What is the most effective strategy for using Antescofo with 2 instrumentalists? I am composing a piece for flute and cello and would like to process each instrument independently.
@larry-nelson : Thanks for bringing this question up and sorry for the late response (vacation time here in France!).
We have had this discussion several times with different composers. I believe the question of using Antescofo with 2 performers boils down to what you want to do in the live electronics supposedly as a result of a dialectical relation with the two performers (and probably among themselves).
Technically speaking, it should be possible to :
-
Use two Antescofo objects in the same piece / patch, each with its own score for 2 performers and 2 lines of electronic actions. This will make the 2 Antescofos completely independent. There are possibilities so that the actions resulting from the two performers be somehow synched by a shared block between the two Antescofo objects. This would be an interesting topic to discuss!
-
Use one Antescofo object with interleaving (but not concurrent) following on 2 performers: Or said otherwise, as if you use one Antescofo object, but you alternative between instruments. Your score would therefore be a mix of the two (to simplify) and things would be as before!
I believe you are more inclined towards the first option! Just use two Antescofo objects. As said before, technically it should be no problem at all… Just make sure they are not receiving the same messages (for example scores) 
We will try to make some simple tests for this here to check if everything works. I will get back to you on this within a week.
Hi Nice People,
We set up two instances of AnteScofo for automated accompaniment of two instruments. This was necessary because allowed us to pay attention in some parts of the pieces to one instrument and switch to the other instrument in other parts.
A premiere live concert performance with this setup took place on May 24, 2015 as part of the International Conference of the Historical Keyboard Society of North America.
Sallynee Amawat (violin), Ryan Gallagher (cello), and a MIDI-enabled harpsichord triggered by AnteScofo played Arcangelo Corelli’s Op. 5, No. 3
You can check a video of the performance at https://vimeo.com/128819657. The actual performance starts at 12:40
Cheers!
@vigliensoni… Impressive! Out of curiosity, did you have any issues with latency (due to mechanical actions) with accompaniment with the mechanical harpsichord?
We had issues on this with Disklavier (used in Stroppa’s opera Re Orso with similar procedures) and it was kinda tough to get rid of those delays (which can be known in advance). Curious to know if you did sth smarter!
Halo Arshia,
Stroppa’s Re Orso seems like a very nice and lively piece!
Regarding the delays, indeed a huge problem. We tried two approaches for playing back the MIDI part. The first approach involved embedding MIDI note-on events as events in the AnteScofo score. This method worked well, however there was a permanent delay, depending on the performance some notes could be skipped, and the overall performance was a bit jittery. As a second approach, we use the real-time tempo and the score position calculated by AnteScofo to play back the MIDI part in NoteAbilityPro. The playback of this approach was much more smooth than the previous one and you also can set a compensation delay in milliseconds for playing back the piece (“NA Ahead”). However, the playback head of NAPro has some “inertia”, and it takes some time in reaching certain point of the score, e.g., the beginning of the piece. Because of this, it can’t handle properly extreme changes in tempo, as in a fast rallentando.
All in all, we ended combining the two approaches. For most of the pieces, we used the NAPro approach, but for the beginning and ending sections we relied on triggering the notes within AnteScofo/Max.
Hope it helps!
@vigliensoni Very interesting and challenging!
What you describe as “playback inertia, taking some time in reaching certain point of the score” for the accompaniment is totally normal. That’s why we have been working on Synchronisation Strategies. For example, what you’d need in certain passages is not to just update the accompaniment using ongoing position and tempo but in a way so that it would converge/end with the musician at a certain point in the score and in the future.
This was actually the topic of Christopher Trapani’s Musical Research Residency project in our team, that ended up in the following paper and sketch:
https://hal.inria.fr/hal-01053836/en
After this experiment, we spent some time to generalize and simplify all these different strategies and make them accessible to everyone. I must admit that we yet need to make a thorough documentation and tutorial of all that. @echeveste made one recently that can be found in this post:
http://forumnet.ircam.fr/user-groups/antescofo/forum/topic/synchronization-strategies-examples/
What you need to overcome the “intertia” you’re talking about is the @target synchronization strategy described in @echeveste’s examples.