Ok so I can fumble around with GPO and set the levels ‘intuitively’. But I would like to understand what is the ‘professional’ way to set levels – if there is a such a thing.
My understanding of the problems and issues is this:
First there are what I a could term the 'real instrument' issues. Real instruments can sound different in different parts of their ranges. A flute, for example, cannot play softly in its upper range, different strings of a cello have different timbres, what’s pp on a tuba is less piercing than the same pp on a triangle. I tend to think of these issues in terms of the musical settings pp mf ff etc.. whereas what I shall term the ‘sound engineering’ concepts I think of in numerical controller values. The real instrument issues have an impact on how the figure or phrase 'sits' in the orchestral moment - if it is toward the front or rear or the sound stage.
Like many others, I use GPO with a sequencer (SX2 in my case). I have a template set up in my sequencer so that for each instance of Garritan there are eight audio channels showing in the mixer. Each Audio channel has its own strip in the mixer.
All the midi channels are hidden in the mixer, for two reasons. Firstly using just the audio fader (not the midi-velocity fader) for each track reduces the levels of complexity when trying to control levels. Midi velocity faders are Not suitable for volume control in GPO because volume is controlled by modulation data (on the whole). Secondly using Audio strips in the mixer gives me access to all my VST plugins – if desired. There is a wider range of midi plugins than there are midi plugins. Also, most importantly, I have set Garritan Ambience as a send and can control the wetness or dryness of any individual instrument using the mixer’s send fx in the audio channel strips.
So far so good (I think!).
Now, say that I were to load eight different instruments into an instance of GPO and have all of them on separate midi channels in arrangement window and with different corresponding audio strip in the mixer. Say that I have paired mid channel 1 with stereo out 1&2, midi channel 2&3 with What would be the best 'pro' way to balance orchestral levels?
As I understand it, for the majority of instruments velocity controls brightness and choice of sample. What I think this means is that if you have a velocity setting of 16, then you get a pp sample (i.e. a recording of an instrument being played softly) if you have a setting of 128 then you get the treble forte sample. These samples would have completely different acoustic properties because acoustic instruments sound so very different in different ranges. Additionally the sample is brightened up a bit using sound engineering FX - EQ.
Depending on the instrument, samples of different volumes can pierce the sound stage (come to the front or rear of the foreground/background) unequally. Some instrument’s pierce through no matter what others are blendable and more inconspicuous in the harmony or chord.
Next we have the continuous controller, ‘modulation’ -which controls volume. I think I am correct in saying that this is used to contour individual notes(?) rather than to a track as a whole. I use this to give volume-ic expression to a musical phrase.
Next there is the question of balancing the sounds of each instrument in the group. Here I am confused. I could either use the volume control offered in KONTAKT, or I could use the audio mixer strip in SX. Not sure what to do here.
If one wants to build up a string section from individual instruments, say 8 violins for first violins, it seems that one needs to set the volume data to very low, in order that you don’t get a disproportionately loud group. One also has to bear in mind that the appropriate sample (pp, mf, fff etc) needs to be selected.
What do you guys do with velocity? My first instinct was to set it all to 64, but would this produce the right samples – I don’t think so.
How is the velocity question controlled?
How are the audio levels controlled?
How is this all done please – am I making things to complicated? Has anyone got any methodology?
My brain is starting to hurt.....