefully filtering below 20 khz with a brickwall filter leaves only the transition band. That would contain the horrid ringing artifacts from the anti-aliasing filter used for recording. What do you get? A very small number of short very low level blips of a whistle sound. So much for the horrid ringing.

bdiament · January 22, 2015

...The same thing from some of Barry's samples showed a fairly clean sound with again some low level tinkling sounds. Just not much up there. And Barry uses microphones good to 40 khz...

Hi Dennis,

It depends on how you define "not much" and it depends on exactly what tracks you sampled--and how much of the track.

By the way, the mics are within 1 dB at 40 kHz. They're "good" to well beyond that. There are tracks on some Soundkeeper discs that have content up around 67 kHz. (If you only tried the .wav samples on the format page, I don't know what you'd see.)

If you want to do serious analysis of spectral and other content in audio files, I would suggest looking into something like SpectraFoo. I have Audacity as well as other "analysis" type software and SpectraFoo shows what the others lack. (Note I'm not saying Audacity or others won't show the info but some programs I've used seem to lack certain sensitivities. I don't know why but suspect the sophistication--or lack thereof-- in the algorithms. I've not encountered anything at the level of Metric Halo's software. It might be out there but if it is, I haven't seen it yet.)

Not that the supersonic content is what makes the difference but typical studio mics are lucky to even approach 20 kHz much less go past it.

Often they "achieve" their treble response with a resonant peak on the diaphragm. So, comparing if you don't see a *significant* difference in bandwidth between a Soundkeeper 4x recording and one made using typical studio microphones, I would suspect the tool used for the comparison.

Best regards,

Barry

esldude · January 22, 2015

Hi Dennis,

It depends on how you define "not much" and it depends on exactly what tracks you sampled--and how much of the track.

By the way, the mics are within 1 dB at 40 kHz. They're "good" to well beyond that. There are tracks on some Soundkeeper discs that have content up around 67 kHz. (If you only tried the .wav samples on the format page, I don't know what you'd see.)

If you want to do serious analysis of spectral and other content in audio files, I would suggest looking into something like SpectraFoo. I have Audacity as well as other "analysis" type software and SpectraFoo shows what the others lack. (Note I'm not saying Audacity or others won't show the info but some programs I've used seem to lack certain sensitivities. I don't know why but suspect the sophistication--or lack thereof-- in the algorithms. I've not encountered anything at the level of Metric Halo's software. It might be out there but if it is, I haven't seen it yet.)

Not that the supersonic content is what makes the difference but typical studio mics are lucky to even approach 20 kHz much less go past it.

Often they "achieve" their treble response with a resonant peak on the diaphragm. So, comparing if you don't see a *significant* difference in bandwidth between a Soundkeeper 4x recording and one made using typical studio microphones, I would suspect the tool used for the comparison.

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

Now I didn't say the ultrasonic content wasn't more extended. It clearly was, maybe I should have mentioned that. However the level of that content is fairly low. When slowed down so I could hear some of it, there still isn't much to hear. I can hear to something like 15 khz. So slowed down by a factor of 4 I might hear something equivalent to 60 khz were it there. I probably would need to boost the level of what is left some to get that audible. Working from memory the analysis showed some real content to about 50 khz and a steep drop past that. The other recordings from old reel tape didn't reach that far. Pretty much nothing beyond very low noise past 25 khz.

bdiament · January 23, 2015

Now I didn't say the ultrasonic content wasn't more extended. It clearly was, maybe I should have mentioned that. However the level of that content is fairly low. When slowed down so I could hear some of it, there still isn't much to hear. I can hear to something like 15 khz. So slowed down by a factor of 4 I might hear something equivalent to 60 khz were it there. I probably would need to boost the level of what is left some to get that audible. Working from memory the analysis showed some real content to about 50 khz and a steep drop past that. The other recordings from old reel tape didn't reach that far. Pretty much nothing beyond very low noise past 25 khz.

Hi Dennis,

What sample were you looking at in Audacity?

The shape of the spectral distribution depends on the program material of course., not on the microphones alone.

As I said, what you see will depend on what part of the Soundkeeper catalog you sampled as well as the tool you use to take the look.

Best regards,

Barry

esldude · January 23, 2015

Hi Dennis,

What sample were you looking at in Audacity?

The shape of the spectral distribution depends on the program material of course., not on the microphones alone.

As I said, what you see will depend on what part of the Soundkeeper catalog you sampled as well as the tool you use to take the look.

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

This was done a while back. Not today. I am pretty sure I looked at all of them for ultrasonic content. I slowed down and listened to filtered versions of the couple that had the most content left after filtering. And yes no disgreement it varies depending on what is there.

As for the tool used for looking. I looked at FFT's for some of it. Mostly looked at the spectrogram view. Not sure what Spectrafoo would show I didn't see this way. Though if you can describe what is shows extra that would be nice. You can add gain to Audacity's spectral view as well as frequency based gain that increases as frequency goes up. Helpful for seeing when content was just a little low in level versus when all that is left is simply noise.

Also, if you can post screenshots illustrating your recordings that sounds like a good educational idea.

Mystic · January 23, 2015

My Vecteur L-4.2 CDP far out performs my MacMini/Oppo 105 for musicality, detail, bass extension (even the HDtracks version of the same album).

I'm a vinyl junkie for the most part, so I do prefer playing physical media rather than scrolling on a screen same as someone else posted earlier.

But maybe this year I may explore some better dac's (looking at a Ayre QB-9 to audition this year).

DigiPete · January 23, 2015

I shunned CD's for decades.

I have better digital equipment now and I love (ripped) CD's.

(Disclaimer: the well recorded ones)

The weak link seems to be filtering and not the CD format as such!!!

EuroDriver · January 23, 2015

My Vecteur L-4.2 CDP far out performs my MacMini/Oppo 105 for musicality, detail, bass extension (even the HDtracks version of the same album).

I'm a vinyl junkie for the most part, so I do prefer playing physical media rather than scrolling on a screen same as someone else posted earlier.

But maybe this year I may explore some better dac's (looking at a Ayre QB-9 to audition this year).

With the OPPO 105D, I have had very good results with using a SOTM SMS-100 in between it and MacMini. There are lots of ways you can use the SMS-100, but the one that sounds really good is to run HQ Player Desktop on the MacMini doing PCM to DSD 128 conversion and have the SMS-100 run in HQ Player NAA mode.

check out out the HQ Player Kickstarter thread

OPPO 105d fronted by SMS-100 however, is no match for an Exasound E22 driven by HQ Player. If you want to be free from fooling with software player then PS Audio DirectStream is a very good sounding option

PeterSt · January 23, 2015

For examples, I used Eric Clapton's 461 Ocean Blvd from 1974. The hires version I had is 96/24 from HDtracks. It does have ultrasonic content. I used a steep filter to remove everything below 20 khz. Not much left after that. I then slowed down the file to 1/4 normal speed so I could clearly hear the ultrasonic content. A bit of low level tape hiss, and occasionally what must have been cymbals on the drumset. Not very loud either.

The point here is that we must try to envision what the higher frequency content actually does in practice, or should do;

If you have a square wave in the music data, this only works out to sounding square when sufficient harmonics can be rendered. Think like Miska's 7KHz example and that the 21Khz harmonic of that wouldn't be audible to us. So we perceive a sine. Same would happen when that 7KHz is bandlimited to 20KHz, then we would measure a sine just the same.

Now the bandlimiting happens at 96Khz so we will measure a square from the 7KHz tone. But mind you, without notice we take it for granted that the level of the 21KHz is the same as the fundamental. With a test signal this will be so ...

People can correct me when needed, but when such sound is captured with a microphone, this all fairly works the other way around; Assumed that the level of the 21KHz fundamental (which plays live in a room) is on par with the 7KHz fundamental to begin with (but anyway that will be depicted by nature) then we would like the 21KHz to be captured by the microphone at the same level as the 7KHz. Ok, this will be wrong-ish, so we'd have to say that the 21KHz has to be captured at the same level that it happens in that room. And this I doubt;

When the distance of the microphone is too far from the sound source, the higher frequencies will have attenuated too far to have a fair level of that tone. This means that relatively the lower frequency 7KHz will be louder, and also be more a sine again caught at that distance.

Nice, but I see not much difference from us listening at that same distance, in the recording room.

Of course we can make that 7KHz now 4KHz, so the 12KHz required can be perceived by us, but the story does not change. So the 4KHz square again will be more a sine than intended (as how we perceive that). Notice that this will show so in the measurement just the same.

Different it becomes when the microphone is close by; Now mentioned 12KHz will be more "on par" and we will measure a 4KHz square (or at least a better one). So we now can also hear it ?

What lacked in the above was the listening distance to the speakers and the same thing will happen; the 12Khz will reach us more attenuated and the 4KHz will still be more of a sine than intended.

Or

Because the lower frequencies spread more than the higher ones, net it will work out the same and higher frequencies do not reach our ears (or distant microphone for that matter) with more attenuation because the lower frequencies also attenuate because of the more spreading.

Or

Because of the same reason (spreading) the lower frequencies meet more reflections and reach us (and mic) again at a higher level than the 12KHz.

Or not with sufficient room treatment.

Or different when directional horns are used for the higher frequencies only. Or ...

Sure, I can make everything more difficult than it is without such thinking. But point is that I tend to agree with Denis' point of view that no such heavy "attenuated" (less well captured) higher frequency content can do much. But we do need to project this all to in-band and how we will perceive transients (or squares) better. I say that this can't happen at all in the higher frequencies anyway because of all the (distortion !) reasons I mentioned. But I also dare say that the higher bandlimited "signal" will improve the LOWER frequencies. Easy enough : that works more easily, if you understood my BS. Harmonic distance is also smaller in terms of frequency and thus attenuation difference / reflections difference is smaller.

Lastly an I hope relevant remark :

Supposed you can lift the level of the tweeter with only 1dB at the very end of the spectrum (20KHz). And suppose that this lifting "linearly" starts at around 7KHz (so think 0.5 dB at 13.5KHz) ... do we know what HUGE differences this makes for spaciousness, width, depth and what not ? we can easily sprout "micro detail". But all what happens is just what I described - the harmonics of fundamentals at 3x lower frequencies are boosted slightly, so the base frequencies become more square and the lot together allows us to locate better. Notice that this is somewhat more complicated to explain, because obviously it is the harmonics only (which again are sines) doing this to us, but because both fundamental and harmonic are related, we suddenly "see" (for us !) smeared lower frequencies pinpointed in space because the higher frequency fundamental bonds them in that space (and still we think we hear the lower frequency only).

Peter

alfe · January 23, 2015

Hi alfe,

I see the link.

I don't see the answers to my questions though. Perhaps I missed them.

Are you affiliated with Crystal?

Best regards,

Barry

Soundkeeper Recordings

www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

Should have one of your master one day and may be you will be my new reference.

If you rip data and get a better sound, the difference is called jitter happily introduced by poor sound card,poor quality hard drives,cable or what so ever during mastering process.

Less difference you hear better is the mastering.

bdiament · January 23, 2015

Should have one of your master one day and may be you will be my new reference.

If you rip data and get a better sound, the difference is called jitter happily introduced by poor sound card,poor quality hard drives,cable or what so ever during mastering process.

Less difference you hear better is the mastering.

Hi alfe,

I believe the differences are due to much more than that. The transport/player must read the disc, keep the laser focused, track the spiral of pits, convert the 8:14 modulated input to binary, perform error correction and ultimately, D-A conversion, all in pretty close to real time and all commonly from a single power supply. More sophisticated transport/players will offer separate power supplies but all must still be performed within the time constraints of playing back the audio.

Perhaps having all these performed outside of the requirement for D-A conversion and playback at the same time, which also allows for continual reads until the data is right, has something to do with it. Clearly, some folks have a different experience but what I've found has always been without exception: I've yet to hear playback from a CD that sounds indistinguishable from the master used to make it. On the other hand, with computer playback, it is a matter of routine. At least that has been my experience, ever since I started using a computer for audio.

So exactly how would a mastering engineer adapt their masters to CD vs. computer?

It's okay if you don't want to answer but since you mentioned it, I'm curious. I would also ask for an example of this so I could try to hear it for myself.

If you ever do hear a Soundkeeper master (the 24/192 .aif version is really *it*), I would hope you enjoy the music as well as the sound.

As I mentioned in my earlier post and as I do with any album I master, they are all mastered for the recording itself, not for any particular format. ;-}

Best regards,

Barry

http://www.computeraudiophile.com/f8-general-forum/encoding-and-decoding-cd-22848/

bdiament · January 23, 2015

Duplicate.

Deleted.

alfe · January 23, 2015

Hi alfe,

I believe the differences are due to much more than that. The transport/player must read the disc, keep the laser focused, track the spiral of pits, convert the 8:14 modulated input to binary, perform error correction and ultimately, D-A conversion, all in pretty close to real time and all commonly from a single power supply. More sophisticated transport/players will offer separate power supplies but all must still be performed within the time constraints of playing back the audio.

Perhaps having all these performed outside of the requirement for D-A conversion and playback at the same time, which also allows for continual reads until the data is right, has something to do with it. Clearly, some folks have a different experience but what I've found has always been without exception: I've yet to hear playback from a CD that sounds indistinguishable from the master used to make it. On the other hand, with computer playback, it is a matter of routine. At least that has been my experience, ever since I started using a computer for audio.

So exactly how would a mastering engineer adapt their masters to CD vs. computer?

It's okay if you don't want to answer but since you mentioned it, I'm curious. I would also ask for an example of this so I could try to hear it for myself.

If you ever do hear a Soundkeeper master (the 24/192 .aif version is really *it*), I would hope you enjoy the music as well as the sound.

As I mentioned in my earlier post and as I do with any album I master, they are all mastered for the recording itself, not for any particular format. ;-}

Best regards,

Barry

Soundkeeper Recordings

www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

You will see in this link the CD spec (Rom because of copyright of RB) that the Physical jitter is part of the spec so having exactly the same sound is wishful thinking.

The answer to your question is dither.

bdiament · January 23, 2015

http://www.computeraudiophile.com/f8-general-forum/encoding-and-decoding-cd-22848/

You will see in this link the CD spec (Rom because of copyright of RB) that the Physical jitter is part of the spec so having exactly the same sound is wishful thinking.

The answer to your question is dither.

Hi alfe,

I believe that in time, we're going to find that there is more to the story than jitter. Just a hunch on my part.

Regarding the answer to the question, are you saying dither is how a mastering engineer adapts a master to CD?

If so, how would that make the CD sound better than playback of the same file from a computer?

(This is going back to what you suggested in post #51 in response to Roch's statement. Or did I misunderstand?)

Best regards,

Barry