Determining the limits of mastering?

[masteringhouse]

Just listened (on computer speakers only) and agree with mshilarious.

I think that a case for MBC could be made here. IMHO the harshness in the upper end of the original is due more to sibilance than an overabundance of 3K. Likewise the bottom end could use a bit of control. The after sounds a bit "muted" due to the 3K reduction and doesn't cut as much as the original. I would try (as a first shot) trying to compress this out rather than reducing it across the board. Then EQ after some of the nasties were tamed.

A while back there was a thread on a mastering contest and I think that both G. and I agreed that it would be better to have a clinic than an example.

G. if you're up for it would you like to try it with this?

 

[mshilarious]

A while back there was a thread on a mastering contest and I think that both G. and I agreed that it would be better to have a clinic than an example.

G. if you're up for it would you like to try it with this?

Yeah, I was organizing that, but I couldn't pull together all the pieces. A contest is really a lot of work, a clinic is probably much easier.

I still have Dogman's track that was supposed to be the entry So maybe if this goes well we can try that next


Anyway, when the download comes up on Soundclick, I'll have a crack at it, maybe more people will too, then Tom can save the best for last

 

[SouthSIDE Glen]

First of all, you guys are absolutely right about the bass. I can handle that one. In fact, I already did; I realized after reading your posts that I sent up an older version of the "during" that was before I pulled back on the bass a bit. That is not an issue with me.

The 3K stuff does remain an issue. I threw a bunch of EQ (paragraphic and graphic, plug and outboard) at it just to get it to where it is now, and it still sounds harsh to me. The high register stuff above 8k is definitely lacking - not suprising, coming from a 20-yr-old tape, and yes, I'd love to bring the cymbals forward, as well as the snare, but there just isn't enough there to grab onto I don't think. Everything between 3k and 8k is so muddled together. Whoever said the drum tracking sucked is right, and I'd personally say the mixing is even worse. To my ears, the vocals and guitars are all sharing too much of the same spectral space and the mix is not only conventional to the boring extreme, but it does noting to try and lift and seperate the instruments at all. The entire upper-mids/lower-his section just sounds crowded and fatiguing to me (you should try listening to an entire CD, it gets REAL fatiguing on the ears), but by the time I can get rid of that feeling, so much of the high mids are scooped out or squashed that it's an obvious lobotomy job that sounds even worse than leaving them in.

As far as the clinic thing, I'm game except for one thing: these aren't my property. I might be able to get permission from the client (he's an old friend of mine, about as old as these tapes), but I'd hate to tell him I'd like to use his stuff as an example of something that sounds crappy . As it is, I really shouldn't have it publically posted as it is; my false justification was that it was in service of getting the job done better, but I probably shouldn't leave it up there permanently.

G.

 

[bblackwood]

No specifics, really, I just stop when it sounds good. Guess it is an experience thing, but I don't really think about it - I just know when it's done.

Sorry, I know this isn't helpful at all.

 

[mshilarious]

Never mind that then, don't want to get you into any trouble.

OK, Soundclick DL is up, this is a much tougher nut to crack than it sounds at first. I tried just about everything once (expect verb ). I ended up with MBC on the bass, and just the slightest hint from 3 to 5 kHz, low ratio, soft knee, no makeup gain, but there are a few offensive drum hits where I thought it helped.

But overall the original track is so bright around 4kHz I couldn't find a way around a big cut there, and a boost at 300Hz and 12kHz as well. The tape noise is broadband, but a denoiser tool seemed to help there without hurting anything else.

After that, the standard compression and limiting added about 1.5dB in loudness, still nothing commercial or anything, but I wasn't really trying for that, actually I can't get it loud at all without it falling apart.

I doubt that's anything you haven't done. I get the feeling the 3-4kHz problem probably should be addressed with effects envelopes, because the snare on the intro sounds fine (other than uneven hits), but when the vocal, hat, snare, and backing vocal are all in on the chorus, it's too much. That much automation is a lot of work though

 

[SouthSIDE Glen]

Never mind that then, don't want to get you into any trouble.

OK, Soundclick DL is up, this is a much tougher nut to crack than it sounds at first. I tried just about everything once (expect verb ). I ended up with MBC on the bass, and just the slightest hint from 3 to 5 kHz, low ratio, soft knee, no makeup gain, but there are a few offensive drum hits where I thought it helped.

But overall the original track is so bright around 4kHz I couldn't find a way around a big cut there, and a boost at 300Hz and 12kHz as well. The tape noise is broadband, but a denoiser tool seemed to help there without hurting anything else.

After that, the standard compression and limiting added about 1.5dB in loudness, still nothing commercial or anything, but I wasn't really trying for that, actually I can't get it loud at all without it falling apart.

I doubt that's anything you haven't done. I get the feeling the 3-4kHz problem probably should be addressed with effects envelopes, because the snare on the intro sounds fine (other than uneven hits), but when the vocal, hat, snare, and backing vocal are all in on the chorus, it's too much. That much automation is a lot of work though

Well, ms, you make me feel good in that your analysis is in almost perfect lockstep with mine .

You're right about the de-noise, I used Sonic Foundry's NR on it and it worked like a charm (as one can hear if they A/B the intro on the two versions).

I also agree on the comp/limiting. My "during" version is pushed a little harder that I would normally like to go, but I brought it to that limit purposely; right about where I have it (give or take a dB or two) is the point where it will really start crumbling if pushed any further, IMHO.

And, yeah, that brightness around 4k is exactly what I'm talking about. I called it at 3-3.5k, but it really is a broadband of mush between 3k and 4.5k (+/-) that is the problem to my ears. I found several resonant peaks around the low 3s which is why I named those, but it's bad up to 4k and a bit beyond as well. I found that some of the most effective "beat backs" were a few narrowband notches in that area combined with a gentle broadband cut. If I went too far with the broadband cut (a la wide-Q or graphic) it scooped too much out. Yet the heavy notching wasn't enough by itself either.

My kingdom for the source tracks ar at least some stems; I'd like to re-mix the whole damn thing and be done with it!

G.

 

[mshilarious]

Well then maybe that's it. I mean the result is much better than the original. Sure, it might get fatiguing three or four tracks into the CD, but I'd have tossed out the original cassette after 30 seconds.

 

[SouthSIDE Glen]

I get the feeling the 3-4kHz problem probably should be addressed with effects envelopes. That much automation is a lot of work though

I've never been afraid of hard work . I'm not sure just wht kind of effects automation you had in mind, though. You got an idea up your sleeve?

G.

 

[mshilarious]

I've never been afraid of hard work . I'm not sure just wht kind of effects automation you had in mind, though. You got an idea up your sleeve?

G.

Yeah, I would try automating a larger cut in that 3-4kHz range, so that it was something like -6dB on the chorus, -4.5dB on the verses, and -3dB during the intro, solo, other parts without vocals. Note that I haven't tried it, so those are wild guesses

 

[FALKEN]

Yeah, I would try automating a larger cut in that 3-4kHz range, so that it was something like -6dB on the chorus, -4.5dB on the verses, and -3dB during the intro, solo, other parts without vocals. Note that I haven't tried it, so those are wild guesses

I love mixing outboard.

 

[pingu]

After trying many processing chains, i have a few that give me the max out of the tunes with the least amount of processing.

I slap the appropriate chain on the song and wiggle.

Ill then print this.

It usually takes about 30 mins.

Now if i like the song and they are paying i might spend more time trying other combos to see if i can gain another 10%.
But you feel that point when you know you are not doing any better.

Also when you first press play i know how far the audio can go.

Sometimes its like oh shit, this mush cant take anything.

Other times its like. ooohhhh this is sweet.

playtime.










ps

i love Southside GLen

 

[LeeRosario]

Any wisdom?

G.

I'm not a mastering engineer by trade (my emphasis being on recording/mix) by any means, but I do have just my two cents to add to the whole thing:


I'm a young guy at 22, but I've been lucky enough to work with really good world class people. Musicians, engineers, producers and mastering engineers. Along the way I have met up with Bob Katz and I've really gathered a perspective on the mastering process that I feel really lucky to have.


Basically it's the two hemispheres of the perceived world: A) What things *should* and we wish to be like and B) What things are really like.

For me personally, I always feel that a perfect world is where everyone knows how to use thier ears and how to tweak out the room for the best sound, and actually knows what they are doing. Then come time to master, the ME can simply enjoy a light workload for a job well done.

Then there's the real world, the reason we need mastering engineers in the first place. I've heard tons of good work in my few years, but I've heard just as much garbage. Surprisingly the garbage comes in mysterious ways. So in that case, a person like a Bob Katz or a Joe Gastwirt are absolutly nessessary in the equation. And not even on such a world class level, because I've met local ME's that do very well.


A person like Bob, who learned under some pretty heavy people (from what I hear), I saw that above anything, experience was the main factor. But you really have to have the observant gene to get by in that world. To think he was first a recording and mix engineer, just like most of us and then found a niche in mastering. Eventually doing it enough to achieve fulltime status.

But you can see how those stepping stones are directly correlated with his vast knowledge of material over time. It makes sense too, knowing hundreds of different preamps, mics, rooms, setups, different personal styles and studio procedures....it's like the mastering engineer is the top of the pyriamid.


Just like a good captain would have been a sailor first, is how a good mastering engineer would of had the extended experience of the studio world. I don't know if the spiritual part of mastering can really shine without the years of embedded production memories and direct experience with that world first.

It's a shame that alot of studio engineers never get that chance to talk one on one with a good mastering engineer. The finalizer. I think it helps in giving that sense of team and less sense of emancipation.

I personally don't see myself becoming an ME, but I enjoy and appreciate what they do for me and for others. So when I think of it like that, it's harmonius to me. I can exsist comfortably out in the studio world interacting with musicians and know that a responsible ME does exsist and will help build off of the team creation. Almost like working for a cool and wise army general rather than a stupid and gun ho one.

I will leave it at this though...if I did get the best explination of being a well rounded engineer, it was to simply "listen to as much music as you can...and of all types".

It does become easier to build off of something that's embedded in your mind rather than make up something from nothing. Which honestly got me nowhere 95% of the time.

I apologize that I have no useful information. I just always enjoy talking about stuff like engineer and mastering engineer relationships.

 

[pingu]

nice post man.

 

[Massive Master]

Hate to jump in late and admittedly didn't read through the thread, so I apologize if I'm off-topic already...

When I'm mixing, I'm trying to tell the mix what to do. When I'm mastering, I let the mix tell me what it wants to do.

The "normal" mode:

I listen to the mix - Digital is great, because I can just jump around the tune. I won't even listen to the whole thing at first - Maybe 30 seconds of it. During those 30 seconds, I try to "visualize" (mentally) what the mix "wants" to sound like - What its potential is - What I imagine it would sound like in its best form. The hard part of course, is trying to keep that visualization realistic - It's easy to imagine what a crappy mix would sound like if it was made with different sounds or what not - This is visualization based on what's presented.

If anything seems odd, anomalies, pops, clicks, etc., I'll probably listen all the way through once and try to fix those right away (I take care of most of that digitally) before I get too familiar with the tune.

Then I shut it off.

Using the visualization as a guide, I set up a chain that I think will get the mix sounding the way it sounds in my head. Whatever pieces, in whatever order, with whatever attack & release times, whatever frequencies cut or boosted by how much and how wide, just about everything except maybe the threshold on the compressor(s) - I just take a guess at those, but I'll have an idea of how much gain reduction I'm looking for from which unit.

Then I turn it back on with that chain engaged and listen again and start tweaking. Make sure the GR is what I was hoping for, maybe wiggle frequencies on the EQ to be sure I'm working on the frequenices I was listening for, etc.

The tweaking is the part that lasts the longest - That last 10% of the tweaking takes 90% of the time. I want to be able to turn that tune on and hear whatever I imagined in my head before. If that's what it sounds like, I make sure the level is somewhere around where the client wants it, tweak again (as that always screws something up) print it and it's done.

If it takes a lot more than that, I'm trying to take the mix someplace where it doesn't want to go... Which may very well be what you're coming up against. If the "harshness" and all is part of what you have to work with, try to quickly (mentally) establish how much you're willing to compromise to get rid of it, and a plan of action to do it. Things that just won't work tend to fail quickly, and you can wind up working yourself into a hole before you know it.

 

[pingu]

Awesome post man.


This shows you know your shit when you have the ability to set up a chain like that after a brief listen.

I think that this method is beyond a lot of us but still very helpful.

 

[SouthSIDE Glen]

Excellent post, John. Definitely some fresh grist for my mental mill. A couple of notes:

...I try to "visualize" (mentally) what the mix "wants" to sound like - What its potential is - What I imagine it would sound like in its best form. The hard part of course, is trying to keep that visualization realistic - It's easy to imagine what a crappy mix would sound like if it was made with different sounds or what not - This is visualization based on what's presented.

This defines exactly the dilemma I'm up against with this project. You may or may not have read these couple of passages from the thread (and that's OK), but I described my situation as one where I can't shake the impression/visualization that I should be able to remove that harshness. I have no illusions on being able to make the mix sound textbook balanced, but I should - it seems to my judgement - at least be able to tame it to the point where it's not fatiguing. But, as mshilarious put it, "it's a much tougher nut to crack" than it sounds at first blush.

My perhaps unsatisfyingly answerable question that birthed this thread was wondering if there were any way other than pure learning from experience to help determine whether a problem in the mix is indeed fixable at the mastering stage and one just hasen't found the right key yet, or whether one is just chasing rainbows.

Now that I look back on that question it seems rather silly, posed in such a way where there is no good answer, perhaps. But what is driving me to pose the question is that I *know* beyond any reasonable doubt that given the tracks I could definitely fix the problem, but with just the mixdown I am having trouble.

This underlines a fact that most of us already know; that just as there is just some stuff in tracking that can't be fixed in the mix, there is stuff in the mix that can't be fixed in mastering. I guess what I am probing for is some wisdom that you guys with more experience might pass along that would help those rest of us determine what is fixable in mastering alone and what is not. It's a tough question to quantify, I'm sure.

...Using the visualization as a guide, I set up a chain that I think will get the mix sounding the way it sounds in my head.

This passage caught me slightly off-guard. I'm somewhat (not entirely) suprised to hear that you'll set up a whole chain at once. I'm not saying that it's right or wrong, but - when mixing, anyway - more often than not I'll have a tendancy to try and break the solution down into more manageable pieces. For example, I usually will not throw EQ and reverb, or EQ and compression on simultaneously; I tend to want to get the EQ settled in first (of course anticipating how the additional links in the chain will affect it) before adding another process. It would seem to me that throwing a whole chain on at once would introduce so many variables at once that I might not know- again, for example - if the bass is too present because the compression is too tight or because the EQ is over-boosted. Maybe it's just a style thing, or maybe you just have more experience and are better and faster at these things (which I wouldn't doubt ).

Just interesting things to ponder...for me anyway.

G.

 

[masteringhouse]

In the case of mastering I agree that you are more likely to visualize a chain based on your goals in regards to the mix.

In the case of mixing however there is a little friend called serendipity that I like to rely on ocassionally in order to make mixes more creative. Getting new sounds often requires experimentation and sometimes you end up with some great stuff by using equipment incorrectly or in ways that you haven't previously used.

For example, I was working with an engineer in England once that told me how he used a delay unit on bass. It wasn't so much that he wanted delay, but the unit was apparently broken and added a very cool type of distortion that wasn't possible otherwise. He just ran accross this by accident one day when the unit f*cked-up. He said that he tried to buy another delay unit for the same purpose that was "technically improved" but it was worthless since it wasn't broken in the same way.

While mastering is a bit more cut and dried (though not completely), I think that mixing is a creative process that can benefit by a bit of trial and error and "thinking outside of the box".

 

[Massive Master]

This passage caught me slightly off-guard. I'm somewhat (not entirely) suprised to hear that you'll set up a whole chain at once.

I know that seems a bit generalized... It just feels right a lot of the time. That's part of the "end game" -- I do know some guys that tend to go with the experimentaion route - Sweeping EQ's, experimenting with this and that... And admittedly, I tend to do some of that during mixing if I don't have a clear end in mind. On the mastering side of things, the very first thing I do is try to get that "end game" in my head. Everything else is just establishing the difference between the source and the destination. In most cases, there's a failry clear path of what it will take to get there.

But there are occasions where it's not so "cut and dried" (as Tom put it) where you have to dig in to some crazy stuff... Your situation is obviously one of those. A lot of "unusual" situations - Noise, crackling, bizarre anomalies, etc., are certainly not as simple as "this needs a narrow cut around 2.5kHz." And it's torture when you hear a decent mix with some goofy crap that you *know* is coming from one lousy track in the mix...

 

[masteringhouse]

I think that some of this also depends on workflow when mastering.

For example if you process on load in (processing when loading the source to the workstation) you will have to commit to your chain earlier than if you process on load out (source already converted in the workstation and processed when cutting final tracks or disc). If you don't commit you will need to redo the load in process and recall all of your earlier parameters for each track, or create multiple versions. If processing on load out, you can save your decision on what chain to use until you are ready to cut the disc, and therefore can "play around" more.

Both have their merits. If mastering from an analog source, I feel it's better to process on load in so that you don't need to go through A/D conversions more times than you need to. If processing from a digital source I usually go load out.