The Hidden Fidelity of Classic Albums

[EMRR]

What is this 'unmastering' you speak of?   I work as a recording engineer, and sit in on most mastering sessions.   

[oldtimer]

This should be good.

[muel]

I'll let Chris speak for himself but here you are... some topics for conversation:

 

[Chris A]

The second thread linked above basically starting out compiling some statistics from the DR Database and related mastering engineers that have decided to openly discuss the Loudness War practices that have inhabited the mastering industry for decades, and was made much more invasive starting in 1991 with the widespread introduction of multi-band compressors into mastering.  If you work in that industry, that's an old story.  About 6 months after I started that thread, there was a JAES article on the subject posted by a couple of French guys.  You can decide for yourself if my thread had any influence on that article being researched.

 

The first thread linked was really a voyage of discovery in undoing the really awful EQ and limiting used on most commercially distributed recordings.  It started fairly innocuously, but has led me to examine every CD or Hybrid SACD that I acquire, and 99% of the time, I wind up having to remove the sometimes unspeakably bad EQ that's been applied to them in order to undo that "commercial sound".  I rely on the statistics of the 1/f energy curve using two different views within the freeware tool Audacity to help me gain insights in removing that EQ and returning the tracks back to something that more closely resembles that which was recorded in the studio before "creative mastering" was applied.

 

I'm about 70% complete in demastering every digital recording that I own (about 1400 discs presently--or 10K tracks). The results have been cathartic: sometimes outstanding and sometimes revealing of some really bad train wrecks that were sold as recordings but that the mastering guys had to do something to recover something that could be sold commercially.

 

Chris

[Chris A]

By the way, you should hear the results of that Audacity-based demastering process on a calibrated setup.  Here's a link to a description of the setup I'm using: https://community.klipsch.com/index.php?/profile/26262-chris-a/&tab=field_core_pfield_14

[EMRR]

Without digesting ALL of those threads, just a bit, I'd say there are many many many more factors in play than I see addressed at a glance, at least in the first few pages of each. 

 

A few, maybe covered and I missed them.  Pardon any run-on sentences! 

 

We are a small minority in an overall market which isn't terribly concerned with great sound, but primarily concerned with audible and intelligible sound under any circumstance.  Not to suggest mastering engineers aren't concerned with good sound, but it is one of a group of decision points based on the position of being a service provider and on client demand.  

 

Many recordings were indeed made with a lack of bass because car stereo over radio was always the primary delivery market.  Bass was not necessarily removed, it was never captured in the first place; equalizers cost money and weren't always plentiful so it made sense to cut out the middle man, so to speak.  Many of the popular kick drum and bass microphones way back when deliver a very muted if not inaudible bottom octave.  Voice communications grade mics were popular for acoustic bass capture in pop music at one time particularly because they emphasized upper mid articulation.  They were chosen for precisely this reason; to emphasize 2nd-3rd octave which would excite speakers with higher resonance points while reducing the power required to drive speakers in the average low powered system.  It can be equalized back in as you've found, but consider it may not have been purposefully captured in the first place.   With pop and rock at least, the further back in time you go, the less bottom end there is to be found.  

 

I've had many experiences where an artist said they wanted the bottom end turned up, only to discover they really wanted the low end turned down and the lower mids turned up.  A lot of this has to do with the crappy sound they are used to on the average home system (boom box/Bose wave/etc).  A lot of the bigger picture is based on the leveled sound of radio and TV which people are used to, and acclimate to as good sound without knowing better.    I had one rock/pop client insist I mix through a high pass filter set at 100Hz, the bottom on that record sounded great and they wanted none of it, you can't make that up! 

 

Mastering is the opportunity for a set of unbiased ears to correct any deficiencies in overall response or dynamic range.  When engineering and mixing, you get used to things you hear which may not translate across all platforms.  Mastering is generally the first time a mix is heard on a high quality system in a decent room outside the originating studio.  Contrast that with the fact that many clients no longer attend mix sessions, and approve final mixes based on emailed mp3's.....on their laptop speakers or iPhone earbuds.    

 

I'll step aside for a sec and say I prefer listening to master tapes, but I think it's mainly because I listen to a lot of my own and am always a little less excited by the mastered version's lesser dynamic range on my studio system.  When I play those mastered versions on mid or low level consumer speakers though, they always translate better, assuming the mastering work was well done.  

 

Well done mastering makes a recording translate better on all possible platforms without negatively impacting higher end systems too much.  It obviously helps lesser systems the most, both from frequency correction and maximizing power output of low powered systems.  That power maximization can be well done, or poorly done, or outright overdone.  Many times the shirts at the label dictate an overdone final.   The average well done maximization barely limits the occasional transient peak without leaning into limiting as a constant, while leaving overall crest factor mostly unchanged AND leaves the last fraction of converter headroom untouched so as to pass on headroom to analog circuitry after then converter, which is an area in which many systems have absolutely no headroom beyond 0dBFS.  That maximization is not necessary on highly dynamic high power systems; most people don't have such at their disposal.  It's also the last chance to fix glaring errors, far more common than you might think, and it's frequently done.  There are back catalog reissues that take the approach of 'modern slam' and those which never touch 0dBFS and display no trace of additional limiting.    

 

My studio speakers are powered JBL LSR28P's.  They reproduce whatever you put into them too well.   You can have wild dynamics from a single element and the rest of the picture doesn't change, therefore giving the ear none of the standard distortion clues that a typical consumer amp/speaker will give.  They sound very consistent at all volumes, unlike most systems.  You have to listen on other systems to see what it's really like in the wild.  For contrast, I have a rackmount broadcast monitor with 1/5" oval speakers and a summed subwoofer in a single 1/75" tall package. I can make the majority of mix decisions within that highly limited bandwidth environment before checking full range balance on the JBL's.  If it's a stringed instrument ensemble trading solo phrases amongst 4 people, it's absolutely the best way to match relative levels consistently.   It has phase coherency metering as well, and I look at a set of Dorrough loudness meters occasionally for overall dynamic range.  My mix room acts overall like a 6dB/oct lowpass filter, so there tends to be a lot more low end than is actually there.   Occasionally I use room correction software to flatten the room, but it sounds incredibly weird and absolutely cannot be toggled on and off, you have to drive natural or flattened exclusively within a session as the disorientation of switching is too great.   These perceptual battles are fought in most rooms, and you shouldn't kid yourself that even the majority of high $ mix rooms don't have them on some level.   The realities around having a room holding a large reflective equipment collection and a randomly changing number of human bodies dictates it at very least.  My control room is nothing to write home about, and I've been in 'high end' studios that were worse.  

 

Vinyl:  you can't limit material pressed to vinyl as much as you can to CD, because the physical process of cutting a metal mother won't allow it.  You can look at that in several ways, from historical to modern.   There's a lot of bad sounding modern vinyl out there that was cut from a CD master rather than a separate vinyl pre-master.  

 

The volume wars can as well be blamed on CD-changers and shuffled playlists.  A lot of people really don't know to use auto-leveling options like Sound Check in iTunes, and don't want to ever touch their volume control, as if it's radio, leveled by AGC's and limiters.  

 

Within pop and rock engineering today you find an increasing use of compression, and lots of it, on everything.  Compressors are being run to extremes that make them act as much like equalizers as anything.  Many masters are already hopelessly flattened before the mastering engineer ever hears them.  Compression on the capture side, compression in the mix, overall compression on the mix.  Here mastering guy, make it LOUDER.  

 

Back to mastering in the modern age, I find there are a large number of people claiming to be mastering engineers who have neither speakers nor room at a minimum entry point.  It's frequently a battle to be sure mastering is handled properly by the client, and I occasionally master records I've done (you should never do this, you're already too familiar with it) so the drummer won't do it on his laptop with earbuds in the back of the van on the way to the next gig.   Really.    There are plenty of great mastering providers who will deliver as dynamic and hi-fi product as possible if it's asked of them; it generally is not asked of them, and their revenue stream is dependent on delivering that which is asked for.    I think I'm relatively rare among engineers I know in that I can go home and check things on La Scala's.   Highly limited recordings certainly sound poor on highly dynamic speakers.  

 

If you haven't heard it, check out King's X - Dogman from 1994.  In my mind, that's one of the first rock records with full blossoming bottom end and a decent amount of dynamic range; it's right at the beginning of the runaway CD volume wars.  It's a great reference for unknown audio systems in my work.  The John Paul Jones solo record Zooma and his collaboration with Diamond Gallas both have deep dynamic bass too.   In general after that point you start to see a lot more low frequency content with a lot more overall volume limiting starting to counteract the gains.  Rage Against the Machine 2nd and 3rd records are good examples.  

 

OK, out of breath for now.  

 

[Chris A]

Change of heart...

[Chris A]

Change of heart.

[jason str]

Just read the list of albums on the first page, have not heard some of these in quite some time due to worn out scratchy records and no replacement CD's.

 

Lately i have using Spotify with an old computer and a few days back i installed one of my better computers due to the single core was slowing things down, to my amazement the new computer sounds much, much better either due to the better sound card or much faster processor (not really sure about this).

 

Really is a shame there is not some basic rules to keep recordings more standardized for those not listening to single dashboard speakers in their old jalopy's.

[Chris A]

37 minutes ago, jason str said:

i installed one of my better computers due to the single core was slowing things down, to my amazement the new computer sounds much, much better either due to the better sound card or much faster processor (not really sure about this).

I've experienced this with a Atom-processor netbook.  I had to take off all non-essential processes and remove the security software (anti-virus) to regain fidelity--thus removing the computer from the network and broadband connectivity.  Still having dropouts in the sound stream periodically, however.  Those re-purposed mobile phone processors just don't have a lot of turns on the rubber band, unfortunately.  Guess I'll have to break down and install Linux (yuck).  Lots of work. I've presently got other things to do.

 

I've also noticed that the bitmap music streaming from my older PS3s sound harsher than the Oppo BDP-103.  I attribute this to internal bitstream conversions occurring in the PS3 that cannot be turned off.

 

Chris

[jason str]

25 minutes ago, Chris A said:

I've experienced this with a Atom-processor netbook.  I had to take off all non-essential processes and remove the security software (anti-virus) to regain fidelity--thus removing the computer from the network and broadband connectivity.  Still having dropouts in the sound stream periodically, however.  Those re-purposed mobile phone processors just don't have a lot of turns on the rubber band, unfortunately.  Guess I'll have to break down and install Linux (yuck).  Lots of work. I've presently got other things to do.

 

I've also noticed that the bitmap music streaming from my older PS3s sound harsher than the Oppo BDP-103.  I attribute this to internal bitstream conversions occurring in the PS3 that cannot be turned off.

 

Chris

 

Interesting conclusion Chris, strange how some things you would not expect to make a difference really change the quality of sound.

 

 

[Chris A]

I would classify the above issues in the realm of "unknown unknowns", mostly.  Some suppositions in my last post above based on the process of deduction...not induction, i.e., these issues could be something other than what I mentioned.  If they are, however, those problems would likely be much worse than the ones I mentioned above. 

 

This is the problem with digital: unknown unknowns that cannot be isolated without significant test gear.  Using non-real-time operating system for real-time performance, when used close to the limits of the hardware...is bad karma.  Not being able to see what's behind the curtain (in the case of Sony's PS3) is even worse, because they locked the game when they removed the Linux-loading capability (in order to protect SACDs from being ripped cheaply).

 

HDMI is riddled with DRM-handshaking issues, not hardware issues.  It's all in the hands of the people providing the content: notably Sony.  I don't hold any of their stock...based on principle.  I hate the way they treat their customers, like I hate cable companies...

[EMRR]

2 hours ago, Chris A said:

Change of heart.

 

Ha!  I wrote a huge response then wi-fi died as I hit submit....it's gone.....   Out of time today, I'll try again later as I can.  

 

Anyway, I'm mainly trying to pull back the curtain a bit, neither defend nor advocate for the many things that happen in the music industry.   I think a lot of the causes vary considerably from what they may seem in broad strokes on the other side of the curtain.

 

 

Here's a couple recent self-financed indy releases I worked on in my terrible terrible room, curious what people think of the dynamic range and response of the mastered product.   With the second one, kickstarter funding partners were provided with an 88.2kHz/32 floating bit file lacking final limiting, but including mastering equalization, specific dynamics (some things compressed in a frequency selective manner, some things actually expanded), and relative song volume balancing.  

 

https://abigaildowd.bandcamp.com

 

 https://davidverga.bandcamp.com/releases

[Travis In Austin]

4 hours ago, EMRR said:

Without digesting ALL of those threads, just a bit, I'd say there are many many many more factors in play than I see addressed at a glance, at least in the first few pages of each. 

 

A few, maybe covered and I missed them.  Pardon any run-on sentences! 

 

We are a small minority in an overall market which isn't terribly concerned with great sound, but primarily concerned with audible and intelligible sound under any circumstance.  Not to suggest mastering engineers aren't concerned with good sound, but it is one of a group of decision points based on the position of being a service provider and on client demand.  

 

Many recordings were indeed made with a lack of bass because car stereo over radio was always the primary delivery market.  Bass was not necessarily removed, it was never captured in the first place; equalizers cost money and weren't always plentiful so it made sense to cut out the middle man, so to speak.  Many of the popular kick drum and bass microphones way back when deliver a very muted if not inaudible bottom octave.  Voice communications grade mics were popular for acoustic bass capture in pop music at one time particularly because they emphasized upper mid articulation.  They were chosen for precisely this reason; to emphasize 2nd-3rd octave which would excite speakers with higher resonance points while reducing the power required to drive speakers in the average low powered system.  It can be equalized back in as you've found, but consider it may not have been purposefully captured in the first place.   With pop and rock at least, the further back in time you go, the less bottom end there is to be found.  

 

I've had many experiences where an artist said they wanted the bottom end turned up, only to discover they really wanted the low end turned down and the lower mids turned up.  A lot of this has to do with the crappy sound they are used to on the average home system (boom box/Bose wave/etc).  A lot of the bigger picture is based on the leveled sound of radio and TV which people are used to, and acclimate to as good sound without knowing better.    I had one rock/pop client insist I mix through a high pass filter set at 100Hz, the bottom on that record sounded great and they wanted none of it, you can't make that up! 

 

Mastering is the opportunity for a set of unbiased ears to correct any deficiencies in overall response or dynamic range.  When engineering and mixing, you get used to things you hear which may not translate across all platforms.  Mastering is generally the first time a mix is heard on a high quality system in a decent room outside the originating studio.  Contrast that with the fact that many clients no longer attend mix sessions, and approve final mixes based on emailed mp3's.....on their laptop speakers or iPhone earbuds.    

 

I'll step aside for a sec and say I prefer listening to master tapes, but I think it's mainly because I listen to a lot of my own and am always a little less excited by the mastered version's lesser dynamic range on my studio system.  When I play those mastered versions on mid or low level consumer speakers though, they always translate better, assuming the mastering work was well done.  

 

Well done mastering makes a recording translate better on all possible platforms without negatively impacting higher end systems too much.  It obviously helps lesser systems the most, both from frequency correction and maximizing power output of low powered systems.  That power maximization can be well done, or poorly done, or outright overdone.  Many times the shirts at the label dictate an overdone final.   The average well done maximization barely limits the occasional transient peak without leaning into limiting as a constant, while leaving overall crest factor mostly unchanged AND leaves the last fraction of converter headroom untouched so as to pass on headroom to analog circuitry after then converter, which is an area in which many systems have absolutely no headroom beyond 0dBFS.  That maximization is not necessary on highly dynamic high power systems; most people don't have such at their disposal.  It's also the last chance to fix glaring errors, far more common than you might think, and it's frequently done.  There are back catalog reissues that take the approach of 'modern slam' and those which never touch 0dBFS and display no trace of additional limiting.    

 

My studio speakers are powered JBL LSR28P's.  They reproduce whatever you put into them too well.   You can have wild dynamics from a single element and the rest of the picture doesn't change, therefore giving the ear none of the standard distortion clues that a typical consumer amp/speaker will give.  They sound very consistent at all volumes, unlike most systems.  You have to listen on other systems to see what it's really like in the wild.  For contrast, I have a rackmount broadcast monitor with 1/5" oval speakers and a summed subwoofer in a single 1/75" tall package. I can make the majority of mix decisions within that highly limited bandwidth environment before checking full range balance on the JBL's.  If it's a stringed instrument ensemble trading solo phrases amongst 4 people, it's absolutely the best way to match relative levels consistently.   It has phase coherency metering as well, and I look at a set of Dorrough loudness meters occasionally for overall dynamic range.  My mix room acts overall like a 6dB/oct lowpass filter, so there tends to be a lot more low end than is actually there.   Occasionally I use room correction software to flatten the room, but it sounds incredibly weird and absolutely cannot be toggled on and off, you have to drive natural or flattened exclusively within a session as the disorientation of switching is too great.   These perceptual battles are fought in most rooms, and you shouldn't kid yourself that even the majority of high $ mix rooms don't have them on some level.   The realities around having a room holding a large reflective equipment collection and a randomly changing number of human bodies dictates it at very least.  My control room is nothing to write home about, and I've been in 'high end' studios that were worse.  

 

Vinyl:  you can't limit material pressed to vinyl as much as you can to CD, because the physical process of cutting a metal mother won't allow it.  You can look at that in several ways, from historical to modern.   There's a lot of bad sounding modern vinyl out there that was cut from a CD master rather than a separate vinyl pre-master.  

 

The volume wars can as well be blamed on CD-changers and shuffled playlists.  A lot of people really don't know to use auto-leveling options like Sound Check in iTunes, and don't want to ever touch their volume control, as if it's radio, leveled by AGC's and limiters.  

 

Within pop and rock engineering today you find an increasing use of compression, and lots of it, on everything.  Compressors are being run to extremes that make them act as much like equalizers as anything.  Many masters are already hopelessly flattened before the mastering engineer ever hears them.  Compression on the capture side, compression in the mix, overall compression on the mix.  Here mastering guy, make it LOUDER.  

 

Back to mastering in the modern age, I find there are a large number of people claiming to be mastering engineers who have neither speakers nor room at a minimum entry point.  It's frequently a battle to be sure mastering is handled properly by the client, and I occasionally master records I've done (you should never do this, you're already too familiar with it) so the drummer won't do it on his laptop with earbuds in the back of the van on the way to the next gig.   Really.    There are plenty of great mastering providers who will deliver as dynamic and hi-fi product as possible if it's asked of them; it generally is not asked of them, and their revenue stream is dependent on delivering that which is asked for.    I think I'm relatively rare among engineers I know in that I can go home and check things on La Scala's.   Highly limited recordings certainly sound poor on highly dynamic speakers.  

 

If you haven't heard it, check out King's X - Dogman from 1994.  In my mind, that's one of the first rock records with full blossoming bottom end and a decent amount of dynamic range; it's right at the beginning of the runaway CD volume wars.  It's a great reference for unknown audio systems in my work.  The John Paul Jones solo record Zooma and his collaboration with Diamond Gallas both have deep dynamic bass too.   In general after that point you start to see a lot more low frequency content with a lot more overall volume limiting starting to counteract the gains.  Rage Against the Machine 2nd and 3rd records are good examples.  

 

OK, out of breath for now.  

 

Thank you for posting this response,  I am always interested in hearing about the way things were done in the tape/lp days compared to the modern Era.

 

If you haven't already seen it there is a great interview with Barry Diament where he discusses dynamic rage, the DR websites, being a pioneer in mastering to CD from "master tapes" including EQ'd Limited copies, etc.

 

It can be found here:

 

http://www.audiostream.com/content/qa-barry-diament-soundkeeper-recordings#60JDsjJMLW4isXAS.97

 

 

[EMRR]

Addressing the last few recent comments, there's a lot of variance in the analog output drive section after the DA conversion which makes a huge difference in sound.   The same basic converter block can sound anything from great to terrible depending on the design surrounding it.  

[Travis In Austin]

10 minutes ago, EMRR said:

 

Ha!  I wrote a huge response then wi-fi died as I hit submit....it's gone.....   Out of time today, I'll try again later as I can.  

 

Anyway, I'm mainly trying to pull back the curtain a bit, neither defend nor advocate for the many things that happen in the music industry.   I think a lot of the causes vary considerably from what they may seem in broad strokes on the other side of the curtain.

 

 

Here's a couple recent self-financed indy releases I worked on in my terrible terrible room, curious what people think of the dynamic range and response of the mastered product.   With the second one, kickstarter funding partners were provided with an 88.2kHz/32 floating bit file lacking final limiting, but including mastering equalization, specific dynamics (some things compressed in a frequency selective manner, some things actually expanded), and relative song volume balancing.  

 

https://abigaildowd.bandcamp.com

 

 https://davidverga.bandcamp.com/releases

 

I'm pretty sure I read about Electromagnetic Radiation Recorders in TapeOp, you are that EMRR?

 

Great to have you on the Klipsch Forum

[Chris A]

In the cases I mentioned, it was all digital.  That actually limits the discussions a bit.  The only way digital sounds bad is:

 

1) clocking issues (i.e., "digital jitter"), and,

2) protocol or format changes in the digital data. 

 

Chris

[Chris A]

Welcome to Travis's country club...

[EMRR]

 

45 minutes ago, dwilawyer said:

 

I'm pretty sure I read about Electromagnetic Radiation Recorders in TapeOp, you are that EMRR?

 

Great to have you on the Klipsch Forum

 

Yes, thanks.  

[Chris A]

13 hours ago, jason str said:

...Really is a shame there is not some basic rules to keep recordings more standardized for those not listening to single dashboard speakers in their old jalopy's.

 

It seems that the following is perhaps oddly unpopular to point out among those engrossed in the internal industry practices: the specific demastered recordings that I refer to above are actually enjoyable to hear now that the mastering is at least partially undone.  Beforehand, I found that the experience of listening to these same recordings was always filled with listening dissatisfaction.  These recordings before demastering actually create stress...at least for me personally.  They no longer retained the natural balance harmonic series of instruments and voices--which creative mastering (as a practice) has clearly forgotten.  The enjoyment of listening to the music was lost somewhere in these processes. When listening on good sound reproduction systems that have been carefully set up, what's been done to the vast majority of these most popular recordings listened to by the greatest majority of customers, the results of apparently industry standard mastering practices I've found to be very unwelcome.  One size has never fit all.  Fortunately, some of these practices are undo-able.

 

The bottom line of this thread (...at least when I created it...): some recordings sometimes slip through the typical web of mastering less damaged.  I find that those recordings more often go on to be much bigger hits--even though the people involved in the typical mastering functions might not have realized or acknowledged this connection between much less mastering and greater sales.

 

Chris