• Register
  • Help
Results 1 to 9 of 9

Topic: "Velocity Dynamic Range" For the Love of Mike

Share/Bookmark
  1. #1

    \"Velocity Dynamic Range\" For the Love of Mike

    Well, what I mean is, I\'ve got the manual in front of me and I can\'t find a definition or description of WHAT this part of the filter does.

    My guess, using the Steinway B as an example, is that the \"high\" setting for the VDR gives you more high end at the top of a velocity range and a more muted sound at the low end of a velocity range. (There are three such ranges in the Steinway B sample.)

    I\'m probably just imagining this. So, what DOES this setting do, anyway?

    (I know, I\'ve probably asked this question before, and I\'m going to do a search.)

    John Grant
    http://artists.mp3s.com/artists/42/john_lewis_grant.html

  2. #2

    Re: \"Velocity Dynamic Range\" For the Love of Mike

    My search shows that I did ask this question about a year ago. But I need clarification. Here\'s an example.

    Let\'s say I\'ve said I\'ve set the lowpass filter at 30, for all three velocity ranges of the Steinway B sample. In this situation, a \"high\" velocity dyn range setting would, I suppose, give me a lot of lowpass filtering for each not played at the TOP of each velocity range and very little lowpass filtering at the lower end of each of the three velocity ranges.

    Now if that\'s true, then a \"high\" velocity dyn range setting coupled with a lowpass filter at, say, 30, would produce pretty abrupt tonal transitions when your going from the TOP of one velocity range to the bottom of the next.

    Right?

    JG

  3. #3

    Re: \"Velocity Dynamic Range\" For the Love of Mike

    Hi John, I bet you knew you\'d find me here...

    > is that the \"high\" setting for the VDR gives you more high end at the top of a
    > velocity range and a more muted sound at the low end

    That\'s right but the full story is a bit complicated.


    > high\" velocity dyn range setting would, I suppose, give me a lot of lowpass filtering
    > for each not played at the TOP of each velocity range

    You\'ve got that backwards. Low pass filtering passes low frequencies, and \'filters, deadens, quiets, attenuates, (etc)\' higher frequencies, so the amount of higher frequencies is reduced usually in a linear fasion by the amount the input frequency in question is over the filter cutoff point. That filter cutoff point in real world terms might be 400 Hz (or whatever) but GS gives you an abritrary value (like 30 Giga-Meta-Hz).

    But typically, a lot of things can affect the filter cutoff point, so it\'s like:

    (\"what you set the base value for on a particular note\" + (\"note on velocity\" x \"velocity sensitivity\")) x (\"Filter envelope sensitivity <note1 >\" x \"filter amplitude <note 2>\")

    <note 1:> On GS, this is always 100% but most synths let you set this as a value from 0 to 1 or better from -1 to 1.
    <note 2:> filter cutoff point moves over time following the note-on
    <note 3:> and usually you can have a low frequency osciallator (LFO) vibrating the cutoff frequency over time too by a variable amount

    So you see filter cutoff frequency can be very dynamic, both with velocity and over time with the envelope and LFO. But for a piano with a full sustain recorded, we typically don\'t need to dynamically simulate a piano string dying down and becoming more mellow because the unlooped recording already does that, so I probably wouldn\'t use an envelope to modulate the filter on a piano program.

    So lets get back to velocity... GS doesn\'t really give you enough rope to hang yourself here, for better or for worse. Let\'s say you had 2 velocity layers and you gave them the same filter parameters, say a base of 30 + high velo sensitivity. So the lower layer might give you a filter cutoff at 400 Hz at 0 velocity and 500 Hz at 64 velocity. The upper layer will simply continue the range, maybe 501 Hz at 65 velocity and 600 Hz at 127 velocity (OK that wouldn\'t be high sensitivity but you get my drift).

    So really what happens to repeating notes that vascillate across the threshold, is that the filter point is barely moved but the sample switches back and forth. Any abruptness there is simply a lessened (filtered) amount in the the differences in harmonic content between the 2 samples, which is probably about as musically correnct as you are going to get. You might try adjusting the sensitivity or something to see if you can get a smoother transition and hopefully a gracefully transitioning \'tone-ramp\' on progressively harder strikes, but given that both samples will surely have many overtones over the cutoff point, I will argue that it will generally sound weird/not natural to pass more high frequencies at a lower strike and then reign them back in on a harder strike (even if there will be more of them present, the effect will probably be worse that simply the sample switch under filtration). Remember you won\'t hear the strikes on top of one another, they will just be a bit different following each other, and a piano is like that more than a sample player. But try it and see, you might be able to get something that works well even if it flies in the face of my pow\'ful logic.

    Hope things are going well. And as usual I hope this makes sense...

    -s

  4. #4

    Re: \"Velocity Dynamic Range\" For the Love of Mike

    Wow I just saw that this forum software ate a bunch of my notation in brackets. It must have thought it looked like html markup. Boy is that retarded, it makes my post make a lot less sense. Sorry I\'m not going to fix it, I won\'t try to figure out how to work around this crud, this goofy site has censorred my URL postings in the past.

    Darned amateurish web programming... I do this stuff commercially and it\'s not difficult to do it right.

  5. #5
    Moderator
    Join Date
    Jun 2000
    Location
    Chandler, Arizona
    Posts
    4,045

    Re: \"Velocity Dynamic Range\" For the Love of Mike

    You may want to go into FAQ\'s at the top of the forum page under \'Post Reply\' and then check out the UBB codes section. I support a forum using the Ultimate Bulletin Board and find it to be quite stable and user friendly.

  6. #6

    Re: \"Velocity Dynamic Range\" For the Love of Mike

    So, let me get this straight. In total layman\'s terms:

    1. A lowpass filter reduces the high frequencies (or more precisely, lets the low frequencies through). That\'s easy enough, even though I managed to get it backwards before.

    2. A \"HIGH\" setting for \"velocity dynamic range\" coupled with a \"lowpass\" filter of, say, \"14\" would serve to excentuate low frequencies especially at the higher velocities within a specific velocity range (because the \"higher\" the frequency, the more the VDR takes effect).

    3. Alternatively, a \"LOW\" setting for VDR coupled with a lowpass filter would also eccentuate low frequencies at the higher velocities, but now only a little bit (because the VLD is set at \"low\".

    4. Here\'s an inference: if the above is true, then a \"high\" VDR setting for each of the three velocity ranges in the Steinway B (or any other sample) might tend to INCREASE the abruptness of the transition from one range to the next. IE as we approach the upper most frequencies of the bottom range and move into the lowest frequencies of the middle range, the former are sounding relatively \"muffled\"--assuming the use of the lowpass filter--and the latter are sounding relatively \"bright.\"

    Right? Or too simple?

    JG

  7. #7

    Re: \"Velocity Dynamic Range\" For the Love of Mike

    Hmm, actually this board didn\'t eat my notation, instead
    it forwarded my un-escaped content which looks like
    invalid tags to the clients. I\'m very surprised that
    this forum doesn\'t escape its content since it means
    I can type content into this form which isn\'t
    rendered as text but could be source code that breaks the
    page. So I apologize for inferring that what is actually
    an unfortunate bug or misfeature was an intrusive feature...

  8. #8

    Re: \"Velocity Dynamic Range\" For the Love of Mike

    > 2. A \"HIGH\" setting ... would serve to excentuate low frequencies especially ...

    No. If we can ignore filter resonance in this discussion (and no one puts resonance on the filter used on a piano because it sounds weird and bad, I won\'t get into why here...) then a filter doesn\'t accentuate anything, it only removes frequencies. A low pass filter with a cutoff at 200 Hz will remove a lot of high frequency content at (say) 6000Hz and with the cutoff at 5000 Hz will attentuate the frequency content at 6000Hz just a bit. The filter will not attenuate frequencies at or below the cutoff frequency, but will attenuate frequecies increasingly with their distance above the cutoff frequency (commonly by 24 dB per octave above the cutoff frequency.)

    If you set a note\'s cutoff at 1000 Hz with low velocity dynamic range, the filter\'s cutoff frequency might vary with velocity like this:

    velocity - cutoff
    0 - 1000Hz
    64 - 1100 Hz
    127 - 1200 Hz

    With high VDR, it might be something like this:
    0 - 1000 Hz
    64 - 4000 Hz
    127 - 7000 Hz

    Anyway, as long as the base frequency and VDR (and filter envelope settings, if you use them, but I don\'t for GS pianos) match across layers for a key, the filter frequency will match at the point of the transition, though the sample will be switched \'under\' this filter. This has worked well for me in the past, and some piano programs by others that tried to get too clever with tweaking the filter response have sometimes had quite imperfect and obvious transitions.

    > then a \"high\" VDR setting for each of the three velocity ranges in the Steinway B (or
    > any other sample) might tend to INCREASE the abruptness of the transition from one
    > range to the next. IE as we approach the upper most frequencies of the bottom range
    > and move into the lowest frequencies of the middle range, the former are sounding
    > relatively \"muffled\"--assuming the use of the lowpass filter--and the latter are
    > sounding relatively \"bright.\"

    Not true at all, I hope my example makes clear why not. As long as the layers are programmed with the same settings, the filter settings at the transition ponts will be close regardless of VDR.

    best regards,
    -sam

  9. #9

    Re: \"Velocity Dynamic Range\" For the Love of Mike

    Well, I can see that I have some personal upgrading to do.

    Thanks, Sam

Go Back to forum

Tags for this Thread

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
  •