Bouncing stems question

[christof]

Hi there,
at the moment I have to compose and programm a lot of film trailer scores for Jeff Rona, his assistant wants me to bounce each track (stems) and load it up to the server in LA.
The engineer of Hans Zimmer will mix my tracks together and do a Hollywood-ready mix.

As this guy probably knows his job very well and is very experienced I want my 'dry' tracks to sound as good as possible.
As I allready have my personal mix I have to reset reverb, panning and so on to do the stems.

Should I also reset all volumes from all VI tracks to zero as I did with panning and reverb depth?

Any hint would be welcome as I have one deadline after another and I am fighting against time.

Thanks,

christof 

[Dietz]

Servus Christof,

Here comes the long version! ;-)

1. Turn OFF all "real" FX like reverb, early reflections, delays and so on. Make notes how _you_ think they should sound (long? short? far? near? "classical"? bright? ...) and hand them over to the mixing engineer.

2. If you deliver stems in the sense of "ensemble groups", make sure that all musically relevant balances and volume rides are recorded within the stem. IOW: Leave volume automation ON! No mixing engineer will know exactly how you expect the single elements of your ensembles to interact (neither will s/he have the time to figure out). For example: Your VI-1-sections is made of Appassionata Violins and Solo Violin. Put them together on one stem.

3. Each stem should have proper gain, i.e. peaks around -1 to -10 dBFS (in 24 bit, of course). If there are instruments which have a much lower gain throughout the whole piece, bring them up to a value similar to the ones mentioned before, but make another note for the mixing engineer. Except for sound-design reasons, don't use dynamic processing like compressors or limiters.

4. Apart from instrument- or ensemble-inherent pannings, the stems should be centered (read: full stereo, NOT mono). Make notes how you think the single parts of your piece are meant to be used, if you have a concept in mind which isn't necessarily a "classical" solution (e.g. VI-1 left, VI-2 right, CB center, harp center behind the CB, etc.). Hand the notes over with the other information you gathered already.

5. Typically, your stems could look like this:

VI-1
VI-2
VLA
VC
CB
Horn Ens
Brass Ens 1
Brass Ens 2
Wood Solo 1
Wood Solo 2
....
Timpany
Harp
....
Synth FX Stereo
Synth Percussion

.... and so on.

One important point is to keep dedicated solo-parts separated from "ordinary" ensemble parts, to give the mixing engineer an obvious reason to take special care for them ;-)

6. Don't forget to make detailed notes about the format (e.g. "24 bit / 44.1 kHz / interleaved stereo / *.WAV"), the number of files and a version-number on the storage medium itself (Data-DVD, FTP-Folder, whatever). Give each single file a version-number and the creation-date _within the file-name_, to make sure that the proper files are used in case of an update at a later date. Give clear indication that all files share the same starting point (the preferred case), or make sure to provide a proper Standard-MIDI-File with a tempo-map if this is not the case (definitely _not_ preferred!).

7. Supply a rough-mix with the same starting-point like the stems, for clear reference what has to go where.



This should take care for most possible pitfalls. All the best!

_____________________

/Dietz - Vienna Symphonic Library

[christof]

Thanks again Dietz for your help and support, I really appreciate it.

I will try to follow these rules hoping everything works out fine.

By the way the bluesky system is such a great improvement for my workflow, I love it.

Ciao

christof 

[Dietz]

Glad you like them! ... they _really_ should pay me something for all the recommendations, by now ... ;-D

/Dietz

[Dietz]

A short clarification after re-reading my meassage form above:

@Dietz said:

[...] 2. If you deliver stems in the sense of "ensemble groups", make sure that all musically relevant balances and volume rides are recorded within the stem. IOW: Leave volume automation ON! [...]

... this means: Keep the (maybe automated) balances of instruments _within a stem_ alive, as well as meaningfull volume automation which is actually part of the musical expression --- but of course _don't_ give the mixing engineer bounced stems with wild volume-rides done on the whole group, like a fade-out at the end, or a pushed solo-part.

Hopes this makes sense, now.

/Dietz

[christof]

Regarding my notes to the enineer:

Instead of writing him a bunch of instructions would you think it would be sufficient to let him listen to my provisionally personal mix as kind of reference?

christof 

[Dietz]

If you sit right next to him - yes. If it's only virtual communication - no, not really. Don't see it as "instructions", it's just a set of production notes. He can't (and won't) stick to them slavishly in any case, but it's always good to know what lead somebody to certain solutions.

HTH,

/Dietz

[hetoreyn]

Give him a stereo reference file anyways .. that way if he screws it all up you can blame him and say ... "Well it sounded okay when I did the mix" :D hehehe

[Rob Elliott]

 Great thread and nice advice Dietz.[:)]

I have a related question for you as well.  I am delivering my VSL based score (in stems) as well on a current film project and you mentioned NO compression when we do this.   I have been putting a 'multiband' on the master buss on my orchestration mixes and have thought it has brought 'life' to the mix.  In addition my final plug on the master buss in a limiter to catch any transient spikes.  Do you NOT recommend this practice.

The tone of the VSL instruments are 'colored' just a little but the sparkle and presence added has been worth it to me in the past.

Open to your thinking here.

Thanks in advance.

Again great thread about 'stems' (doing this more and more lately.)

Rob 

[Dietz]

Hello Rob,

as a matter of fact, I use multiband compression a lot in the main output-bus in my mixes. Also, the limiter a as last instance makes much sense.

What I meant to say in the postings above was not to "finalize" single stems ("sub-mixes"), because it will hinder the stems quite a bit from interacting with each other at a later stage of the production. Plus, it takes aways the possibilty of choosing other options or settings later on.

In the digital audio-world, it doesn't really matter if you "tape" a signal with or without effects. Provided you have the same processors or plugins at hand, the result should be the same. (... of course, this is not true in the analogue domain or in an hybrid digital-analogue setup).

*******

For me, it is the beauty of the VSL-concept that you can do all those things, but you are not _forced_ to do so. In other words: If it works for you and me, it's great, and it's even better if other people found other ways of achieving their goals. :-)

/Dietz

[DG]

Not to disagree with any of the above posters, but I would never deliver a totally dry mix of anything to an engineer. Even the most experienced ones have trouble with samples that are recorded the way VSL does theirs. With the traditional way of mixing there is always some dry signal in the mix. I absolutely loath this, so if I am forded to deliver a dry mix, or stem, I don't follow instructions. I place the instruments on a scoring stage, so that there can be nothing naked in the mix. I'm sure that some of you will exclaim some horror at the thought of printing reverb, but as long as you choose an IR (in Altiverb) like Clinton and activate stage positioning for depth there is absolutely no problem in then using this instrument with a send to a longer reverb.

DG

[christof]

Hi Dietz,

could you give us a hint how we can get the most out of multiband compression or limiters?

Or is it just a matter of taste?

I adjust these plug ins just  with my ears, I mean that I play around till it sounds good to me, but may it sound also good on another sound system?

This may be a philosophical matter.

christof 

[Rob Elliott]

 Couldn't agree more Dietz on the 'stems' and master bus points you make.  Thanks for the clarification.  Your knowledge this 'side of the production' is so well appreciated to at least this customer[:D]

One last question (sorry for the slight deviation for the stems topic) - but I have found the Voxengo stuff (Soniformer 2 - 'multiband' ; and Elephant 'limiter' be quite a good 'value' - but I am probably ready for a possible step up with this higher profile project I am currently working on.   What multiband are you using (and limiter for that fact?)

Thanks again for the clarity on the stems issue and mixing.

Rob 

[Dietz]

Christof - talking about compressors is like dancing about architecture ;-) . I can't think of a more powerful tool in audio than a Multiband-Compressor (or - in other words - a dynamic EQ). Understanding single-band compression and a good idea of EQing helps a lot to get the grip on MB-comps.

Rob - Voxengo has a very good reputation when it comes to plugins. Especially the "ELephant" seems to be a highly esteemed tool amongst many people. I have to admit that I didn't try it on anything serious yet, though. For mixing, I tend to rely on a combination of hardware processors (like the t.c. System 6000 with it's MD5-algorithm, or the Finalizer 96) and professional mastering. :-)

For the typical quick fix, Waves' L3 Multimaximiser is a tool I used quite often during the last few years. This is not because of sonic superiority, but simply because I get a good approximation of how a _could_ sound in the end.

HTH,

/Dietz

[Rob Elliott]

 Dietz - there is a L316 maximizer.  Is that the one you were talking about?   Also - this doesn't appear to function as a 'multiband' - right?

Thanks in advance.

Rob 

[Dietz]

No, I'm talking about the L3. Although 5 bands are actually too much ;-)

-> http://www.waves.com/Content.aspx?id=255

HTH,

/Dietz