• Register
  • Help
Page 1 of 2 12 LastLast
Results 1 to 10 of 11

Topic: KS versus multiple channels for sound families

Share/Bookmark
  1. #1

    KS versus multiple channels for sound families

    I'd appreciate a review my understanding about the pros and cons of using KS (keyswitching) versus multiple channels for sound families.

    As an example, consider Flute Solo, for which there are three GPO samples for Flute Solo Vibtrato, Flute Solo Non-Vibtrato, and Flute Solo Flutter. There is also a Flute Solo KS, which defines pitches for switching between the three flute solo sounds: vibrato, non-vibratro, and flutter.

    Suppose that for an entire flute solo track, I intend to use only two of the three flute solo sounds, say, Vibrator and Non-Vibratro. Suppose also that I have a MIDI sequencer that supports multiple channels per track. (This second assumption assumes away the problem of having to have separate tracks per each distinct channels.) There are two ways I could prepare the flute track: (1) using KS and (2) using multiple channels.

    Using KS has an advantage that I could record keyswitches in real-time at my MIDI keyboard, and instantly hear as I play the changes between flute sole vibrato and non-vibrato. If I compose or arrange by improvising, then this immediate real-time feedback has a real advantage.

    However, since I'm only using two out of the three Flute Solo KS samples, I'm wasting memory and load-time to load the unused third Flute Solo Flutter sound. If in the same track I assigned the vibratro and non-vibtrator to distinct channels associated with slots in the GPO Player, then the Flute Solo Flutter sample would not be unnecessarily loaded.

    Is my above analysis of the pros and cons of KS versus using multiple channels correct and fairly complete?

    Also, do I understand correctly that every KS sound in GPO is made up of separate sounds that I could load individually if I wanted to so as an alternative?

    Thanking in advance for your review,
    Mark Walsen

  2. #2

    Re: KS versus multiple channels for sound families

    Uhm, as long as you aren't going to use the controllable tremolo/trills then you should be okay. If you want to use them, then you have to use the keyswitch section/solo strings. I'm thinking there's another reason to use the keyswitch strings, but I can't remember what it is right now.

    The RAM increase by using the flute keyswitch in negligable if you are going to be loading two of the included patches anyway.

  3. #3

    Re: KS versus multiple channels for sound families

    One additional advantage of doing it the multi-channel-single-track way is that you can start playing your project from whatever position you want to and get the correct sound right away. With the keyswitches you have to take care that you play from the position where the correct KS are set.

    Erik

  4. #4

    Re: KS versus multiple channels for sound families

    Quote Originally Posted by Joseph Burrell
    Uhm, as long as you aren't going to use the controllable tremolo/trills then you should be okay. If you want to use them, then you have to use the keyswitch section/solo strings.
    What are the details about the issue with using tremelos/trills and multi-channels in the same track?

    Note, I'm not proposing that in the same track two voices would be simutaneously planning different samples. At any time location in the track, only one channel would be active. Maybe this provision addresses the issue you have in mind.

    Quote Originally Posted by Joseph Burrell
    The RAM increase by using the flute keyswitch in negligable if you are going to be loading two of the included patches anyway.
    I wasn't aware of this. Aren't the various alternative KS sounds real separately recorded samples? Or, are some/all of GPO's KS sounds simulated with audio wave transformation magic?

    Cheers
    -- Mark

  5. #5

    Re: KS versus multiple channels for sound families

    I've written an MFX plugin (on request from Sean Dunn) to switch MIDI Channel for subsequent notes based upon CC. Sort of like keyswitching, but you also get to leverage the "Use controller lookback" (or whatever) feature for starting a sequence in the middle. Also doesn't put a note in your track if you're printing out sheet music later.

    (This method also has the benefit of allowing me to use a randomizer to change channels via chaos function.)

    But yeah, there are "multi-channels per track" hosts like SONAR where you can highlight a range of notes and tell it to change their channel, and then you just set the default output channel of the track to "none".

    I'm still deciding whether keyswitching versus channel switching is better. I guess part of the decision is up to how you have your Direct From Disk mode tuned.

    If you're interested in this MFX plugin, I hope to have it released soon: gotta find time this weekend...

    - m
    Free MFX and other plugins: http://www.TenCrazy.com/
    Markleford's music: http://www.markleford.com/music/

  6. #6

    Re: KS versus multiple channels for sound families

    Quote Originally Posted by markwa
    Suppose also that I have a MIDI sequencer that supports multiple channels per track.
    Quote Originally Posted by sfiks
    What sequencer is that?
    I'm the developer of MidiNotate Composer (www.notation.com). Currently, it supports only one channel per track. But I'm in the process of implementing multiple-channels-per-track support, as well as specialized keyswitching support, specifically to make GPO users (and Gary) happy.

    To be honest, I don't recall what MIDI sequencers support multiple channels per track. If anyone knows, that would be useful information to provide in this thread.

    Cheers
    -- Mark

  7. #7

    Re: KS versus multiple channels for sound families

    Erik,

    That's a very good point you have made. A GPO-friendly application should "chase" keyswitch events, just as a smart MIDI sequencer must chase MIDI bank select and program changes when the user starts playback in the middle of the song.

    I'm currently implementing GPO-friendly KS support in MidiNotate Composer (www.notation.com). One of the requirements here is that keyswitch events are chased.

    Cheers
    -- Mark

    Quote Originally Posted by efreitag
    One additional advantage of doing it the multi-channel-single-track way is that you can start playing your project from whatever position you want to and get the correct sound right away. With the keyswitches you have to take care that you play from the position where the correct KS are set.

    Erik

  8. #8

    Re: KS versus multiple channels for sound families

    Quote Originally Posted by Markleford
    But yeah, there are "multi-channels per track" hosts like SONAR where you can highlight a range of notes and tell it to change their channel, and then you just set the default output channel of the track to "none".- m
    The multiple-channel-per-track support I'm currently adding to MidiNotate Composer is less general than what you have described for Sonar. In MidiNotate Composer, the track would only play one channel at any given time. There would be "channel change" events, much analogous to MIDI bank select and program changes. Like the keyswitch support I'm implementing, channel change events would be "chased" if you start playback in the middle of the song. The intent of MidiNotate Composer's multiple-channel-per-track support would be to support cases such as a GPO sound change from flute vibrato to non-vibrator, or the classic example of string arco to pizz.

    From a GPO user's perspective, do you think it would be a significant limitation if MidiNotate Composer implemented "channel change events" rather than the more general "assign channel to any note in the track" approach in Sonar?

    The main case where perhaps the "assign channel to any note in the track" option is useful is for drum tracks. For GM, this is not an issue, because MIDI pitches distinguish drum sounds. With GPO, this could be an issue. However, MidiNotate Composer supports one-line staves for drums, where which stave is a different track that can be assigned a distinct MIDI channel.

    Therefore, I'm thinking that it isn't necessary for MidiNotate Composer to have the more general "assign channel to any note in the track" feature that Sonar has.

    Cheers
    -- Mark

  9. #9

    Re: KS versus multiple channels for sound families

    Mark, you misunderstood. The programmable trills/tremolo are only available in the keyswitch string instruments.

    As to the flute and other keyswitch instruments...

    The flute has three separately recorded sample sets. Vibrato, Non-Vibrato, and Flutter. These are real, recorded samples. However, GPO is designed in such a way that the footprint of these instruments is relatively small (filtering, blah blah.) Thus, you don't really need to be concerned about the effect of loading 2 patches versus the one keyswitch patch, since impact on RAM of these particular patches is negligable in this scenario. Now this doesn't hold true universally. Best way to know is to look at the individual RAM requirements for the instruments you need (just look on the Kontakt Player when that instrument is loaded and selected) and compare that to the keyswitch version.

  10. #10

    Re: KS versus multiple channels for sound families

    I had the same question as markwa and didn't quite find a satisfactory answer in this thread so I thought I'd throw in my own two cents worth. I don't know if these two cents are part of the answer or just restating the question.

    I am using GPO/Sonar/Kontakt 2. I tend toward a mix of entering notes in staff or piano roll view and recording from keyboard depending on the part. I tend I do my "sketching" with the KS instruments because it seems handier for experimentation. But when I'm done experimenting and it comes time to finalize my instrumentation, I'm mystified as to which approach is better. I personally find one major disadvantage to each.

    1. Using separate channels and tracks for each articulation leaves the score difficult to read because the different articulations are scattered across different staffs, even though it's really just a single piccolo part with three different articulations. This drawback may simply be related to Sonar's limited staff view and printing. Other users may have ways around this.

    2. Using keyswitching leaves a bogus score because of the extra, out of range control notes. I read somewhere that there was a way to remap these control notes to other (i.e. CC) events in Kontakt 2, but I don't remember where I read it and haven't yet tried to figure out how to do it myself.

    If somebody suggested a slick workaround to either of these two drawbacks, I would probably head down that road and stop asking myself which approach was "better". Until then, I'm just winging it and living with the drawbacks.

    Thanks if anyone has ideas.

Go Back to forum

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •