The Missing Octave(s) - Audacity Remastering to Restore Tracks

[Chris A]

Sometimes when I run across a particularly good CD to recommend, I post it in this thread.  One of the most iconic American composers of the 20th century was Aaron Copland, whose most noteworthy compositions include Rodeo, Appalachian spring, and Fanfare for the Common Man, all of which are on this album.  This CD also has very high DR (dynamic range) ratings, 14 for the entire album, and is good to listen to without unmastering, requiring only bass restoration...the EQ unmastering curve that I used is shown below. 

 

The conductor is Antal Dorati, who is a personal favorite conductor, and the spark that created the modern Dallas Symphony Orchestra from its restart after WWII.  His interpretations of Copland's famous works are more measured, distinctive, beautifully interpreted (IMO), and in keeping with the composer's intent than the wider known recording by the New York Philharmonic with Leonard Bernstein.   Highly recommended.

 

 

 

[muel]

I'll have to check this out!  It seems I have about 6 or 7 versions of Fanfare and I find that the recording quality simply does not matter at all to me if the interpretation of the music is bad.  Bernstein destroys this music in my opinion!  I'd like to give him the benefit of the doubt and assume he had a dinner engagement later that evening and was in a hurry to rush through the performance.   Blechhh!  

The version I enjoy the most both for the sound quality and "most Copland like" interpretation is Michel Tilson Thomas conducting the San Francisco Symphony on, "Essence of America: Aaron Copland."  It sounds much better and the music is played just like I heard on "Copland conducts Copland."

 

 

[Chris A]

Another album that is probably much more familiar to the audiophile community--Sara K.'s Play on Words:

 

 

This one is much more educational than others that I've demastered.  First off, this is considered an "audiophile-quality disc" by many, and one of the tracks is used by the recording label to show off on their "Ultimate Demonstration Disc".  It's interesting that they would pick a track that is so heavily EQed (although I have to say that it doesn't seem to be significantly altered in any other ways--more on this later.)

 

Secondly, once the tracks are demastered using the example EQ curves that I provide below, the tracks become much, much more ambient and "nightclub-like" in their overall sound.  I would hazard a guess that the studio that was used for recording was more like a night club in terms of its acoustics than a series of dead-acoustic recording booths.  It's interesting to me that Chesky would choose to record in such a live environment if they later chose to "deaden" the sound, as they have for this entire album.

 

Third, the curves that I provide below result in much more midrangey results than the disc--due to the ambient nature of the recordings themselves.  This can be adjusted to taste if you so choose, simply by adding a second EQ curve after you apply the first, anchoring the 20 Hz and 20 kHz end points, then pulling down in the middle of the curve at around 900 Hz by a few dB, then listening again to it's overall sound--until you arrive at something pleasing for your environment.

 

Fourth, I find myself turning up the tracks much higher in order to capture the new sounds that are revealed by the unmastering curves.  I find this part very interesting.  This album is much more engaging than "Jazz at the Pawnshop - Volume 1" (after demastering) and has similar dynamic range.

 

Fifth, and lastly, if you want to know how to remaster recordings that are too ambient as-is, the curves  inverted using the "Invert" button on the bottom of the Equalization box dialog--will show you how to redo the EQ to "enhance" your tracks to achieve a more commercial-sounding flattened resulting sound.  (This is true of all the unmastering EQ curves that I've provided--just hit the "invert" button to see what the mastering engineer actually did to the EQ for the track in order to achieve a "commercial sound".

 

 

Sara K Play on Words EQ curves.XML

 

Chris

 

[Chris A]

One of the thoughts that recently hit me like a rock while unmastering recent Phil Collins and Warren Hill CDs from the 80s and 90s) is that, after unmastering all tracks and while listening to the CDs in their whole afterwards, I realized that it reminded me strongly of the sound that I remembered when listening using computer desk loudspeakers and headphones. Most direct-radiating loudspeakers distort, exhibiting compression distortion and modulation distortion even under very moderate volume levels. I believe that most people have come to ignore that effect while listening to them.

 

When I de-emphasized certain narrow frequency bands. i.e., 1.5, 4, 6 and 10 kHz, which were boosted across the entire tracks relative to the other levels, and restored attenuated bass frequencies below 100 Hz, that the sound presence on the Jubilees was strongly reminiscent of that I remember the tracks sounding like on these poorer performing loudspeakers with the bass tone control knob turned up to about the 2 o'clock position (nominal is at 12 o'clock--a very informal way of stating that their bass performance was exaggerated over "flat" levels).  The nice thing is that the sound of the tracks becomes even better than I've ever heard them, but after unmastering EQ has been applied.

 

Let me explain why this thought (or realization as it may be) is so troubling:

 

1) If it is true that mastering engineers are using poorer performance loudspeakers as the basis of their EQing efforts (i.e., one size fits all" EQing), then a great part of what is happening is that the music tracks themselves are being tuned to only being listened to on these particular models of loudspeakers.

 

2) Much of what I hear on the Jubilees as really poor sounding tracks that need to be corrected--is really just a mastering engineer using something like a pair of Yamaha NS-10Ms in the near field to master the EQ to "sound good", plus making the tracks almost unlistenable when using truly hi-fi loudspeakers.

 

3) The reason for this is that the mastering engineer is relying on this compression and modulation distortion, even at moderate SPLs, to make the tracks sound "good".  Let me say that again: the mastering engineer is relying on you listening to these poorer performing loudspeakers to make his (popular) music tracks sound good.  If you listen to them on loudspeakers that don't distort like that, it won't be a very enjoyable experience.

 

Fortunately, most of what mastering does is reversible using re-EQing to unmaster, except for use of compression algorithms which irretrievably damage tracks' hi-fi potential.  Some of what I see on popular music tracks is a fairly massive amount of limiting (clipping), which can partially and mostly successfully be undone (i.e., reconstructed), losing only the higher frequency harmonics clipped off by the limiting algorithms.  I believe that some of this is due to mastering engineers not wanting the sound to be "flat and loud", thus clipping is the preferred choice.

 

But it still occurs to me that perhaps the mastering engineers are leaving themselves a way out of their "mastering for maximized loudness" by limiting, because this effect can be partially undone later by the (educated) music consumer, but compression algorithms really cannot be undone.  Undoing the clipping eliminates the lions share of introduced odd-order harmonics that occur due to the clipped peaks, and makes the sound much more pleasant and much less fatiguing, in my experience. It also reduces the average loudness level of these "mastered for loudness" tracks to something more reasonable that's more in line with other non-clipped music tracks.

 

Chris

[Chris A]

On 1/31/2015 at 9:51 AM, dwilawyer said:

...mastered by Bernie Grundman, one of the best in the business. Unfortunately, later on some poor guy had to master it for Cd and, based on your curve, it appears he was given what is referred to as the EQd limited master, the vinyl master, the one with the bass cut and the upper end boosted...

I keep running into this: some of the oddest mastering EQ curves that I've unmastered (in terms of "weird" and "sounds not very good before unmastering") was by Mr. Grundman's studio.  The latest in this series of really odd mastering EQ curves is from the 1991 Kiss Under the Moon by Warren Hill.  Compared to this artist's next couple of albums (Devotion and Truth) the sound of this CD is seriously "off" as mastered.

 

I wonder what the deal is?

 

EDIT:  Perhaps Grundman is/was getting the "mixing disasters" that other mastering labs couldn't do much with.  I find that the Grundman-mastered discs that I've unmastered usually aren't the best discs after unmastering.

 

Chris

Edited September 6, 2016 by Chris A

[Schu]

I havent read this entire thread, but on average how long does it take for you to do an entire LP?

Would you consider a commission or two for a fee?

I am getting kind of curious about this whole remastering technique... thoughI have to be honest, NONE of my computer files, no matter what resolution they are, seem to sound anywhere near as good as my good Vinyl or removable media discs.

[Chris A]

It takes me a couple of hours, maximum.  If the same EQ curve was used across the album (usually classical and non-pop/rock/folk), it can take me a few minutes.

 

I really don't need a commission.  Just let me know the following:

 

1) which artist and album name, and exact release number or bar code, and

 

2) some idea of how your system response is set up, i.e., do you boost certain frequencies and attenuate others, or do you use a "house curve"?  If you don't know this, then I can provide neutral tracks that assume a perfectly flat response across the board...like I use at home. 

 

If I already have the disc(s) on hand, I can share the results with you via Google Drive--given that you give me your Google name or gmail address so that I can access you to the directory.  If I don't own the disc(s), and the discs can be obtained fairly inexpensively on Amazon marketplace, I can spring for a couple or three myself...they're usually $4-$8/disc used for the inexpensive ones.  If they're more expensive discs that I don't currently own then you can forward the discs via U.S. mail.  I'll unmaster them and send them back after I'm done. 

 

Chris

[Chris A]

55 minutes ago, Schu said:

I have to be honest, NONE of my computer files, no matter what resolution they are, seem to sound anywhere near as good as my good Vinyl or removable media discs.

 

I can understand this.  Virtually all CDs that I unmaster that were originally released after 1991 have severe clipping, usually 2-4 dB, and sometimes even worse even if you don't see the "red lines" (i.e., where the waveform touches full scale or "1") when opening them.  Most clipped CDs have been further attenuated by a fraction of a dB from maximum in order to not draw too much attention to the fact that they've been clipped heavily.  These CDs look like they've had a haircut (in the waveform view).  Even an average of 1 DR Database point when comparing two versions of the same album (about 1 dB of crest factor difference) from disc-to-disc or disc-to-record is apparent in the resulting unmastered sound quality.

 

Chris

[Marvel]

Joni Mitchell's "Blue" is awesome after being unmastered, Chris. I'm not at home at the moment, but one of the tracks with mostly/all piano. You can hear all the overtones with the sustain pedal down. It all seems mellower (not rolled off), clear as can be. This was on the JBL4311s in the living room. The whole album drew me in...

 

Bruce

[Don Richard]

On ‎9‎/‎5‎/‎2016 at 10:57 AM, Chris A said:

2) Much of what I hear on the Jubilees as really poor sounding tracks that need to be corrected--is really just a mastering engineer using something like a pair of Yamaha NS-10Ms in the near field to master the EQ to "sound good", plus making the tracks almost unlistenable when using truly hi-fi loudspeakers.

 

 Some mastering houses may use small monitors for mastering purposes these days, but a first class mastering studio will have something like this Westlake monitor that they will use for some projects:

 

http://www.westlakeaudio.com/Speakers/Professional_Series/sm-1.html

 

Westlake also has subwoofers and other professional quality goodies available. No association with this company.

[Chris A]

42 minutes ago, Marvel said:

Joni Mitchell's "Blue" is awesome after being unmastered, Chris. I'm not at home at the moment, but one of the tracks with mostly/all piano. You can hear all the overtones with the sustain pedal down. It all seems mellower (not rolled off), clear as can be. This was on the JBL4311s in the living room. The whole album drew me in...

 

Personally I was impressed with the results when I finished working on it, without having to update the results later.  I do realize that there is a large factor associated with the sound of each person's setup but this disc seemed to stand out. 

 

It had a consistent mastering to hide certain frequencies, which is common to virtually all solo guitar/voice and piano/voice discs that I've run across.  It must be some "unwritten rule of mastering" to attenuate 100-300 Hz or even 50-450 Hz for all solo piano or guitar tracks.  I've found it amazing how much more engaging the sound is after unmastering to restore those frequency bands. I've found it always makes a large difference.

 

I've got the keyboard works of J.S. Bach done by Angela Hewitt (Hyperion, 16 discs) and all had basically the same unmastering EQ curve.  After unmastering, I now look forward to listening to them...which is a big difference.

 

Chris

[Schu]

I am going to go back and re read this entire thread over the next couple days... this is very interesting stuff

Maybe I should learn to do this. Is your primary de mastering tool still audacity?

[Chris A]

Yes.

[Chris A]

A link to a "best of" unmastering thread, which is a spin-off:

Chris

[tromprof]

This thread has been around for a while but I just got around to reading it and playing with a few favorite files. A lot of fun and the results aren't subtle.  I tried this on the original Star Wars soundtrack opening and the low brass lines really came out. It is fun to hear the London Symphony players really struggling with the parts (having sweated through them many times myself).

Thanks Chris!

[Chris A]

I've found that virtually all classical/orchestral recordings are attenuated, starting below 450 Hz--which is about tuning fork frequency--since at least the late 1950s.  That's half the piano's fundamental frequencies that are being attenuated.

 

I remember you once saying that recordings (implying classical) never sounded like the real thing.  There's a really big reason for that.  After undoing that attenuation, the recordings cause me to play them again and again spontaneously.  It's like getting a completely new music collection. 

 

And amazingly, most "audiophiles" never have looked at their stereo recordings.  If they did, I think that they'd be horrified.  It's at least 10x worse than they think it is.

 

Chris

[garyrc]

7 hours ago, Chris A said:

I've found that virtually all classical/orchestral recordings are attenuated, starting below 450 Hz--which is about tuning fork frequency--since at least the late 1950s

Does there tend to be more attenuation as the frequencies get lower, i.e., like the mirror image of what the Harmon researchers found that most listeners like?  Or are these recordings attenuated by the same amount at 450 on down to, maybe, 20 Hz ?

 

By about how many dB are these recordings attenuated, at, say, 100 Hz and 40 Hz?

 

[Chris A]

39 minutes ago, garyrc said:

...like the mirror image of what the Harmon researchers found that most listeners like?

Not me.  I like it flat, please...no EQ...just noise removal.

 

39 minutes ago, garyrc said:

Or are these recordings attenuated by the same amount at 450 on down to, maybe, 20 Hz ? By about how many dB are these recordings attenuated, at, say, 100 Hz and 40 Hz?

This is a typical unmastering curve (i.e., the inverse of the mastering curve used on the CD)...if there is one, since these curves vary all over the map.  Note the characteristic ramping from about 450 Hz down to (typically) 40 Hz, or lower to 20 Hz if there is a pipe organ, like in this case.   Argo/Decca  414 042-2 King's College Choir:

 

[Chris A]

Here's another one...this time piano: Bach Keyboard Works (16 CD box set).  This one looks mild...but it isn't.  Listening to the unmastered tracks compared to those on the CDs--is like the difference between the living and the dead.  Note the boosted highs above 10 kHz that needed to be tamed, too:

 

[Chris A]

Another one from this thread: