Combining a live recording with VSL samples

[PierreFunck]

A film student asked me to provide big 50s-Hollywood-style music for his short animation film; there is no budget for a dedicated orchestra recording session. But since I'm studying film music in Zurich, I have the opportunity to make a recording session with a very small orchestra (1.1.1.1 / 1.1.1.0 / 1 perc / 1 harp / 1 piano / 1.1.1.1.1), and I plan to "heighten" the recording with my VSL samples.

I think there are basically two approaches:

(1) Orchestrating for a full orchestra (double or triple wind instruments, lots of divisi in the strings), recording the "most important parts" with the small orchestra at my disposal and then completing the missing parts with VSL;

(2) Making an orchestration that "works" for the small orchestra (single wind instruments, no divisi in the strings), and then just doubling the recording with VSL section samples (e.g. woodwinds à 3, orchestral strings) to yield a "big" orchestra sound.

Has anyone got some experience with this? Which procedure yields better results?

Thanks a lot for your help!
Pitt

[PaulR]

On (1) I would suggest that would be good provided that the placement and reverb on the samples pretty much match up.

On (2) the same as (1) but not quite sure where the big orchestral sound is going to come from.

[Simsy]

Hi

I do this all the time. I find if possible to give any of the solo or single parts to live instruments, for example Flute, oboe etc. Basically woodwind or brass, piano. Then add a few live strings to the VSL samples to add a bit more realism. Finally, after the recording is done, I would mix and run them all trough the same reverbs.

It really will produce a great effect.

Thanks,
SImon

[PolarBear]

I'm not sure if it's practical as I haven't done it myself yet, but if you have access to 1.1.1.1.1 strings it might be a possibility not to record everything in one go, but have perhaps 3 violins instead only 1 playing the 1st violin part and then having them played the 2nd violins part in another go. Being a double bass player myself, I doubt it will enhance realism more but adding in another violin instead of solo double bass and solo viola - if their parts are not exposed that is. (You could overlay some Solo Strings package where you don't have live recordings nevertheless without compromising too much.) That way you could perhaps even have all divisi parts recorded and the sample parts augmented or enhanced in expressiveness where needed. 

The idea here is, that you focus on recording those parts, where samples usually struggle to sound realistic. In that vein I think it's a tough task to mic the piano with the same pristine results as can be found not only in the VSL piano library. Harp could also be substituted well by samples if not exposed in the composition. So maybe getting another wood or even better brass instrument in for that may also be worth concerning.

Compositionally I'd start with what the piece is asking for, if you want it to be fully orchestrated you should also compose for a full orchestra... it's a balance thing as you would compose differently for chamber ensembles than for full sections.

Just an idea, maybe not even worth the 2 cent...

PolarBear 

[Basstardo]

There's the production type called "mixed recording" that is very common in film music productions and used by many composers out there.

There's the possibility to prepare the "mockup" on your computer and then in studio you record over it the live music. That requires that all musicians need headphones. They have to hear the signal from the sequencer (because of intonation reasons) and a metronome click that it's definately in sync.

Or I heard of a professional film composer that there's also another possibility that is very much used in film music because of budget reasons it fits perfect to your situation.

In case of having only a small orchestra available there's the possibility of making "overdubs" within each instrument groups.

For example if you have 2 contrabasses they can easily record (along to the prepared sequencer tracks) the bass parts and then record the same part again. You have immediately the sound of a 4-head bass group or if you do the same again and again 6 or 8 basses.

The same with the cellos, violas, violins and you've a full string orchestra group that sounds really powerful.  Then you can go on with the brass and the woodwinds. But it is very important this way to start always with the lowest (bass) instruments and then the higher instruments. Otherwise you can get serious problem with the intonation! And don't forget to make click tracks (not only for rhythmical parts, also and especially for free, rubato passages!)

There are 3 advantages in this way of production: You spare money, you can reach (especially with a very rhythmical music) rhythmical precisio, you will have a clean intonation and have lot of possibilities when it comes to mixing (as you have all the tracks seperate and dry).

The 3 negative things: for example the oboist has never really played together with the violin, so in a way it will never have that same kind of "depth" (both as a sound and on the emotional side)  as a real large symphonic orchestra.

Another advantage is that you can decide later easily which sound source you prefer to use or you can mix them (for example: if the intonation of the live cello is not that good for a moment you can easily control everything and for example fade in the sampled cellos, and so on...) That's really an interesing way of working and with a little luck you'll be very happy with the result.

That way of recording is really very much used nowadays (not only in Germany, also in big Hollywood productions).

I wish you all the best and I'm interested to hear your experience!

Best wishes,

Andie 

[Angelo Clematide]

@PierreFunck said:

I think there are basically two approaches:

(1) Orchestrating for a full orchestra (double or triple wind instruments, lots of divisi in the strings), recording the "most important parts" with the small orchestra at my disposal and then completing the missing parts with VSL;

(2) Making an orchestration that "works" for the small orchestra (single wind instruments, no divisi in the strings), and then just doubling the recording with VSL section samples (e.g. woodwinds à 3, orchestral strings) to yield a "big" orchestra sound.

None of the two approaches produce a usable recording.

The first problem is that single violin, viola, cello and Cb do not produce a chorus. Next problem would be that such a recording would miss the usual spread of phantom-source of all instruments across the image and stage width. The third problem would be that the recording of the real ensemble would miss the presence and three dimensional depth. I often have to integrate pre-recorded music into an arrangement and mix, but the way you want to do it does not integrate.

.

[Basstardo]

 The process I tried to explain is explained much more in detail in "Komposition für Film & Fernsehen" ("composition for film and television") by Munich based film composer Norbert Jürgen Schneider (or also called "Enjott" Schneider). I'm not really sure if that book has been translated to english yet, but for all german native speakers it is a very inspiring book and not only interesting for film composers, lot's of information about psychology, receiving music, musicology, etc.

Best wishes,

Andie

[Rob Elliott]

I am just now expanding my recording booth to handle MORE live instruments.   For me to date what has worked it get the arrangment and orchestration RIGHT with VSL and then 'sweeten' the weaker mock up parts with as many live instruments as possible (time and money).

For me it usually has recording live vlns, va and vc.   My expanded room will be able to accomadate 6/2/4.   To avoid phasing issues if parts are 'doubled' I have them play parts with slight different dynamics and 'pan away' to suit.

Not perfect but it is amazing how much this gives realism to you cues.  When we have to deliver a 'package deal' this is the way to go.

Of course I'll do the same for vln and brass solos as well (VSL's WW's rarely demand my replacing them with live.)

Rob 

[Angelo Clematide]

Multitracking of the same parts does not produce any phasing when the tracks are played back together.

To get a real chorus effect when doubling strings, the proceeding is that the string players losen the strings and retune again, this will produce the required effect and the reality of 12 or 18 viiolinists is achieved when only doubling with 6 violinists, of course the same is true to celli, and contrabass, guitar, it also works for wind instrumens, in other words, never track doubles of the same physical instruments without retuning in between.

Later panning, especially narrowing the stereo width on both or only one side, does not produce any comb filter effect when you recorded with the right stereo microphone technique, for example Blumlein.

[DG]

@Another User said:

To get a real chorus effect when doubling strings, the proceeding is that the string players losen the strings and retune again, this will produce the required effect and the reality of 12 or 18 viiolinists is achieved when only doubling with 6 violinists, of course the same is true to celli, and contrabass, guitar, it also works for wind instrumens, in other words, never trackl doubles of the same physical instruments without retuning in between.

Multi-tracking 6 players does not sound the same as 18 players. It sounds bigger than 6 though.

 

In my experience, re-tuning does nothing, as players will automatically try to play in tune and make the same sound as on the previous take.

DG