Thanks...

I did try that, but not with great results. This is a staccato type passage, and though I did try it using alternation with the short bows, they just can't seem to keep up. They are late all the time because there is a little ramp up in the sample....this has been a tough one. Don't know if it can be done.

I did post an example of what I'm trying to do at

www.homepage.mac.com/oceantracks called "repeat example" if anyone cares to check it out and advise...I'm sure it's possible to do this I just don't know how..

Tom

I think it sounds fine!

But that's just my opinion.

Where did you get the vocal samples? It sounds so realistic! [:D]

@DG said:

Firstly select the correct rep instrument, making sure that the speed you want to play at is faster than the speed at which the reps were recorded.

Tom, if your result is not satisfying you (it would be to me, but different people have, thank god, different taste) I'd give it a try and do it against the norm, involving major tool programming time. Load a rep that is slightly faster or a bit more than the passage you are playing. Then set up a program that only contains the starting note, then a program that only contains one middle note and so on, until you set up the last that only has the end note. So one sample per line only triggered - you need to keyswitch for every sample. This is to get the faster played notes to do the passage without them being triggered at the original recorded, faster tempo.

Another thing I'd try if this is not something I would be willing to spend time on: Move all current rep notes a few ms so that they are triggered beforehand, and to compensate that redraw the attack with volume curves in your sequencer. Also possible more tedious if reprogramming the gig-file itself for that with x-fading attack and sample (KingIdiot did that for some GOS strings back then). The only thing I would do is to simply move the notes a bit forward in the arrangment if you think they are too late. The not-so-exact attack at this tempo is only more real to me. Another thing to look at is the volume of the passage in general, as not all notes of the repetitions will be hit at the same volume in reality, so little adjustments with volume could help this passage also.

A variant to metod 1 two paragraphs above is to load the two faster programs or the one you are using now plus a faster one and variate between them if you think you're hearing a machine gun effect. Though not necessary in my opinion as the reps have enough samples and I don't hear that effect at all.

Horse Opera, don't you think it could be "Lena"? [:D]

All the best,
PolarBear

@Horse Opera said:

I think it sounds fine!

But that's just my opinion.

Where did you get the vocal samples? It sounds so realistic! [:D]

OH HEY...No that's the REAL thing, not me[;)] That is the example of what I am TRYING to do....yeah I wish it sounded that good!

Tom

OH! I thought the music was by VSL and the singing was live! My mistake!

-BW

Another method, ignoring the performance rep samples (which take up a lot of memory and require too much rigidity of treatment to achieve non-rigid music, imo):

A fast way to do fast triplets is to use a fast attack sample, three tracks and intelligent copy, paste, and edit. Record quarter notes. Select them. Reduce the length to less than 2/3 (around 30%). Copy, paste at 1/3 note forward (320 ticks) in a second track. Select the same area on the second track, modify the length and velocity level slightly. Paste at 2/3 note forward (640 ticks) in a third track. Select the area on that track, modify the length and velocity level slightly. De-quantize to taste using tempo control, randomization, or shifting note start times for some of the notes. Alter panning very slightly on the three tracks for some variation in sound. Use a similar but different patch on the middle track for more variation. If you use pitch wheel, you could even add a tiny bit of variation in pitch.

If you are writing for strings and planning for down-up-down| up-down-up then consider using 4 tracks with the same strategy as above, and the down patch on tracks one and three and the up on tracks two and four.

Be aware that some instruments "tune up" the sample in the first short part of the recorded sound --- and that first short part is all you get when using a patch as above! So don't use that kind of instrument ...

gugliel,

That's a pretty interesting approach! What would the difference be in loading an alternation patch and using say a stac1, 3s1, and 5s1, or whatever and tweaking the notes the same way? Is there not as much flexiblity and control once the alternation has been recorded?

-BW

You could use the alternation patch, of course. I just like the ability to change a whole group of notes (one third of the passage) by selecting one of three tracks and panning or shifting or pitching or changing dynamics.

The problem with the alternation patches, at least on the strings, is that at very fast tempos they just can't keep up. They lag. And even shifting them...well, I don't know, it doesn't seem to work. There seems to be a slight ramp up in the samples. I know they could be edited I suppose but I'm not real good at the Giga Editor.....[[;)]]

Thanks for the suggestions and I'll keep trying.

I hope that maybe VSL team will sample some triplets at maybe a hundred different BPMs so were are all covered. [[;)]]

Just kidding, just kidding...

TH

Tom,

just to be sure, when you use the rep tool, are you setting the predelay offset on your MIDI track to match the offset on the tool? If the answer is 'no', that might help reduce the lag. If it's 'yes', then maybe another approach is needed...

here's how I'd do it:

for triplets, I'd set up the matrix with 3 or 4 different 3- or 6-note patterns (making sure only one ends with the 'single hit' note - for use on the last set of reps). Switch between these sets often, if the rep line is long, and make sure you end on a 'single hit' (if the last note is on a '1' or '2', you could do the math and use a 4- or 5-note line).

if it's fast, don't play it in. either way, make sure you go back into the edit window of you sequencer and play with the Note Velocity data - a MIDI export from Sibelius will only give you equal velocities. Put extreme-looking accents where they need to be, and if they're too big on playback, draw them smaller!

remember that MIDI delay on the rep track - otherwise that line will play at eg. 20ms behind the rest of your sequence. If the rep tool has 'First Keyswitch Delay' of 20ms, set your MIDI track to -20ms to compensate.

hope this helps?

Peter