Another AM Swam Violin Example

Discussion in 'SAMPLE Talk' started by robgb, Oct 12, 2018 at 7:49 AM.

  1. robgb

    robgb I Have Strong Opinions

    2,257
    2,097
    Apr 6, 2016
    Working with samples over the years, I've discovered that, to some degree, learning to play a library can be as laborious as learning an actual instrument (okay, not quite, but you get the idea). The more I work with Audio Modeling's solo violin, the better it sounds (to my ears anyway). I still have a ways to go, but I think I'm getting better at making it sound "natural."



    EDIT: AM Violin Test-003 is a new version with the EQ/Comp gaffs removed. I had foolishly posted the first version with my headphone correction filter and another EQ/Comp still on the master bus, which is why the first version had an odd tonal quality.



    EDIT 2: Changed the SWAM impulse to Electric Violin and added a Stradivarius impulse I bought just to test what an external impulse would do.

     

    Attached Files:

    Last edited: Oct 14, 2018 at 9:12 AM
    novaburst likes this.
  2. I like music

    I like music Senior Member

    142
    89
    Oct 26, 2015
    I like it. I'm sure these used to sound better before I bought them. Is it true that they changed some IRs after the whole SM/AM split?

    I actually mocked up Schindler's list (I know, I know) using AM violin + AM woods, and I think it was HW strings. Might dig it out and add it here. I don't play instruments in, I draw them. This is because of a) the lag and b) I can't play even the simplest of things. But with AM instruments I actually think that the drawing approach is pretty good. Does add a lot to the time it takes to do stuff, but with the crazy control you have, you can mock up 20 different performances of the same piece, and I believe they would sound different enough.

    I wonder if they'll ever manage string sections one day ...
     
  3. muziksculp

    muziksculp Senior Member

    3,761
    1,066
    Dec 26, 2004
    Los Angeles
    Still sounds odd, and synthy to my ears.
     
    meradium and Vovique like this.
  4. ZeeCount

    ZeeCount Senior Member

    93
    39
    May 16, 2015
    New Zealand
    There is this buzzsaw like quality to the tone of the AM strings that once I first heard it I've never been able to unhear. I also find the vibrato adds some unpleasent resonances to the tone, that I have never been able to remove, no matter what processing I have tried.

    This is not to say that the AM strings are unusable: I've used them in plenty of projects, I just find for me I have to be very careful where I place them in the mix.
     
    Last edited: Oct 12, 2018 at 5:22 PM
    robgb likes this.
  5. muziksculp

    muziksculp Senior Member

    3,761
    1,066
    Dec 26, 2004
    Los Angeles
    I think AM-Cello sounds more realistic than their Violin, but imho. both need further improvements.
     
  6. shawnsingh

    shawnsingh Senior Member

    62
    70
    Jun 2, 2018
    As a violin player - sounds pretty great, actually, nice work =)

    Like others said, there is a buzzy quality - but actually I think it might be only a symptom of a different underlying issue: I feel the body of the violin just isn't resonating the right way across frequencies. Ideally that would be solved by having a better IR (does AM violin use IRs to represent the body of the violin resonating?). With the right IR, not only can it fix the spectrum balance, but also how long each frequency resonates for. The next best thing would be to do EQ. So I toyed around with EQ, and I found a setting I personally thought sounded better on your clip:

    (using VSL Equalizer)
    band 1: -8.4 dB centered at 11.5 kHz, a wide Q (Q=0.35)
    band 2: +5.3 dB centered at 460 kHz, a medium Q (Q=0.63)
    band 3: +4.9 dB at 2.6 kHz, narrow Q of 3.16
    low cut 12 dB/octave, positioned at 240 Hz (this one might not matter as much, not sure)

    The first two bands create crudely a straight line from +5.3 at 500 Hz down to -8.4 dB at 11 kHz, which I felt made the instrument resonating sound more natural. Then, for the close-mic sound that you had, the third narrow band helps add just the right amount of clarity without nasal or buzzy feeling.

    Along with this EQ change, I wanted to hear the subtle crunch noises of bow changes, and also there's a subtle timing issue in a real player between bow changes, finger changes, and hand positioning. Those three things never get timed perfectly which is part of what would make a violin performance sound real. Its kind of analogous to having a virtual woodwind player have infinite breath. I don't know how AM violin works, but I bet you could emulate a bow change by having the player (a) lighten bow pressure just before the change, (b) but there would be a slight instantaneous spike of bow pressure to emulate that the string gets crunched on the exact moment the bow changes direction, then (c) back to light bow pressure and back to normal. Again, that's just a wild guess knowing how I would try to perform smooth bow changes as a violin player. And also fingering messiness, particularly when a player's hand has to move up and down the neck of the violin.

    With all those things I think you could get a pretty crazy realistic sound.

    Cheers!
     
    mike753, Zee, Saxer and 1 other person like this.
  7. shawnsingh

    shawnsingh Senior Member

    62
    70
    Jun 2, 2018
    Took another listen after 5 mintues away, and I think the "band 2" that I suggested could be reduced a lot, like maybe just +2 dB instead of +5
     
  8. kavinsky

    kavinsky misty orchard in the middle of Czechoslovakia

    424
    288
    Jul 4, 2012
    I don't like that filter sweep effect when you ride the dynamics.
    But yes the transitions sound pretty smooth.
    I'm not sure why one would go through such a hassle of programming a solo line though. It would take you a lot more time than hiring your violinist friend to do a 1 hour session and it probably would cost not as much as the plugin and your time, and will inevitably sound superior.
     
  9. Vovique

    Vovique Senior Member

    434
    188
    Aug 11, 2013
    In all honesty, a far cry from a natural violin sound
     
    kavinsky and muziksculp like this.
  10. OP
    OP
    robgb

    robgb I Have Strong Opinions

    2,257
    2,097
    Apr 6, 2016
    Folks, you're absolutely right about the tone sounding unnatural. I just realized I posted a clip that I accidentally rendered with my headphone correction filter on the master bus (Sonarworks), plus another EQ/Comp I had been testing out. Unfortunately, I foolishly rendered offline and didn't listen to the MP3 afterwards and now that I do can certainly see what people are talking about. I'd repost with the untreated violin, but I closed the project and didn't bother saving it since it was just a test. My bad. When I get some time, I'll repost here with a new properly EQ'd test. Let this be a lesson in not paying attention to what you're doing...

    That said, I think the playability of the instrument is pretty clear, since this was played in real time.
     
  11. Hanu_H

    Hanu_H Senior Member

    792
    291
    Nov 7, 2011
    I think it sounds great. Not as good as real violin but really impressive for a VI. I can also hear some synthiness in few places but the transitions and the fluidity of the lines is way better than any sampled instrument could deliver. If you would hear this in a composition, I am sure it would fool most of the people. This is the future of virtual instruments.

    -Hannes
     
    mike753 and robgb like this.
  12. OP
    OP
    robgb

    robgb I Have Strong Opinions

    2,257
    2,097
    Apr 6, 2016
    There's a weird filtered "sheen" to it that shouldn't be there. I wish I had saved the original...
     
  13. OP
    OP
    robgb

    robgb I Have Strong Opinions

    2,257
    2,097
    Apr 6, 2016
    A new quick and dirty one without the EQ gaffs.

     

    Attached Files:

    Sid Francis and shawnsingh like this.
  14. OP
    OP
    robgb

    robgb I Have Strong Opinions

    2,257
    2,097
    Apr 6, 2016
    The difference in sound is my fault, explained above. The second version is out of the box (with reverb added). I actually think the sound has improved with each new upgrade. There used to be MORE IR choices, but I think they kept the best ones.
     
    shawnsingh likes this.
  15. stfciu

    stfciu Senior Member

    54
    30
    Aug 13, 2017
    Much better. Especially the attack seemed harsh and ruined the overall impression. The vibrato could be a little bit modulated in my opinion.
     
    robgb likes this.
  16. novaburst

    novaburst Senior Member

    1,850
    713
    Dec 21, 2015
    the 2nd one sounds very good,
     
    robgb likes this.
  17. Vardaro

    Vardaro Senior Member

    184
    67
    Feb 16, 2014
    Near Paris, France
    Interesting sound but absolutely nothing like any violin I have ever heard or played.
     
  18. shawnsingh

    shawnsingh Senior Member

    62
    70
    Jun 2, 2018
    For me personally, it sounds close enough to the real thing. Among all the other SWAM demos, @robgb this one has finally sold me on SWAM solo strings. Still a few performance nuances that feel off, but they are very minor and don't detract from the authenticity of the violin tone on the new version. :)
     
  19. OP
    OP
    robgb

    robgb I Have Strong Opinions

    2,257
    2,097
    Apr 6, 2016
    I have yet to find a single sample library of any kind that sounds like the instrument it's supposed to be "replacing." They can only be approximations of the real thing. As for a violin library, I own several and this one is far and away the most authentic sounding. But, of course, it will never sound like the real thing to trained ears. The best we can hope for is close. And, fortunately, most of the people who will hear our work can't tell the difference. In fact, a couple years ago someone on this forum slipped the SWAM violin into a comparison of several solo violin libraries against a real violin, and this was the one that most people thought was the real thing. But, obviously, as a player, you know the subtle nuances that give any fake away. Just as I do with guitar libraries.
     
    mike753 and Motr3b like this.
  20. pmcrockett

    pmcrockett Senior Member

    423
    294
    Nov 3, 2014
    Saint Louis
    Here's my take on the first phrase of the second example. This is with the old IRs because I haven't gotten around to updating yet. Performed with a Leap Motion controller with very minor touch up editing afterwards and a bit of extra reverb. I'm also using a Reaper script that lets me select bow direction by pressing/releasing the sustain pedal, effectively bypassing the natively implemented bow control methods, all of which are extremely awkward.

    I think I've avoided a lot of the synthiness that people have noted by staying primarily in lower dynamic ranges. The high dynamics sound best when used sparingly and for accents; I find that just sustaining them almost always sounds bad.

     

    Attached Files:

    rottoy, Saxer and kavinsky like this.

Share This Page