Skip to main content

Mastering Classical Music

I know that there is no real standard for the apparent loudness in popular music mastering (i.e. the loudness wars), but is there any standard in mastering classical music?


JoeH Wed, 08/16/2006 - 09:42
Not in the way you might think; that is, there is nothing "published" or circulated or written about in places like this.

There is an overall approach with classical/acoustic music that is quite different from the rock/pop genres. Although there can and often is multitrack involved in the tracking process, certain types of signal processing are not used, or if they are, they are used sparingly, and only if the results are transparent. (You never want to hear pumping or any kind of limiting on a singer's voice, for example, yet there may be times when something has to be done to tame a wild soprano in an otherwise perfect opera recording.)

Just as the "storage-medium-formerly-known-as-vinyl" affected the entire chain of recording leading up to it, the classical (and often jazz) medium itself strongly affects what can and cannot (or should and should not) be done to the final product. Some things that are routinely done to a modern rock/pop recording would never be applied to classical/jazz - hard compression, deep EQs, phasing, doubling, etc.

Some "Normalizing" of levels is of course done at times to even out track by track consistency, but it's NOT done to simply make everything loudest, again, as with rock/pop mastering. Ditto for mild peak limiting, esp if there are random, rogue transients that are driving down the overall usable level. (My approach has always been: If it does not hurt the overall impact of the music, then it's worth attempting - but ONLY sparingly, with the option to back out if I think the transients are being adversely affected.)

Many clients who listen to both classical AND rock can sometimes be put off by the fact that some classical material doesn't sound as "Loud" as the rock stuff. This is of course deceptive because of the things done to the pop/rock stuff in the name of loudness: "My stuff is louder than yours, therefore it's better!" On the other hand, classical recordings can certainly have peaks and even sustained passages that are just as loud as any rock recording, but they also get extremely quiet at times.

A good mastering engineer understands that his/her audience knows how and where to listen to material like this. (Hint: It's not nec. in the car or at a monster truck rally.) Therefore, crunching levels just to impress someone just doens't fly in classical music. (Classical radio stations of course do some comp/limiting for their listening audience's varied environments, but of course NOTHING like what the other stations on the FM dial do to their stuff.)

One effective way to establish levels in an acoustic/classical recording is to put the loudest track first, if at all possible. That's often a natural occurence anyway, with overtures and symphonies and curtain-raiser material. This tends to let the listener set a comfortable listening level/zone, and puts the overall sonic landscape in perspective throughout the disc. (You don't want them getting up and down to adjust levels all through the CD, either!) When you're mastering a disc full of material this way, it's then a lot easier to work with levels relative to that loudest track. (PPP material can then be left alone, heard the way it SHOULD be, instead of cranked up to silliness.)

Another tricky example is extremely quiet material, such as small quartets and chamber music, esp ancient instruments that do not speak or sound as loud as more modern instruments. Does one release a CD that lives at around -15 to -20 for most of the playing time? (Probably not!) Overall levels may be a lot higher here, but the consistency in dynamic range from track to track stays the same, and there might indeed be an entire track that's insanely quiet compared to the rest of the work, if the music calls for it.

Again, setting an overall level for the work is critical here; peaks may still be set to -.01 db below full zero, or it may be far less than that, but the real trick is in how the material has been recorded in the first place: Solid musicianship, on good instruments, in a good space with great mics and pre's. That's going to make for a good recording (and a more constant RMS value overall) - more than any after-the-fact mastering. An ME cannot go cranking levels, EQ'ing or comp/limiting a classical release because the musicans are lame. It just doens't work that way in classical music, and listeners would know it instantly.

There's a host of things that simply cannot be allowed due to the very nature and purity of the music. Unless it's a live recording, just about everything else is verboten: No chair squeaks, instrument noises, coughs, page turns, car or truck rumble (You'd be surprised how many studios still DO let this in, and how easily it is to hear it in a solo violin passage on good monitors that go all the way down to 20hz), no abrupt gain changes, and nothing obvious in the electronic/DSP realm.

More than anything else, classical mastering is a gentle, organic process that checks for the obvious things, like overall EQ, usuable dynamic range, distortion on any peaks, track sequencing, ID timings, fade outs/tails, etc. Of course, it's an arguably higher standard brought to bear on such naked, exposed music with "no net" of saftey after-the-fact. (Classical producers do NOT walk around saying to themselves: "Oh, this will sound much better after it's been mastered properly." Just doens't work that way in that genre. ;-)

Thankfully (and mabye most importantly?), classical/acoustic/jazz mastering still depends on an ever-decreasing phenomenon in our lives: Quiet contrasted with Loud. Both are just as important to an excellently mastered classical/acoustic/jazz recording.

JoeH Wed, 08/16/2006 - 09:55
Another point that just occured to me: The clients (usually music directors, conductors, composers, instrumentalists etc.) will AMAZE and astound you with what they are able to hear, so don't get any funny ideas about making aribtrary level changes without permission. :twisted:

I have found, to my utter amazement, that most of these folks hear gain changes (associated with triple forte vs triple pianissimo) like no one else in the biz. Radio compression drives them bonkers - to them, their hard-fought dynamics control has been ruined, and taken out of their hands. In the loud world of rock and pop, it's amazing to hear people who still base their entire careers on not just hitting the right notes, but how softly they can make it, esp when turning on a dime in the middle of a killer passage. Take that dynamic range away from them, and you'll get some angry clients (and listeners!) 8-)

Cucco Wed, 08/16/2006 - 12:21
Amen brother Joe! Preach on!

Joe's pretty much nailed it. When mastering for classical, there are no hard and fast rules except - "Don't let your hand in the mix be heard."

Personally, when I mix and master levels on a classical CD, I do so at a set gain level in my monitoring/mastering chain.

For example, I know that, with my Rotel Preamp turned up to "70", I will get a sound which (by the K-Scale) sounds just right - peaks are loud and present, quiet sections are audible to hear things like the bubba-trucks or harley's flying down the road in the background.

For heavy classical (err..romantic and/or 20th century material), the overall level is louder than chamber works. I'm not afraid to leave dead space at the top of the dynamic range for my chamber works and even some choral material. To me, a chamber ensemble sounds wierd when pushed too high in the recording itself.

Personally, I like to listen (using that fixed gain principle) as though I were in the audience and see what sounds natural (or as natural as it could given that you're squeezing hundreds of musicians through 2 speakers.)

As for limiting - I must admit, I do it. Mainly, it's to pull down percussion. Often times, the percussive attacks and transients are incredibly loud and do not represent the actual or perceived volume of the ensemble at that point. So, I have no problem pushing those transients into a limiter and bringing them down into the sound of the mix. Again, if I hear them disappear or even change noticably, it's too much - start over...

That being said, when I record orchestras, I give myself enough headroom that these transients are well within the scope of the medium. I don't like my converters or my DAW doing the limiting for me because it ran out of 1's and 0's.

It's far easier for me to raise the overall level and push the transients into a limiter than it is for me rebuild peaks manually...(Ugggghhhhh...done this before - it sucks!)

Oh, and by limiter, I usually mean (with classical and almost only classical) a digital brick-wall limiter. My goal with the limiter is not to change the sound but merely affect those transients which are significantly higher than the rest of the program material. I've found that most hardware-based limiters affect the sound even when only performing peak limiting with high-threshold ratings.

And one last thing - Joe is absolutely right about the volume changes. It's easy to do some judicious gain-riding and get the dynamics YOU want, but those very often do not agree with the dynamics that the conductor wanted. He or she WILL notice and likely get VERY pissed off. Unless they ask for the album to be "louder" or "fuller" - don't screw with the levels.

Did that kind of answer the question, or did we just make things more confusing??


Tapwater Wed, 08/16/2006 - 18:15
Good discussion

Those are really good answers. More than I would have expected.

I have a client that insists that there is some kind of documented standards for mastering classical music. I tried to explain to him that it doesn't really exist or if it does, no one really adheres to it. I just wanted to make sure that there wasn't something I was missing.

It's good to know there isn't really a loudness war to speak of in classical music. That is so frustrating when I get things to sound great and then the client wants it louder. It's a difficult to explain to them that making it louder will ruin the music.


Thomas W. Bethel Thu, 08/17/2006 - 05:17
Must be an urban myth circulating somewhere out there as I got asked th same question about a year ago when a faculty member from the local college called us to ask the same question. She said one of her colleagues had told her about it. Maybe it is somewhere on the web published by a record company or some stereo/ hifi magazine. To the best of my personal knowledge there is no such "standard".

Good answers by JoeH and Cucco as usual.

JoeH Thu, 08/17/2006 - 10:31
thanks guys, to be honest, I a lot of what I've written is a blend of what I've learned over the years, and then fine-tuned from some very good colleagues comments here: Jeremy, Ben, Tom, David; you guys know who you are, doing some excellent work! ;-)

I too have done the same kind of peak limiting Jeremy describes, and as long as it doens't hurt the sound or its impact, it can do some good in raising the overall level when needed.

I'd love to hear more from your client, Tapwater, about what they think they've heard or read. I'm not being combative, I really DO want to know what's out there, in terms of what the general public is thinking. (Ever read some of those reviews and/or comments in those fringe "audiophile" magazines? Hooo boy... I agree with a lot of the technical stuff in some of those reviews, but some of the subjective commentary on how something is recorded, mixed or mastered are pretty out there. It gets obvious quite quickly that a lot of those folks have no real-world experience in the recording-side of classical music. Its usually the wording that gives it away...)

As for "standards" in acoustic music mastering, I think it's gone through the same changes (analog to digital) along with the rest of the recording industry. The tools may be a little different/newer, but the goals are the same. Just listen to any well-done classical recording from the pre-digital days, and you'll wonder how the hell they did it!

I find it interesting that most recorded classical music (when you go back to the source, that is) sounds almost as good back then as it does now. THe mics used back then were arguably as good then as now, and many (including vintage Neumanns, etc.) are still used today. Classical/Acoustic music recordings really set the standard for everything else that followed, and as anyone can read about with the histories of Les Paul, the Beatles, LA recording scene, etc., there were gradual shifts away from the "pure" approach in all parts of the process (Not that that is good or bad, it just IS).

I feel a little bad for the folks who have to fight in the loudness wars; we have this incredible gear out there now, with specs to die for, great sound, and DAWs that supply the kind of flexiblity and bandwidth we could only dream about back in the analog tape days. From a whisper to a roar, with perfect storage and playback, with no artifacts or coloration of any kind!!!!

What do they end up doing with it? Making MP3's with a dynamic range of about 5 db.

I was at IPR studios in MN/St. Paul two weeks ago visiting a friend who teaches there, and played some DVD tracks for a couple of students. I had some jazz and bluegrass artists on a video DVD with a simple dolby AC3 soundtrack. (which of course had all been done at 24/48 multitrack, mixed for picture, etc.) Their first words were: "Wow! Dynamic range." It was quite telling. :(

Cucco Sat, 09/02/2006 - 17:47
jase wrote: If it has been recorded well then it should not need processing.Classical music is the only music that i have ever cut flat (to vinyl)

Sorry, I have to disagree to a point.

It not only has to be recorded well, it must be performed well also. As well, the room had better be damned good too. All of these things together can make for a perfect, no-post-processing-required disc.

Just a thought.


ghellquist Mon, 08/21/2006 - 14:38
Fascinating. Illuminating. Good writing.

Currently playing around with some recording that will not make CD to see what I can do -- Bartoks Concerto for Orchestra. Starts ppp and passes several ffff peaks. Really something to sharpen your fangs on.

Diversion: I played the bass trombone part, it has a verrry special glissando that really needs a now dead trombone type to be played, a true F Bass trombone. The gliss is from H to F and cannot be truly played on a modern Bb/F/Eb bass which I have. Can be faked, but I am not happy with the result.