Jump to content
IGNORED

Interesting article with Steven Wilson


Recommended Posts

Thanks for posting this. It is an interesting article.

 

I am of the opinion we do not benefit from more than 48/24. But his idea, that there is a tiny difference (unlike those who call it night and day) is not extreme. 96/24 isn't too crazy for a number of reasons. Mr. Wilson apparently isn't saying more than 96 is of any benefit.

 

I very much applaud his idea to skip mastering. Do your mix, and a flat transfer to the final product. Mastering seems to be very nearly ruining more than 90% of currently released music.

And always keep in mind: Cognitive biases, like seeing optical illusions are a sign of a normally functioning brain. We all have them, it’s nothing to be ashamed about, but it is something that affects our objective evaluation of reality. 

Link to comment

 

Hi Paul,

 

Mr. Wilson is, erroneously in my view, equating mastering with knob twiddling. (The same view could be taken of any part of the process, including the recording and mixing. It depends who's doing it.)

The fact is, if his recordings weren't mastered, there would be nothing for anyone to hear.

 

In order for CDs to be replicated, the manufacturing facility must receive the project in a certain format. All the tracks must be in their proper sequence, with the desired amount of spacing (if any) between tracks.

 

In addition, relative levels between tracks must be established -- unless one doesn't mind having to constantly adjust the volume level as the album progresses from song to song. (A poorly mastered album might have the gentle ballads sounding louder than the all-out screaming rockers.) One of the functions of proper mastering is to make the album a cohesive whole.

 

Final level must be established too, particularly if the mixes went through analog. Any A-D converter I know will exhibit its best performance and lowest distortion when the maximum peak level does not exceed -6 dB. Some colleagues I've spoken with would say -10 or -12 dB. In order to make the most of the delivery medium, the level will often be raised. (I'm not referring to the modern trend of compressing dynamics to the point where even the "low" parts are loud. I'm saying even with full dynamic range preservation, the level of a properly done mix will tend to go up with a proper mastering.)

 

Notice I have not mentioned any EQ or other processing a mastering engineer might utilize to make the most of a set of mixes. I'm just talking about making the best straight transfer of a collection of songs in order to create a unified album.

 

And speaking of EQ, if most audiophiles saw how most (certainly not all) records were made, got to hear what the most popular microphones do to the sound, saw where the microphones were placed, and heard the monitoring on which all the sonic decisions were based, they would understand that most recordings need help.

 

But again, even for the best "flat transfer", if there is no mastering, there will be nothing for the listener to hear.

Send the raw mixes to the replication facility and what will they do? They will send it to their in-house "mastering room" to be properly formatted for production. If this isn't done, they can't get to the next step.

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

Link to comment
Hi Paul,

 

Mr. Wilson is, erroneously in my view, equating mastering with knob twiddling. (The same view could be taken of any part of the process, including the recording and mixing. It depends who's doing it.)

The fact is, if his recordings weren't mastered, there would be nothing for anyone to hear.

 

In order for CDs to be replicated, the manufacturing facility must receive the project in a certain format. All the tracks must be in their proper sequence, with the desired amount of spacing (if any) between tracks.

 

In addition, relative levels between tracks must be established -- unless one doesn't mind having to constantly adjust the volume level as the album progresses from song to song. (A poorly mastered album might have the gentle ballads sounding louder than the all-out screaming rockers.) One of the functions of proper mastering is to make the album a cohesive whole.

 

Final level must be established too, particularly if the mixes went through analog. Any A-D converter I know will exhibit its best performance and lowest distortion when the maximum peak level does not exceed -6 dB. Some colleagues I've spoken with would say -10 or -12 dB. In order to make the most of the delivery medium, the level will often be raised. (I'm not referring to the modern trend of compressing dynamics to the point where even the "low" parts are loud. I'm saying even with full dynamic range preservation, the level of a properly done mix will tend to go up with a proper mastering.)

 

Notice I have not mentioned any EQ or other processing a mastering engineer might utilize to make the most of a set of mixes. I'm just talking about making the best straight transfer of a collection of songs in order to create a unified album.

 

And speaking of EQ, if most audiophiles saw how most (certainly not all) records were made, got to hear what the most popular microphones do to the sound, saw where the microphones were placed, and heard the monitoring on which all the sonic decisions were based, they would understand that most recordings need help.

 

But again, even for the best "flat transfer", if there is no mastering, there will be nothing for the listener to hear.

Send the raw mixes to the replication facility and what will they do? They will send it to their in-house "mastering room" to be properly formatted for production. If this isn't done, they can't get to the next step.

 

Best regards,

Barry

Soundkeeper Recordings

www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

 

Completely right Barry.

The title of this article should be " interesting BS"

 


Link to comment
Final level must be established too, particularly if the mixes went through analog. Any A-D converter I know will exhibit its best performance and lowest distortion when the maximum peak level does not exceed -6 dB. Some colleagues I've spoken with would say -10 or -12 dB.

 

Many D/A converters also behave the same way. There are some that flatten out at -10 dB, but most have distortion significantly increasing from -10 dB upwards.

 

There's no reason to push converters to the max...

 

Too bad the hi-res version of the two latest albums seem to have low-pass cut at 30 kHz.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
Hi Paul,

 

Mr. Wilson is, erroneously in my view, equating mastering with knob twiddling. (The same view could be taken of any part of the process, including the recording and mixing. It depends who's doing it.)

The fact is, if his recordings weren't mastered, there would be nothing for anyone to hear.

 

In order for CDs to be replicated, the manufacturing facility must receive the project in a certain format. All the tracks must be in their proper sequence, with the desired amount of spacing (if any) between tracks.

 

In addition, relative levels between tracks must be established -- unless one doesn't mind having to constantly adjust the volume level as the album progresses from song to song. (A poorly mastered album might have the gentle ballads sounding louder than the all-out screaming rockers.) One of the functions of proper mastering is to make the album a cohesive whole.

 

Final level must be established too, particularly if the mixes went through analog. Any A-D converter I know will exhibit its best performance and lowest distortion when the maximum peak level does not exceed -6 dB. Some colleagues I've spoken with would say -10 or -12 dB. In order to make the most of the delivery medium, the level will often be raised. (I'm not referring to the modern trend of compressing dynamics to the point where even the "low" parts are loud. I'm saying even with full dynamic range preservation, the level of a properly done mix will tend to go up with a proper mastering.)

 

Notice I have not mentioned any EQ or other processing a mastering engineer might utilize to make the most of a set of mixes. I'm just talking about making the best straight transfer of a collection of songs in order to create a unified album.

 

And speaking of EQ, if most audiophiles saw how most (certainly not all) records were made, got to hear what the most popular microphones do to the sound, saw where the microphones were placed, and heard the monitoring on which all the sonic decisions were based, they would understand that most recordings need help.

 

But again, even for the best "flat transfer", if there is no mastering, there will be nothing for the listener to hear.

Send the raw mixes to the replication facility and what will they do? They will send it to their in-house "mastering room" to be properly formatted for production. If this isn't done, they can't get to the next step.

 

 

I'm pretty sure Mr. Wilson is specifically talking about compression, EQing and any other techniques that mess with the sonic signature he envisioned. Getting track levels and spacing correct along with the correct bit depth/sample rate conversions excluded. I'm quite certain Mr Wilson would consider these common sense tasks that need to happen prior to physical media replication and would have simply done it himself.

Roon Rock->Auralic Aria G2->Schiit Yggdrasil A2->McIntosh C47->McIntosh MC301 Monos->Wilson Audio Sabrinas

Link to comment
I'm pretty sure Mr. Wilson is specifically talking about compression, EQing and any other techniques that mess with the sonic signature he envisioned. Getting track levels and spacing correct along with the correct bit depth/sample rate conversions excluded. I'm quite certain Mr Wilson would consider these common sense tasks that need to happen prior to physical media replication and would have simply done it himself.

 

Yes, seemed pretty clear to me. Get levels and such lined up. Then he wants the mix to sound on the media exactly that way. Doesn't want mastering to polish it in a way that changes the sound. As you wrote, compression EQ etc.

And always keep in mind: Cognitive biases, like seeing optical illusions are a sign of a normally functioning brain. We all have them, it’s nothing to be ashamed about, but it is something that affects our objective evaluation of reality. 

Link to comment
Many D/A converters also behave the same way. There are some that flatten out at -10 dB, but most have distortion significantly increasing from -10 dB upwards.

 

There's no reason to push converters to the max...

 

Too bad the hi-res version of the two latest albums seem to have low-pass cut at 30 kHz.

 

No need to push the converters with the 24 bit format, but if the output format is 16 bits, then there is good reason to normalize the levels so as to make full use of limited resolution on the medium. If a DAC has problems this can easily be fixed up by the use of digital volume control in the playback chain without any loss of resolution (assuming 24 bits of resolution in the DAC).

Link to comment

One of my problems on all audio forums and publications is the tendency to talk about small differences as if they were massive improvements.

 

I understand that people get excited about something that improves the SQ of their results, but I think a lot of times, the change that "lifted the veils and made everything come from a truly black background" is something closer to - "I now hear small details a bit better and a slight edge has been removed from the highs".

 

I think hi-res may be in this category.

 

As far as Steve Wilson, I own a couple of his remasters and am very happy with them. I also think his comments about "no mastering" were made with the assumption that most mastering involves a lot of alteration with "knob fiddling".

 

Thanks for posting this. It is an interesting article.

 

I am of the opinion we do not benefit from more than 48/24. But his idea, that there is a tiny difference (unlike those who call it night and day) is not extreme. 96/24 isn't too crazy for a number of reasons. Mr. Wilson apparently isn't saying more than 96 is of any benefit.

 

I very much applaud his idea to skip mastering. Do your mix, and a flat transfer to the final product. Mastering seems to be very nearly ruining more than 90% of currently released music.

Main listening (small home office):

Main setup: Surge protector +_iFi  AC iPurifiers >Isol-8 Mini sub Axis Power Conditioning+Isolation>QuietPC Low Noise Server>Roon (Audiolense DRC)>Stack Audio Link II>Kii Control>Kii Three >GIK Room Treatments.

Secondary Path: Server with Audiolense RC>RPi4 or analog>Cayin iDAC6 MKII (tube mode) (XLR)>Kii Three .

Bedroom: SBTouch to Cambridge Soundworks Desktop Setup.
Living Room/Kitchen: Ropieee (RPi3b+ with touchscreen) + Schiit Modi3E to a pair of Morel Hogtalare. 

All absolute statements about audio are false :)

Link to comment
Hi Paul,

 

Mr. Wilson is, erroneously in my view, equating mastering with knob twiddling. (The same view could be taken of any part of the process, including the recording and mixing. It depends who's doing it.)

The fact is, if his recordings weren't mastered, there would be nothing for anyone to hear.

 

In order for CDs to be replicated, the manufacturing facility must receive the project in a certain format. All the tracks must be in their proper sequence, with the desired amount of spacing (if any) between tracks.

 

In addition, relative levels between tracks must be established -- unless one doesn't mind having to constantly adjust the volume level as the album progresses from song to song. (A poorly mastered album might have the gentle ballads sounding louder than the all-out screaming rockers.) One of the functions of proper mastering is to make the album a cohesive whole.

 

Final level must be established too, particularly if the mixes went through analog. Any A-D converter I know will exhibit its best performance and lowest distortion when the maximum peak level does not exceed -6 dB. Some colleagues I've spoken with would say -10 or -12 dB. In order to make the most of the delivery medium, the level will often be raised. (I'm not referring to the modern trend of compressing dynamics to the point where even the "low" parts are loud. I'm saying even with full dynamic range preservation, the level of a properly done mix will tend to go up with a proper mastering.)

 

Notice I have not mentioned any EQ or other processing a mastering engineer might utilize to make the most of a set of mixes. I'm just talking about making the best straight transfer of a collection of songs in order to create a unified album.

 

And speaking of EQ, if most audiophiles saw how most (certainly not all) records were made, got to hear what the most popular microphones do to the sound, saw where the microphones were placed, and heard the monitoring on which all the sonic decisions were based, they would understand that most recordings need help.

 

But again, even for the best "flat transfer", if there is no mastering, there will be nothing for the listener to hear.

Send the raw mixes to the replication facility and what will they do? They will send it to their in-house "mastering room" to be properly formatted for production. If this isn't done, they can't get to the next step.

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

 

 

Hi Barry

 

Thanks for your comments and insight. I just thought the article was interesting especially from an advocate of high resolution material. I have to admit I have many of the mixes done by Steven and I like the sound of them, I have not really listened to many of the surround recordings but I think it will be interesting to take a listen to some of them at some point.

 

Cheers

 

Paul

Stereo Source: Auralic Aries + Mytek Brooklyn DAC+

Surround Source: Windows PC

Pre-amp: Mark Levinson ML380s, Anthem D2v

Speakers: ATC SCM50A (L/R/C), C4 (Sub), SCM20-2A (LR,RR)

Link to comment
Many D/A converters also behave the same way. There are some that flatten out at -10 dB, but most have distortion significantly increasing from -10 dB upwards.

 

There's no reason to push converters to the max...

 

Too bad the hi-res version of the two latest albums seem to have low-pass cut at 30 kHz.

 

Hi Miska,

 

Which albums are you referring to?

I wonder why anyone would cut the bottom from any release, much less exclusively with the hi-res version.

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

Link to comment
I'm pretty sure Mr. Wilson is specifically talking about compression, EQing and any other techniques that mess with the sonic signature he envisioned. Getting track levels and spacing correct along with the correct bit depth/sample rate conversions excluded. I'm quite certain Mr Wilson would consider these common sense tasks that need to happen prior to physical media replication and would have simply done it himself.

 

Hi Dr Tone,

 

I would hope you are correct. However, what I took from the article was that he burned the raw mixes to a disc and sent that to the replicator.

 

Mastering does not necessarily change anything about the sound. To equate mastering with messing up the mixes is, in my opinion, to say that one has known some not-so-good mastering engineers. Put another way, I know of many albums that were mastered without compression (I never use it) or EQ or any other alteration of the source. Sometimes the source sounds fine as it is. It still needs to be mastered if it is going to be replicated.

 

I can remember one particular mastering session I did where the client was present and as I was evaluating the album during the first listen, at one point I mentioned that one of the mixes sounded perfect to me as it was. The client asked "So, you're not mastering that one?"

I explained that the decision to *not* turn a knob is as valid a mastering decision as the decision to boost or cut frequency x by x db.

 

So, while compression, EQ or any other processing to alter the sonics, are components that *might* be used in a mastering session, these should not be confused with the process of mastering, which can very often occur without any of the above ever coming out of the tool box. Not wanting the processing and calling it "not mastering" is like not wanting a lot of salt added to a soup and calling it "not cooking".

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

Link to comment
Hi Barry

 

Thanks for your comments and insight. I just thought the article was interesting especially from an advocate of high resolution material. I have to admit I have many of the mixes done by Steven and I like the sound of them, I have not really listened to many of the surround recordings but I think it will be interesting to take a listen to some of them at some point.

 

Cheers

 

Paul

 

 

Hi Paul,

 

To be clear, I only want to point out that mastering is not knob-twiddling - unless one has only experienced not-so-good mastering folks. And I wanted to point out that mastering is not simply the last step in production (where one *might* apply things like EQ), it is also the first step in manufacturing. Without it, we would not have discs (or tapes) to listen to.

 

I can fully understand Mr. Wilson's desire to have the sound he creates during mixdown carried forward to the finished product without alteration.

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

Link to comment
Hi Miska,

 

Which albums are you referring to?

I wonder why anyone would cut the top from any release, much less exclusively with the hi-res version.

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

 

I read that wrong the first time.

Post corrected.

 

Also, I just saw your post in the other thread.

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

Link to comment
Many D/A converters also behave the same way. There are some that flatten out at -10 dB, but most have distortion significantly increasing from -10 dB upwards.

 

I have noticed this too and I thought it was "just me". Is this just your (and mine) opinion and experience or is this actually documented somewhere as a shortcoming of D/A converters? I have always recorded at around -12 dBFS as this gives the best final result based on extensive listening tests, but I was extremely surprised to find this remained the case for playback as well. I can take a 24 bit file of classical music where the levels are where I originally left them when recording, adjust the volumes upwards with respect to a peak value of, say, -0.3 dBFS (so it is absolutely nothing more than increasing the volume in the digital domain and working at 64 bit floating precision) and it sounds worse even though the volume adjustment is mathematically perfect.

 

I have never been able to understand why and decided to trust my ears and live with it. So I would be interested to see if there has been any actual "scientific" testing in this respect. My material simply sounds more truthful to the source if I do not tamper with the levels apart from ensuring the overall volume of the transfer is consistent to allow for issues with the original source and then to only increase peak volumes to around -5.9 dBFS for my 16 bit masters. And I just leave it at that because that preserves the sound quality of the original the best.

Link to comment

Have you read this anecdote by J. Gordon Holt?

 

 

Spokesmen for the CD record manufacturers have claimed "There is no possibility for sonic degradation!" or "Science does not know how to make a better-sounding record," or "The CD sounds exactly like our master tapes," or...Hey, wait a minute. What was that last one? Exactly like the master tapes, eh? And what, pray tell, do your master tapes sound like? Like your CDs, perhaps?

"Exactly like them."

 

Well then, let's examine the pedigrees of those tapes? How were they recorded?

"Why, we just put a bunch of instruments in this great big hall, and we use microphones to convert all those squiggly little sound waves into..."

 

No, no. What microphones do you use to make your master tapes?

"Well, we usually like AKG and Neumann condensers, and sometimes our chief technician brings in his Acme electret because...."

 

Hold it right there! Why don't you use Schoeps or Coles or B&K mikes? They're supposed to produce the most accurate sound.

"Oh, we don't want accuracy; we want each mike to have its own sonic flavor, so we can choose the one that gives the best sound from the instruments it was aimed at."

 

You mean, they weren't just aimed at the whole orchestra?

"Oh dear, no. That's so primitive. Bell Labs was recording that way back in the 1930s. We find it easier to get the kind of sound we want that is—that our market researchers want—if we pick up every instrumental group separately, from closeup, and..."

 

You mean you multimike?

"Of course. Doesn't everybody?"

 

Some don't. But tell me, what do you feed those mikes to? A Neve console, maybe?

"What's a Neve? No, ours was custom-designed by a local firm that services touring rock groups. It's a honey! Twenty four inputs, six independent foldbacks, separate bass and treble controls, artificial reverb on each input. Weighs a ton and a half! Give you an idea how good it is, it cost us $27,000 before installation."

 

Uh-huh. What digital mastering recorder do you use?

"Oh, we don't master in digital. Do you know how expensive that stuff is? Heavens no! We have this old Ampex 350 tubed recorder. Helluva reliable machine. Would you believe, that thing has run for over 13,000 hours and it's never needed any attention at all. How's that for reliability! Oh, the heads need cleaning every few months, but that's normal for any professional recorder."

 

Yes. Well, who converts your tapes to digital?

"Oh, we have that done by the Japanese firm that stamps our CDs. They're really good, y'know? There's this little Japanese gentleman, I think his name's Osha or something like that, but he's got real good ears. He can hear 18kHz! He trims up our tapes before he transfers them to PCM-1601 format—you know, adds a little bass here, puts some bite into those violins—the real thing sounds so dull, you know."

 

So after your programs go through your public-address mixer and equalizer and onto your never-maintained Ampex, they get processed a second time before they become CDs?

"That's right. We put a lot of care into our product, because we want it to be good. After all, that's what CD is all about, isn't it?"

"Science draws the wave, poetry fills it with water" Teixeira de Pascoaes

Link to comment
I have noticed this too and I thought it was "just me". Is this just your (and mine) opinion and experience or is this actually documented somewhere as a shortcoming of D/A converters? I have always recorded at around -12 dBFS as this gives the best final result based on extensive listening tests, but I was extremely surprised to find this remained the case for playback as well. I can take a 24 bit file of classical music where the levels are where I originally left them when recording, adjust the volumes upwards with respect to a peak value of, say, -0.3 dBFS (so it is absolutely nothing more than increasing the volume in the digital domain and working at 64 bit floating precision) and it sounds worse even though the volume adjustment is mathematically perfect.

 

I have never been able to understand why and decided to trust my ears and live with it. So I would be interested to see if there has been any actual "scientific" testing in this respect. My material simply sounds more truthful to the source if I do not tamper with the levels apart from ensuring the overall volume of the transfer is consistent to allow for issues with the original source and then to only increase peak volumes to around -5.9 dBFS for my 16 bit masters. And I just leave it at that because that preserves the sound quality of the original the best.

 

If you change the gain at different parts of the playback chain you will change the noise floor and distortion and this applies to both digital and analog portions of the playback chain. The actual sonic effects will depend on the specific implementation of the components used. Many DACs are designed to have good specifications on paper, especially on the all important S/N ratio. This may even result in digital overload in the filters. Plus the analog output amplifiers are going to have more distortion at high levels, but there will be a S/N benefits.

 

If you understand the tradeoffs involved for your particular equipment you will be able to maximize the performance of your system when playing your record collection. If you use player software that has a digital volume control that would be a start as this will give you at least two volume controls you can work with.

Link to comment
Is this just your (and mine) opinion and experience or is this actually documented somewhere as a shortcoming of D/A converters?

 

It is result of my measurements with the hardware I've been playing with. And you can also check out the level-vs-distortion plots for example HiFi-News publishes. Some other magazines also publish similar measurements for the equipment they review.

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
Which albums are you referring to?

I wonder why anyone would cut the bottom from any release, much less exclusively with the hi-res version.

 

Bottom? No, top...

 

"The Raven That Refused to Sing" and "Hand. Cannot. Erase."

 

See:

http://www.computeraudiophile.com/f14-music-analysis-objective-and-subjective/steven-wilson-hand-cannot-erase-23655/

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment
It is result of my measurements with the hardware I've been playing with. And you can also check out the level-vs-distortion plots for example HiFi-News publishes. Some other magazines also publish similar measurements for the equipment they review.

 

Thanks Miska and Tony for the replies. The analogue side of it is easy for me to understand since I have always been wary of analogue output stages ever since we saw the first CD player with the 2 volt output versus what we had all been used to for years prior (say 0.5 to 1 volt maximum, usually courtesy of the tape deck and tuner). I never saw any good reason for the outputs to suddenly be boosted when we first saw CD (apart from maybe louder is "better" and to preserve as much S/N ratio as possible) and to honest I don't think it has ever really been very good for the sound.

 

That is why I am interested in this because I am curious as to how much the analogue side contributes to this issue and how much the pure digital side does.

 

If I take a sample in my DAW and increase the volume by 10 dB then save it, then take that saved sample and reduce the volume by 10 dB, that sample still nulls out perfectly with the original one, so it is as I say - literally perfect - I am not destroying anything or reducing the quality by doing this. So it then has to be the DAC - whether that be the digital / analogue sections or both.

 

As was pointed out, it does not really matter when dealing with 24 bit material because to leave it as is makes no difference. But it is an issue when I produce my CD masters, since not everyone will understand or appreciate the issues and not everyone has transparent digital attenuation. And in the end, I produce a 16 but master for CD only - not for playback on a computer or other form of digital transport. I don't even know of a CD player that has such controls and most people who spin disks do not have the luxury to alter volume in the digital domain since their systems tend to be of the traditionalist hifi ilk.

 

I have tested my 16 bit CDs on high quality Rega CD players, for example, and the "loss" of quality on the CDs where I have done nothing apart from increasing the volume to a peak of -0.3 dBFS is quite obvious on a revealing system. They sound noticeably better and more transparent if I just leave the original levels where they are. It is a pity because the nulling experiment I explained above shows I have nothing to lose and everything to gain by peak normalising prior to the 16 bit conversion. But the irony is it does not produce better sound when you stick it into a CD player connected to a "traditional" hifi system.

Link to comment

I am of the opinion we do not benefit from more than 48/24. But his idea, that there is a tiny difference (unlike those who call it night and day) is not extreme. 96/24 isn't too crazy for a number of reasons.

 

I was going to reply to this yesterday but got sidetracked. My opinions are fairly close to yours. For final playback I don't really see any dire need to go beyond 24 bit, 48 KHz, however I still consider this to be marginal - right on the cusp of a perceptible and unavoidable loss of quality. Yes, I can hear the differences in a direct comparison between 48 kHz and 96 kHz, but not to the point where it really concerns me unless all I did all day were A/B comparisons.

 

Perhaps 25 years ago, I would have said 48 kHz is not enough (Telarc had right the idea of using 50 KHz). I say it is just "OK" these days, however, because I believe filtering and the understanding of it has improved dramatically over the years to the point where 48 KHz gives us just (barely) enough room to work with, without having a significant audio effect on the actual program material. But not really back in the old days.

 

For example, if I downsample my 96 KHz masters to 48 KHz, I can be reasonably ham-fisted about it (in relative terms) even to the point of often picking, for example, the highest quality default setting with iZotope. This still gives me a very useable result with very little complain about. Try going down to 44.1 KHz (even from an 88.2 KHz master) and all heck breaks lose. It's like trying to hit a target at the firing range with a crook sight and rapidly changing wind direction. I can hit all around the bullseye - sometimes incredibly close. But I can never hit it bang on. There just isn't enough room to work with. Same deal with the 16 bit thing - the losses from both are about as bad as each other so they are basically cumulative in terms of the practical outcome.

 

Mind you, I am talking about esoterics here. If someone put one of my MoFi Gold CD-Rs into their CD player they would probably be amazed by how good it sounds. That is until they heard where it came from...

Link to comment
Thanks Miska and Tony for the replies. The analogue side of it is easy for me to understand since I have always been wary of analogue output stages ever since we saw the first CD player with the 2 volt output versus what we had all been used to for years prior (say 0.5 to 1 volt maximum, usually courtesy of the tape deck and tuner). I never saw any good reason for the outputs to suddenly be boosted when we first saw CD (apart from maybe louder is "better" and to preserve as much S/N ratio as possible) and to honest I don't think it has ever really been very good for the sound.

 

That is why I am interested in this because I am curious as to how much the analogue side contributes to this issue and how much the pure digital side does.

 

If I take a sample in my DAW and increase the volume by 10 dB then save it, then take that saved sample and reduce the volume by 10 dB, that sample still nulls out perfectly with the original one, so it is as I say - literally perfect - I am not destroying anything or reducing the quality by doing this. So it then has to be the DAC - whether that be the digital / analogue sections or both.

 

As was pointed out, it does not really matter when dealing with 24 bit material because to leave it as is makes no difference. But it is an issue when I produce my CD masters, since not everyone will understand or appreciate the issues and not everyone has transparent digital attenuation. And in the end, I produce a 16 but master for CD only - not for playback on a computer or other form of digital transport. I don't even know of a CD player that has such controls and most people who spin disks do not have the luxury to alter volume in the digital domain since their systems tend to be of the traditionalist hifi ilk.

 

I have tested my 16 bit CDs on high quality Rega CD players, for example, and the "loss" of quality on the CDs where I have done nothing apart from increasing the volume to a peak of -0.3 dBFS is quite obvious on a revealing system. They sound noticeably better and more transparent if I just leave the original levels where they are. It is a pity because the nulling experiment I explained above shows I have nothing to lose and everything to gain by peak normalising prior to the 16 bit conversion. But the irony is it does not produce better sound when you stick it into a CD player connected to a "traditional" hifi system.

 

Did you try the boosted CD (+10 or even +0.3) on a crappy boombox? I thought the loudness wars were all about making crappy stereos sound better by being louder.

Analog: Koetsu Rosewood > VPI Aries 3 w/SDS > EAR 834P > EAR 834L: Audiodesk cleaner

Digital Fun: DAS > CAPS v3 w/LPS (JRMC) SOtM USB > Lynx Hilo > EAR 834L

Digital Serious: DAS > CAPS v3 w/LPS (HQPlayer) Ethernet > SMS-100 NAA > Lampi DSD L4 G5 > EAR 834L

Digital Disc: Oppo BDP 95 > EAR 834L

Output: EAR 834L > Xilica XP4080 DSP > Odessey Stratos Mono Extreme > Legacy Aeris

Phones: EAR 834L > Little Dot Mk ii > Senheiser HD 800

Link to comment
Did you try the boosted CD (+10 or even +0.3) on a crappy boombox? I thought the loudness wars were all about making crappy stereos sound better by being louder.

 

Don't forget that loudness wars not only imply using all the digital headroom available, but also compressing the heck out of the material. I would never touch any sort of compressor.

 

Anyway, I've got a bedside Sony CD player / FM Radio in my bedroom and it is worth about $200 new :). The CDs where I have the level maxed (peak -0.3 dBFS) prior to the 16 bit conversion sound better, but that is because louder does tend to sound better on these cheap things (and in the car). Plus the amplifier circuitry used in portable bedside CD players tend to be relatively noisy, so the louder the signal going into them, usually the better so long as you don't end up with obvious overload distortion.

 

On a very revealing audiophile setup, however, the subtle sonic differences between a CD mastered to max out at -10 dBFS versus - 5.9 dBFS versus -0.3 dBFS are actually quite easy to hear, even though the one at -0.3 dBFS may(?) be technically the "best" one.

 

I think Barry once said years back that if you max out at -5.9 dBFS prior to the 16 bit conversion you are using all the available bit depth along with the advantage of not overloading things. Whatever he said (I may have misquoted or misunderstood) I've found this myself in practice, which is why I normalise using peaks only to -5.9 dBFS (on my DAW the peak normalise works precisely the same way as a "straight" volume increase but is easier to use since you don't need to go into the file statistics to find the existing peak value).

 

It's interesting to note that years ago when Chesky starting bring out those audiophile classical CDs (perhaps these were amongst the very first "audiophile" digital products), the levels are actually quite low. And I know opinions vary but I thought they sounded extremely good. Compare that to any classical CD / download - high res or otherwise - these days and they use every little last 0.1 of a dB.

Link to comment
Why Mastering Isn’t Needed In High-Res Audio | Real HD-Audio

 

maybe we should get Steven Wilson, Mark Waldrup and Barry together to see how we could progress the debate ?

 

Hi Paul,

 

I think it strange that some folks don't differentiate between mastering and bad mastering.

Other than that, as I said in posts #3 and #12 in this thread, without mastering, we don't have discs or tapes to hear.

 

And as I said in #12:

Not wanting the processing and calling it "not mastering" is like not wanting a lot of salt added to a soup and calling it "not cooking".

 

I don't really see the debate as mastering is not simply the last production step, it is also the first manufacturing step. If one doesn't take the first step, there will be no subsequent steps and hence, nothing will be manufactured.

 

Let's see. If I want to get home from a remote location, I'll get in my car and head out to the nearest road going in the right direction. Now, if I choose to avoid swerving wildly across lanes and perhaps going right off the road into a ditch, would that mean I'm not driving? ;-}

 

Best regards,

Barry

Soundkeeper Recordings

http://www.soundkeeperrecordings.wordpress.com

Barry Diament Audio

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


×
×
  • Create New...