Stems?

[astronmr20]

Ive sent a lot of mixes out for mastering, BTW, but I was always reluctant to stem because I felt like the mastering engineers might not have a lot of "practice" with stems, making it a waste of extra time/ $.

[dcollins]

You mix, I master.

Stems for emergencies (or bill padding) only.

And please, they are called "separations."

DC

[astronmr20]

Seperation... as in seperate me from my $ (;

I get it.

[Bob Olhsson]

Provided they are real stems, I can't imagine anybody having the slightest problem. If it's really just an incomplete mix, the mastering engineer is being put in the position of making mix decisions and that's a slippery, expensive slope.

Mastering decisions are about the presentation of the mix. Mixing decisions are about the presentation of the music and the performance.

It's virtually impossible to judge the former until the latter has been decided on and that decision generally requires that people have lived with the mix for some time.

Real stems break the final mix down into stereo or surround components that combine at unity gain into a perfect mix. They have the potential to retain the desired qualities of the mix in the final master better than using overall eq. and dynamics treatment but they also have the potential of destroying subtle relationships between mix elements unless great care is taken in making them. It's not really a technique for beginners.

[Masterer]

send stems as backup [talk to your ME first, if he/she gets real nervous you may have to "rethink"].

They can be helpfull if you are unsure of the mixes or just really paranoid.

Another [possibly even better] approach is to send a few mixes to your ME as a test. Your ME can help evaluate the mixes and send you back some mastered tracks for you to check out. This lets you know you're on the right track [or not] and you can tweek accordingly before you send the rest of the project to your newest bestest buddy - your ME. [He in turn will know that you dig how he is mastering your tracks and everyone can get on the same page]. The only thing you need to do it this way is a little extra time.

[Darius van H]

I've been doing a lot of stem (er, i mean seperation - that reminds me, i must call my hair stylist) mastering recently.

There's basically 2 ways to do it:

1) the way Bob described - stems bounced then lined up at unity.

2) Bring your computer to the session.

....Method 1) has the advantage of being able to process the stems analog...(this method also has the dissadvantage that people can mess up the bouncing, resulting in missing FX on the Himalayan nose flute etc etc).

.....method 2) is cool because you can tweak the problems directly at source, without being limited to the choice of which stems were bounced......(most people mixing ITB don't have a soundcard with multiple digital outs - and the mastering session is not the time to be installing new soundcards and re-assigning busses!!!)

Stem mastering is not for Bob Clearmountain or Serban Ghenea......nuff said.

It is fraught with danger, and i have had a couple of spectacular dissapoinments (where stem mastering just didn't seem to add anything)

But i've also had a far greater number of stem sessions where the results where far superior to anything we possibly could have got doing it normal styleee.

Mixing is mixing, mastering is mastering - agreed.....but we can also be open to new ways of thinking/working in conjunction with the new technologies and the desires of our customers.....etc.

Stem mastering doesn't need to cost any more then normal masterin.....how long does it take line up a few stems or plug in a computer? - if the client wants to get finickety and start tweaking everything, then of course the cost will rise - and then you start mixing - so self control and a firm grip is required.

[bob katz]

Originally Posted by Darius van H

I've been doing a lot of stem [size=1]

Stem mastering doesn't need to cost any more then normal masterin.....

That's the only part of this discussion I have to disagree with. It ALWAYS takes more time, and time is money. But you can get a better product. To elaborate on what Bob Olhsson said, if you need to tweak the vocal up or down after mastering processing, it's a lot easier with stems and more precise than getting an 0.5 dB vocal up mix if you need something different. But I definitely agree mastering should not be an open-ended extension of mixing. UNLESS the client is asking for mix help, and then all hell can break loose.

Other examples: Client is having trouble getting "definition" on the bass instrument, but otherwise his mix is excellent. As Bob O. pointed out, overall eq in this case can be a compromise, but a slight eq on the isolated bass instrument can do it with "no compromise." It's a slippery slope, and I try to audition client's mixes before they come here, make suggestions on what THEY can do better if there are issues, because the end results client can (almost) ALWAYS sound better if the client fixes it in the mix at his end.

Hip Hop. I've had four stems with great success, rhythm, vocal, chorus, and lead instruments. Yes, they represent EXACTLY the client's mix, but it's hard to anticipate what can happen to a dense hip hop mix these days if we're asked to push the level and the mixes were made very "clean" to begin with. The more severe the processing you do in mastering, the more the mix starts to alter and if the vocal needs a hair of EQ or level to come through after aggressive emphasis on the beats or anything in the mix, the stems really do help. Especially if the mixes are really "clean" and the client would like the overall mastering to be a little more "dirty". And stems with hip hop truly simplify the process of creating a capella, TV, and other iso mixes.

And to repeat: YES, it takes more time and money. Figure on the average an additional hour for an album with 2 to four isolated stems. Better results than if it was a full mix? I'd say, 9 times out of 10, yes, but only if the client discusses it beforehand with the mastering engineer, sends samples for pre-evaluation, and everyone is on the same page beforehand. I do not want to make a business out of it (like the guy who invented "separation mastering"---what a marketing piece of hype)---I only suggest it when it truly proves to be the best solution and/or a remix is not possible. I do advocate that mix engineers mix to stems or to "the gang of four" each and every time.... just to save your ass before you move onto another project. Even if the stems are not used (most of the time they are not) it was worth the trouble.

BK

[busytbp]

Actually i do a lot of Stem Masterings and i see no problem with it!

A lot of mixings today, specially here in Germany, are made in Producer based Studios.

The Budgets are spend mostly on promo instead of sending the Artist (or the Beatmaker) to a good studio, where an experienced Engineer take control of the mixing. and as you all know Budget is a big Problem today!

Mixes made on Cubase or Logic, and as you now, a beatmakers always underrate the Vocals in his mix, they now how the Bassdrum or the Snare have to sound, but the Vocals are always a big Problem, specially in the Genre i´m Workin (HipHop).

So on Stems it´s easyer for me to take care on the right Levels, what´s the Problem with Mixes made in Project Studio.

I just had this problem today, where i had to Master a Production, wich was made Homestudiostyle, so after i received the Mixes and took a listening, i decided to talk to the Producer about Stems, that i have a better chance to get everything out of his Songs, if i have control over his Mixbusses.

Long Story, but to make it Short, after i finished and put the Final Masters on Server for him, he was more than satisfied, cause he heard what a Stem Mastering can do to a flat Sounding Mix.

But anyway, I see nothing wrong with Stems at all, plus it´s all about communicating with the artist or producer.

We have a new Generation of Producers comin up and as the last Instance before the CD gettin in the Store we have to take care of the Music and that means we have to deal with Issues like that.

And the money thing... it is easier to explain the PM that Stem-Mastering was envolved to make the record even sound better, as instead of tellin him, that the Mastering Hour took longer cause the mix wasn´t sounding good.

[astronmr20]

Thanks.. all makes good sense and good points all around. I just wanted to get the opinion of some ME's.

Yes.. I did know that stems were unity.. so that the ME would know exactly what your intended mix was originally, but would still have seperated parts.

[Bob Olhsson]

Originally Posted by Masterer

...Another [possibly even better] approach is to send a few mixes to your ME as a test...

I do a lot of work this way. To me, collaboration is by far the best approach.

[busytbp]

Originally Posted by Bob Olhsson

I do a lot of work this way. To me, collaboration is by far the best approach.

Testing is a good one, if the artist or the producer taking time to decide where they wanna go for mastering and havin an eye on the record-companys deadline.

[Darius van H]

Originally Posted by bob katz

That's the only part of this discussion I have to disagree with. It ALWAYS takes more time, and time is money.
BK

Actually you're right - that was a bit of marketing on my part, because i tend to soak up the extra cost of stems myself - for the simple reason that if you do an excellent job for someone with stems, you've got a customer for life!

[Jerry Tubb]

Originally Posted by Bob Olhsson

...If it's really just an incomplete mix, the mastering engineer is being put in the position of making mix decisions and that's a slippery, expensive slope.

That's where I've been for the last two days. Up to 32 tracks of Pro Tools HD placed into sixteen "stems" summed through the Dangerous Music 2-Buss box, then through our mastering chain... making mix tweak and mastering EQ decisions... whew, talk about multi-tasking... think both halves of my brain have turned to jelly. Wasn't my choice, but it's what the producer wanted, didn't trust his monitoring environment, so we did it.

I'm quite happy to be working with "mere" stereo mixes tomorrow.

JT

[RainbowStorm]

Separation mastering is in my opinion a better approach to mastering, but it requires more time and time is money, exactly like Bob said.

I really don't agree about what Bob said about mastering shouldn't be an extension to mixing, I think it should these days and I think it makes sense too. I am fully aware of the evil behind a mixing mastering engineer, that's not my point and I understand your point Bob. That's what lock protection is meant for. My point is that the mastering process, no matter type, should always result in something better rather than something worse by going through a flexible process and I think the mastering engineer becomes unnecessarily limited in what he can do with the material if he is only handled a stereo track or some separations.

"Fix it in the mix" I think has these days become a result of the mastering engineer's limited flexibility in dealing with the material and to some extent a good explanation of mastering work that have failed and some conservative approaches on record production. In my opinion a mixing engineer should not have to prepare the mix composition in any particular way for any process that takes place after it, because that's when I think the real loss would take place. For this reason I don't really like when mastering engineers say pan it like this, compress it like this, set a different effect ratio, record it in this way or that way... It sure makes life easier for the mastering engineers to get an already almost perfect mix to work on, but at the same time the mastering engineers' tasks are relfected onto the mixing engineer, who in fact might strongly disagree about the quality of the mix and the way it was made. Let's say we would have two or more very conflicting mastering situations and follow this kind of approach, that would mean that each mastering engineer would tell the client to do a different remix for a certain final product type which of course would not work if the client would have many different mastering engineers working on the material and they all require different things about the mix.

In the future we will have network driven scenarios, which means a certain type of mastering engineer could easily setup the client's identical setup as the starting point or as a reference point for software or hardware communication. At that moment the mastering engineer's equipment might technically communicate with the gear through open standards like XML to be able to target the source material in a certain way most efficiently. If the source material components then would be isolated from the rest of the world it would soon be very limiting on the usability. This is already the case with the wav format in itself and for that reason it is not a good long term format. In the future the business will need to have full access to single samples coming from a certain mix component no matter when in the consumption process chain that requirement takes place. I believe more in a lock protection or interface approach that enables the mastering engineer to do deeper cuts when necessary.

So overall, I'm sceptical about the "fix it in the mix" approach.

[Bob Olhsson]

I totally disagree that it is "better."

It's a different way that sometimes gives better results when a mixer hasn't got all of the information they want available at the mix session for some reason. If it turns into an extended mix session, the mastering part can suffer. If there's no break between the mixing and mastering sessions, the mix and the master can really suffer. This is why I said it's probably not a good choice for beginners.

Mastering should not be able to "improve" a truly good mix. It just changes the presentation slightly to fit the listening context that the market has determined the recording must dwell in. It's like a picture frame. You don't put dark tinted glass over a great picture but non-glare glass might be just the ticket for some. Shiny glass might be better in a gallery with different lighting.

What a mastering engineer brings to a project is fresh ears. In the words of the late Denny Purcell, "you only get to hear a recording for the first time once!" The further removed a mastering engineer becomes from that first impression and their gut-level response to the mix, the less they have to contribute.

[bob katz]

Originally Posted by Jerry Tubb

That's where I've been for the last two days. Up to 32 tracks of Pro Tools HD placed into sixteen "stems" summed through the Dangerous Music 2-Buss box, then through our mastering chain... making mix tweak and mastering EQ decisions... whew, talk about multi-tasking... think both halves of my brain have turned to jelly. Wasn't my choice, but it's what the producer wanted, didn't trust his monitoring environment, so we did it.

I'm quite happy to be working with "mere" stereo mixes tomorrow.

JT

That was clearly "the mix that didn't happen". In cases like those I've separated it into a "mix day" and a "mastering day." You can't do justice to a production trying to think out of both sides of your head at once.

In the case of hip hop I'm primarily mastering. We get "the sound" with the stems at unity gain. Then we listen carefully, and if we love the sound, but the vocal has gotten a hair buried in the mastering by the beats or the rhythm or melody line, we make slight tweaks. But I never put on my "mix head'. I'm still using my mastering head.

As soon as you have to put on your mix head, stop mastering, it's counterproductive, in my opinion. And then you can concentrate on getting a great mix, unencumbered by issues of loudness and "maximizing" and all that b.s.

If the budget is too tight to separate it into two days, at least take a lunch break. Reduce it (commit it!) to one, two or 4 stems, and then patch in the mastering processing. That would be my recommendation. Plus, if you are mixing in the mastering room, the EQs will be so good and so right that you won't need to do much mastering if any*. You can sell them on that!

* other than levelllng the songs and MAYBE some peak limiting if you have to "maximize it". Or a minor set of eq tweaks to help the songs go together, or to compensate for the effects of any mastering compression. If I've done the mixing in this great room, I definitely need much less time in the mastering stage. You can hear the tiniest EQ changes in a good mastering room, so your EQs and overall tonality is likely to be right. But get away from issues of "is the reverb on the second vocal loud enough?" If that's what you're doing in the mastering stage, something is going deadly wrong. It's time to let go :-)

BK

[Andy Krehm]

Lots of good points on this thread but there is one that I'm surprised no one has mentioned.

It is my observation that the stereo mix (of a good mix!) almost always sounds better than the sum total of stems when re-combined. What people forget is that "elements" of the mix (stems) that are printed separately do not hit the gear at the same volume as the stereo mix and therefore sound a little different when combined. This is especially true in regards to compressors. They especially don't work the same way when the signal going into them is dropped.

Having said that, I was working on a gospel album last week where the producer had sent me stereo mixes and two stems, the lead vocal and the music track including background vocals. As usual, the stereo mix had more qualities that I liked than the combined stems but on one track, where the lead vocal was too loud and too bright, I used the stems with eq and volume adjustments rather than working with the M/S of the stereo file.

So, definitely handy, from time to time, but I don't encourage my clients to use them unless they are very unsure of their mixes or they want to take the time to create them, pay me to load them, just in case.

[airmate]

Originally Posted by RainbowStorm

In the future we will have network driven scenarios, which means a certain type of mastering engineer could easily setup the client's identical setup as the starting point or as a reference point for software or hardware communication. At that moment the mastering engineer's equipment might technically communicate with the gear through open standards like XML to be able to target the source material in a certain way most efficiently. If the source material components then would be isolated from the rest of the world it would soon be very limiting on the usability. This is already the case with the wav format in itself and for that reason it is not a good long term format. In the future the business will need to have full access to single samples coming from a certain mix component no matter when in the consumption process chain that requirement takes place. I believe more in a lock protection or interface approach that enables the mastering engineer to do deeper cuts when necessary.

i would agree that sometimes it would be better to do a new mixdown instead of needing to make heavy adjustments during mastering which always have some negative trade-offs.

but i definitely doubt that technical progress automatically means a progress sound wise. anything that prevents people from making (early) decisions (like in a scenario you describe) can be a VERY bad thing.
what about the old fashioned way: some people with a good monitoring system making educated decisions on what their productions should sound like...?
this has been working a few decades ago, and this should also work in the future.

i'm not at all convinced that it can be a good thing that mastering engineers more and more have to take over mixing jobs.

in my opinion the best mastering still is this kind of mastering where you just have to some sweetening (if any). all major problems should be solved during mixing...

that being said, it's a great thing that mastering facilities get more and more options and tools to do their job.

but think of the effect hd recording vocal comping capabilities, autotune and the likes has made out of the grand art singing used to be a long time ago...
to me it seems the next lost art is going to be the ability to deliver a proper mixdown.

people just get lazy as an effect of all this technology... dfegad

[Bob Olhsson]

Originally Posted by Andy Krehm

It is my observation that the stereo mix (of a good mix!) almost always sounds better than the sum total of stems when re-combined...

Getting a stem mix right is not trivial. When it is right, it does sound exactly the same.

[Darius van H]

Originally Posted by Andy Krehm

It is my observation that the stereo mix (of a good mix!) almost always sounds better than the sum total of stems when re-combined. What people forget is that "elements" of the mix (stems) that are printed separately do not hit the gear at the same volume as the stereo mix and therefore sound a little different when combined. This is especially true in regards to compressors. They especially don't work the same way when the signal going into them is dropped.

when done right, a (ITB) summed stem mix will 100% cancel (with flipped phase) the stereo mix. I've never had a client bring stems where this was the case though......but that's not the fault of stem mastering. Once you start bringing 2-buss processing into the equation, things change, of course. Or using an analog mixer. In my experience, the stem version often sounds better then the stereo mix.

[busytbp]

Originally Posted by Darius van H

In my experience, the stem version often sounds better then the stereo mix.

I totally agree!!!

[johnlink]

Originally Posted by dcollins

And please, they are called "separations."

At http://www.johnvestman.com/separations.htm I just read that separations are different from stems. Whatever they are called, they seem to me to be exactly the same thing. Is that right?

John Link

[masteringhouse]

Originally Posted by johnlink

At Mastering with Separations - The Format of the Future I just read that separations are different from stems. Whatever they are called, they seem to me to be exactly the same thing. Is that right?

John Link

They're the same thing in my book, it's just semantics and what you decide to include in the "submix".

I've had a good deal of success using stems in mastering, especially over trying to fix issues with M/S. One thing that I would recommend however is to ask the mix engineer to use the group sends for the stems, and send these out to the stereo bus for any bus processing then including that stereo mix with the stems.

To me it's a best of all worlds, the processed stereo mix is available along with the stems if needed. Much more versatility than up/down mixes in stereo, and using the original stereo mix as a reference should allow any ME worth his salt the ability to get very close if need be.

That said, the better the mix engineer, the less stems are needed.

[Ben F]

Stems (stereo masters) is the correct term. If you are mastering a track for 5.1 then you are given 6 stems that are mastered and bounced to 6 stereo tracks, then sent for encoding into AC-3, DTS or whatever codec is required.

For mastering, I only allow the client a maximum os 2 stems- Music and vocals. This can make an enormous difference to the finished product. If the mix was done ITB, I generally ask for a vocal stem without EQ/compression, so I can use decent analogue processing on the vocals separately.

If they are having trouble with a mix then I ask for a maximum of 8 stems and I charge for a mix, and use a different studio with a Pro Tools HD system and outboard designed for mixing, that is half the cost per hour. This really enables you to nail all the problems on the head before mastering. It can be really satisfying to take one of your own mixes into mastering and it sounds right! I generally give one of the other mastering engineers a listen just in case I may of missed a problem.

I think mixing challenges you as an engineer, and mastering definately improves your mixing ability. In the end it's the result that counts, whichever way you choose to go about it...

[bob katz]

Stems vs. "separations?" It's all semantics.

Vestman tries to invent the term "separations" to describe "wet stems which when printed at unity gain produce the original mix".

This is the official definition of a "stem" anyway! To describe the alternate, why not use the term "dry stems", or better, "unmixed or partially mixed tracks". Basically, Vestman is trying to point out that in mastering he wants to get proper stems so that if nothing needs to be done to raise or lower the vocal, for example, then it becomes regular mastering with a little more power to fix things that may be altered by the mastering process. All well and good, but why invent a new term?

In film work, a stem is, by definition, a submix which when added together with other submixes at unity gain and with no processing, will produce the intended original mix. Those stems therefore must contain any effects and all panning that were used. That's the definition of a stem, and it is simpy confusing for Vestman to try to invent a new term "separation".

BK

[Andy Krehm]

Originally Posted by Bob Olhsson

Getting a stem mix right is not trivial. When it is right, it does sound exactly the same.

Hi Bob:

Can you elaborate on this? Like I said in my previous post, my experience is that stems that involve bus compression will sound different if printed solo, with no gain change, and then combined at unity.

I had one client who always sent stems and the stereo mix and the stereo mix always sounded better. He was an experienced mixer and there really was no need to send stems anyway.

So, what advice should I give to my clients who want to create stems that won't sound different when combined after the fact?

[johnlink]

Originally Posted by Andy Krehm

So, what advice should I give to my clients who want to create stems that won't sound different when combined after the fact?

I think the answer is to not use any processing on the mix bus.

John Link

[bob katz]

Originally Posted by Andy Krehm

Hi Bob:

Can you elaborate on this? Like I said in my previous post, my experience is that stems that involve bus compression will sound different if printed solo, with no gain change, and then combined at unity.

I'm not Bob, but I am Bob :-). So I'll take a shot at this. Andy's right! If you apply bus compression on the stereo bus then the whole concept of stems is a long shot. Even if you (the mix engineer) consider the stems to be a safety, things aren't going to sound the same when mixed together. I'd suggest "vocal up" and "vocal down", (usually 1/2 dB is good) options. The rest of the typical, including, a capella, instrumental, and TV will not sound the same, but it's a good idea to at least archive them and document what was used on the mix bus as well. Not that you will ever need them, hopefully, but Murphy says if you don't bring your umbrella, it will rain.

[bob katz]

Originally Posted by Ben F

If they are having trouble with a mix then I ask for a maximum of 8 stems and I charge for a mix, and use a different studio with a Pro Tools HD system and outboard designed for mixing, that is half the cost per hour.

An excellent approach. But I might stretch it just a hair further than Ben :-). In discussion with a client, and after hearing their mix, I've sometimes suggested they also split out the bass instrument, for example, if they're having trouble getting it right. Then, of course, it's going down that slippery slope towards mixing. But I can (mentally) handle vocal, instrumental, and bass instrument in the mastering room with the mastering processing without losing focus on the mastering. It usually only takes me about 5 minutes per song to get the bass instrument level and if necessary a separate EQ or compression so I don't lose focus. That's assuming that the client has sent me a bass track that he has already mixed in context but maybe needs a little help with EQ or minor tweak in level.

Then there have been times when the client supplies a special effect, such as radio static, and wants to run it under and between two other tracks. That can usually be handled in the mastering context without losing focus on the mastering.

But when it becomes a concept album, with applause, laughter, or involves extensive editing or segues, I too would move into the mix room. Or at least mix it first in the mastering room without the mastering processing, and leave the stems in the EDL in case we have to touch up (perhaps that's a dangerous thing, individual cases apply).