• Register
  • Help

View Poll Results: Where do you see the future of sampling?

164. You may not vote on this poll
  • Physical Modeling of/and Actual Sample Data

    116 70.73%
  • Terrabytes of Articulations (loads of actual sample data)

    32 19.51%
  • Other (please feel free to make a prediction)

    16 9.76%
Page 1 of 6 1234 ... LastLast
Results 1 to 10 of 57

Topic: Physical Modeling or Samples?

  1. #1
    Join Date
    Oct 2000
    Orcas Island

    Re: Would you like to see DIVA in GPO?

    I segmented out this poll from an unrelated thread for those who want to respond.

    Gary Garritan

  2. #2
    Senior Member
    Join Date
    Nov 2004
    Davenport, IA (US)

    Re: Physical Modeling or Samples?

    I think Loads land loads of samples first...then maybe physical modeling...id just prefer samples though

    And maybe a GOB to go along with GOS, that'd be cool

  3. #3

    Re: Physical Modeling or Samples?

    I may sound stupid for asking, but what exactly is "physical modeling" ?
    Sean Patrick Hannifin
    My MP3s | My Melody Generator | my album
    "serious music" ... as if the rest of us are just kidding

  4. #4
    Senior Member squoze's Avatar
    Join Date
    Aug 2004

    Re: Physical Modeling or Samples?

    Physical modeling would be if you have a program that creates the sound of a, say, tuba, using math and audio signal calculations, based on, perhaps, the actual construction and air flow in an actual tuba, vibration of the lips, etc.
    This is opposed to a sampler, which basically plays back sounds recorded from actual instruments.

  5. #5

    Re: Physical Modeling or Samples?

    What creates the most acurate impression , a painting or a photograph ?

  6. #6

    Re: Physical Modeling or Samples?

    Physical Modeling has still a long way to go. As to use it along with samples, well, imo, most of the times it adds a synthy sound to them instead of improving them. I hope new developments prove me wrong soon

  7. #7

    Re: Would you like to see DIVA in GPO?

    Quote Originally Posted by Garritan
    I segmented out this poll from an unrelated thread for those who want to respond.

    Gary Garritan
    I think physical modeling with real samples is the only way we'll ever be able to take the next quantum leap in sampling technology. I don't know if it's possible to make a satisfactory sounding model of samples that can completely mimic an instrument, but the benefits of having one merit the effort.

    thats my two cents worth,

  8. #8

    Re: Physical Modeling or Samples?

    Quote Originally Posted by guybrush threepwood
    Physical Modeling has still a long way to go. As to use it along with samples, well, imo, most of the times it adds a synthy sound to them instead of improving them. I hope new developments prove me wrong soon
    Very true. This page has some good sounds from a custom made chip for the Yamaha VL70-m physical modeling module. http://www.patchmanmusic.com/turbovl.html


  9. #9

    Re: Physical Modeling or Samples?

    Hi all,

    At the moment I find that physical modeling is very limiting in the type of sounds that are available.

    Whilst there are huge possibilities of what can be achieved there are relatively few people that are programming up good patches for it. And It has an enormous learning curve.

    Also, nobody creates patches with the type of timbres or instruments that I like. Especially accurate ethnic instruments

    I have often wondered whether some sort of resynthesis, where you fed the computer program a *.wav file and it spat you out a physically modelled model would be possible. This would be fantastic because the number of patches would grow exponentially becase anybody could create patches without having to have the knowledge of all the mathematics and PhM ins and outs.

    I have no doubt that one day we will be using physical modeling and the accuracy of tone and the ease of tools will be there, but now we are still a way off.

    Ben Horwood

  10. #10

    Re: Physical Modeling or Samples?

    (Gary, did you start this just to tease me? This IS my baby...)

    'Proper' physical modeling (i.e., not Tassman analog synth modeling, but full emulation of resonating bodies, closer to a super-VL1) should be capable of extremely realistic results, providing the proper controller resources are available to twist the timbre and effect to emulate the real instrument - breath, reed pressure or trumpet buzz, etc.

    Personally, I feel this is where the important new instruments will come from. The VL1 and VL70 are old technology, with processors and memory many steps below what is currently available. Part one is redesigning the concept (fundamental PM processes-model a brass tube 1" in diameter, with a flare to 3" starting 6" from the end, with a parabolic curve. Add a buzz generator made from a bamboo reed of structural strength w and length,width, thichness x,y,z and opening p . Etc.

    Part two is the wind controller to offer intuitive, real-time control of multiple parameters to effect the change in characteristic. The WX5 and WX7 were good beginnings, but MUCH more work remains to be done here. Replace the sax keys with a piano-style keyboard, so the output is polyphonic. Aftertouch (z-axis) and side-to-side vibrato (y-axis, like the Yamaha YC-45D organ.)

    The problem with samples is they are photographs, not movies. Sample development is like trying to make a feature film by using photographs rather than motion picture cameras. Current state-of-the-art is quite good, but limited by the sample library, not the musician.

    The problem with PM is the amount of work it takes to learn how to obtain musical sound. The solution here is to create an easily-understood GUI interface that anyone with a basic knowledge of traditional instruments can relate to: resonator drop down box - tube, string, membrane. (Pick tube to select next set of options) - material - oak, mahogany, zebrawood, tin, aluminum, steel, etc. size: L,W,H, OD, ID, etc. Vibrator/generator- type-trumpet-style, single reed, double reed, pick, etc. Size, position, etc. rather than lists of numbers to type in, these are presets with variations of appropriate type (list of woods, metals and other parameters looked up from 'timbral' database, easy to program...)

    Expand this into a full-blown interface, add the appropriate presets and you've got a powerful, flexible instrument capable of realist articulation.

    May not sound like a specific player, but will sound like YOU. (I think
    It's all about the music - really. I keep telling myself that...

Go Back to forum


Posting Permissions

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