Gary Lindsay Big Band Demo
OK, here's the Sonar5 file of the big band demo. I had to re-organize a few things in the file because I use a dual monitor setup and had various windows spread across them. This one is saved in single monitor form. First warning: Unless you have a very fast computer you will first need to freeze one or more of the JABB VST instances before attempting to play this Sonar file. I made this demo on a 2.4Ghz PC and I had to freeze one of the instances to complete the mix. The Sonar5 file can be downloaded here: www.garritan.com/JABB/LindsayBBDemos/GLindsay1a.cwp
Without going into a massive amount of detail, here is how I created this demo:
1. I loaded all needed instruments by section. JABB instance #1 = saxophones; #2 = trumpets; #3 = trombones; #4 = rhythm. Drums were split out to give more flexibility in mixing. Having them separated by section allowed me to easily solo sections to check balances and blending.
2. Did all necessary configuration to setup tracks, labeling, MIDI channel assignments, window layout, mixing layout, aux send configuration, reverb settings, etc.
3. Imported the reference mix to the "Mix" audio track at the top (now empty.) I have omitted this track from the Sonar file because it is copyrighted material. It is available as part of Gary Lindsay's book, Jazz Arranging Techniques. Go to www.lindsayjazz.com if you wish to obtain a copy.
4. I used the music on page 217 of Lindsay's book to read as I played in the parts. The music contains the parts for the saxes and brass but not the rhythm section. Consequently, I had to take the rhythm section parts off by ear from the reference recording. I did the bass part first, then the basic drums (primarily a temporary hi hat part.)
5. I played each of the horn parts in separately to get the basic notes and timing. I made no attempt to play sustain pedal (tongue/slur) in real time.
6. Once the notes were recorded I placed some preliminary mod wheel data in the tracks and set about balancing the sections. I A/B'd constantly with the reference mix to get an approximation.
7. Then the real work began. I started with the sustain pedal since one of the most important things is to get the tongue/slur patterns right. These are too intricate and specific to be played in live and if you get it wrong it will be a dead giveaway that samples are being used. I took each part note-by-note using the notation, the reference mix, and my experience as a player to get articulations as close as I could to being idiomatically correct and convincing.
8. The next thing up was the mod wheel data. Again, every part was taken note by note and phrase by phrase as I shaped the dynamic give and take. Lots of A/B comparisons with the reference mix during this process.
9. Pitchbend data was then added where appropriate for saxophone scoops and certain special case applications (like adjusting the tuning of the last note in tenor 2. to reduce beating)
10. Vibrato was then added to key parts like lead trumpet. This step is extremely important to get a believable interpretation. It helps to know how a lead trumpet player would apply vibrato (this isn't part of the notation.) After getting vibrato into the primary instruments (lead trumpet, lead alto, 1st tenor) I added it sparingly to other parts where I thought it necessary. Remember, vibrato is a combination of both speed and intensity shaping. For the purposes of this demo I dealt primarily with vibrato intensity. Speed was set to an approximate value at the beginning of the track for each instrument. No two instruments were set to precisely the same speed. Had I taken more time I would have done detailed note-by-note editing to vibrato speed in each part as well.
11. Detail work by section. I took instruments in combinations of two to see how they were working together making, especially, timing adjustments to the notes as needed. The point is to get them playing together but never perfectly together. This is a good place to mention that quantizing was never used. Once two instruments worked well, together I added a third, and then a fourth, until I was reasonably satisfied with the sound of the section. Sometimes instruments were functioning more for the contribution to the total sound than how they sounded when soloed.
12. Same drill except I now examined combinations of instruments from different sections where they had parts together.
13. Then I moved on to all horn sections together using the reference mix to make balance adjustments at the mod wheel level. Used a little EQ in Sonar to brighten the trombones slightly.
14. Back to the rhythm section. More dictation from the reference recording to get the detailed drum parts in with particular attention to the ride cymbal and kick. The reference mix had the drums pretty low in the mix so it was something of a chore to sort it out. Some of the snare and toms work was very challenging to decipher.
15. Worked on the timing relationships between ride cymbal, hi hat, and bass to get things swinging as much as possible. Made a couple of tempo adjustments. Also, attended to accents and punches.
16. Full ensemble. More balance work but heavy concentration on accents and dynamics compared to the reference mix. Interplay between sections had a high priority.
17. Adjusted reverb and panning to resemble reference mix.
18. Sent preliminary mix to several trusted people who made valuable suggestions, almost all of which I incorporated into the mix. Also made a few final passes at things I thought were weak - mostly having to do with some remaining phasing issues between parts. This was mitigated with more motion in volume and pitch at the trouble spots.
19. I called off editing (it could have gone on much longer) the day shipping of the product was announced. I prepared things for posting and did just that.
What did I leave out by stopping at that point? Well, JABB is very deep in controllers. I just used the basic ones described above and some of those I didn't elaborate with as much detail as I could have. I did not take the time to use VAR1 or VAR2; nor detailed use of length control on a note-specific basis; nor modify tone control (where it could have been valuable, especially at the beginning and end of the piece in the saxophones.) Some features, like keyclicks and airflow, were unnecessary because such things are only rarely heard in an ensemble situation like this. Other than the little bit of EQ in the bones, I didn't really get into any of the sound shaping tools that could have been applied in Sonar. I used just a single send for the reverb using one instance of Sonar's Pantheon reverb. I could have gotten into impulsing each player into a different seat on the soundstage. Lots of things could have been done but I was satisfied that I had reached the point where the potential of the library was clear. The work I have described took place on and off over a period of several days. Some of the work could have gone much faster. This is especially true concerning playing the parts in one by one. My keyboard skills will never win any prizes so that is one of the slower steps of the process for me.
Some general advice: Like most things, your biggest asset is knowledge. Knowing how the parts should be played (and seeking out examples and advice when you aren't sure) will give you the best chance for success. When you see that something isn't working well, try to determine exactly WHAT isn't working well and then attack that specific thing with the tools at hand. I've tried to give users as many tools for shaping sound as the technology can now provide but it is up to the user to apply the tools in effective ways. Above all, try to have some fun with this library. Remember, we didn't design this library to replace real players, we designed it to give composers and students a convenient tool to get a reasonable idea of what a chart will sound like when put in front of real players. Another tool in the compositional arsenal. Like GPO, it is primarily intended as a sketch pad. The fact that in skilled hands (and with time invested) the tools are present to make it sound quite convincing is a bonus.
I intended to keep this fairly brief but I see that, once again, I failed. Gary considers my tendency toward writing entire books (when answering questions) something of a character flaw. He's right. Kind of makes the need to do a more detailed analysis later less compelling . . . Oh well.