Physical modelling synths - are there any actually convincing ones?

Discussion in 'Samplers, Synthesizers' started by Xeraser, May 5, 2022.

  1. Trurl

    Trurl Audiosexual

    Joined:
    Nov 17, 2019
    Messages:
    2,480
    Likes Received:
    1,453
    I've thought about getting a breath controller. I think it would really help with wind instruments. But when you get good with exp pedal and mod wheel they work pretty well. And yes, I often go back and do vibrato on a 2nd pass. And then redraw all the controller values :rofl:
     
    • Like Like x 2
    • Agree Agree x 1
    • List
  2. Xeraser

    Xeraser Producer

    Joined:
    Dec 9, 2020
    Messages:
    342
    Likes Received:
    127
    Location:
    Japari Park
    What about the opposite? Why would people with such high budgets bother using physical modelling and "wasting" their time trying to make it sound good when they can just hire an actual player? Maybe they're used for sketching but I doubt they use them for actual production.

    Yeah, that's one of my main issues with the SWAM instruments. If I have to do a shitload of tweaking, microediting, automation and processing.. I might as well dust off my beloved VSL and G-Player.

    I think it's less about MIDI controllers lacking bit depth and more about MIDI controllers just not feeling natural enough to allow you to convincingly mimic a human's performance. This is why things like Yamaha's MIDI/digital trumpets/saxes/etc really intrigue me. The included samples and the programming on those seems pretty terrible and unnatural (I swear I've heard better samples on a Motif Classic) but I wonder how they'd fare as MIDI controllers for physmod synths, particularly the SWAM ones. I think that setup would still make more sense than finding an actual player and/or buying and trying to actually learn to play whatever instrument you need. I know for a fact I wouldn't be able to play a real trumpet due to my busted lungs but maybe I could play a virtual one.
     
  3. orbitbooster

    orbitbooster Audiosexual

    Joined:
    Jan 8, 2018
    Messages:
    957
    Likes Received:
    550
    I think that just saying in or out (or off) doesn't help much.
    Do you need it as solo or "in the mix" instrument?
    That could change drastically your choice: if in the mix, almost anything with the current audio tech could be acceptable/good.
    For solo, you sharp your ear and try to understand if dynamics, timbre, playability linked to hardware are good.
    I played a lot of real pianos, and found Pianoteq convincing (even better than some real one!), though not for all instruments, but at least for some, and really damn good.
    I'm happy that there are companies that go the physmod way, I think that in future (if there will be one :rofl:) sample libraries will fade away.

    I could be wrong, it seems really similar to FL Sakura, have a look.
    That doesn't mean that you blissfully clog up your storage or buy hdds with no expense limits.
    Also, thinking that most if not all libraries are not scalable, you have to take the whole bunch even if you use just a bit of it (I want to thank all minified libs makers!)

    That is the real point. See below.

    I know there are a lot of clips (DR. MIX :wink:) where breath controller is used for SWAM, but in my opinion is unnatural, and even a real clarinet player tried my SY-99 breath controller, and found very difficult to translate in real dynamics, beside is pretty tiring.
    I think that such instruments should (could) be played with hi level controllers like Seabord, OR at least with a controller that allows 2 or more params in real time.
    Personally I made a small layout (TouchOsc) with XY pad, vibrato and expression, that worked pretty well, now I'm thinking of 2 XY controllers that would allow 4 params in real time.
    I found results to be acceptable, in some case good.
    photo_2022-05-06_12-31-44.jpg
     
    • Agree Agree x 2
    • Love it! Love it! x 2
    • Disagree Disagree x 1
    • Interesting Interesting x 1
    • List
  4. Xeraser

    Xeraser Producer

    Joined:
    Dec 9, 2020
    Messages:
    342
    Likes Received:
    127
    Location:
    Japari Park
    Assuming we're not going the cracked route, physmod software isn't cheap either. All SWAM libraries cost 1400€, all Pianoteq packs cost 800, etc. For how much work you need to do to get the results you want out of them (and time you need to spend) using sample libraries is a no-brainer. Most full orchestras don't cost above a thousand and about the biggest one I've seen is almost 700GB in size. But you do get A LOT of stuff.

    Also, yes, I blissfully clog up my storage with just about every library that interests me. Always have, always will. If 100 bucks for storage are a preposterous sum for you to spend on your work/hobby.. perhaps consider getting into something else. I suggested buying one 100€ 1TB NVME, maybe two for futureproofing. Not an entire NAS setup with a petabyte worth of storage.
     
    Last edited: May 6, 2022
    • Like Like x 2
    • Dislike Dislike x 1
    • Agree Agree x 1
    • List
  5. DoubleSharp

    DoubleSharp Platinum Record

    Joined:
    Oct 20, 2015
    Messages:
    235
    Likes Received:
    166
    That's a super important point. I'm using an iPad with a 25 key keytar in live/rehearsal scenarios. I don't expect much but sometimes it can be surprisingly effective. Often various horns for solo/melody instrument but I'm hoping to try my hand at arranging horns. (To be honest, I have a mix of Apps, not just physical synths. Haven't got around to iOS SWAM yet.)

    I'm not saying bitdepth of midi is defintely the issue but I question whether 0-127 is enough for a human interfacing to easily and accurately be expressive. Especially when/if energy(sound) is logarthmic. I suppose this depends on the synths interpretation of the incomming midi.

    I've recently managed to get a RockBand3 Pro guitar which has a really interesting guitar midi controller implementation, each fret is divided by string.(stringbends are not picked up by the implementation) Which is how it knows the note value, and the velocity is trigged by the string pluck. The strings are muted/dampened via the device mute mechanism or a piece of light sponge.

    It takes a lot of physical tinkering, depending on the style of the music. EG raising the pickup on the guitar. Slackening the strings makes it more responsive. The thinner strings tend to have less volume/velocity. Requiring a strong pluck. Each string is on a different midi channel and there are ways of intercepting the signal. I've been looking into Midiflow which looks to provide a pretty simple interface for remapping and managing midi CC values.

    https://www.midiflow.com/documentation/#routings-modifiers-note-remapping-velocity-curve

    I suppose my point is that even something like the guitar, there are a tonne of strange unexpected things that can happen from controller to synth. It's not immediately obvious that playing notes outside of the physical range of the instrument are going to potential make it sound unrealistic.
     
    Last edited: May 6, 2022
    • Interesting Interesting x 1
    • List
  6. DoubleSharp

    DoubleSharp Platinum Record

    Joined:
    Oct 20, 2015
    Messages:
    235
    Likes Received:
    166
    How do you go about doing this? 2nd pass business.

    I'm wondering if would be effective trying to do this in realtime. IE play the midifile and then overrule the CC values. Not sure how easy it would be in practice trying reroute/intercept and then rewrite the midi. Makes me wonder how easy it would be to merge midifiles.
     
  7. clone

    clone Audiosexual

    Joined:
    Feb 5, 2021
    Messages:
    5,948
    Likes Received:
    2,527
    It's very easy if your DAW/sequencer will let you put the midi track into Latch Mode, for automation. You can play the file back and change any parameter/values the device can make use of. Hopefully that is what you mean, anyway :)
     
  8. Dblurgh

    Dblurgh Ultrasonic

    Joined:
    Jan 27, 2022
    Messages:
    62
    Likes Received:
    27
    Are you looking for physical modeling synths that accurately try to simulate particular real instruments or just physical modeling in general that you use to create your own "physical" instruments? As for physical modeling that targets specific instruments, MODO Bass is absolutely amazing. Literally sounds like real bass. Other than that I'm really, really interested in Chromaphone 3. Looks crazy good.
     
  9. DonCaballero

    DonCaballero Producer

    Joined:
    Feb 6, 2016
    Messages:
    160
    Likes Received:
    90
    Pianoteq is pretty mindblowing. I often prefer sampled pianos with more character, but the randomize feature is amazing as you can quickly create unique instruments that are highly responsive and "realistic".

    Modo Bass is also fantastic.
    Applied Acoustics has great stuff, though not hyper-realistic.

    One of my favorites from 20 years ago is Emagic EVP73. It's far from realistic, but has a nice sound that often sits better in a mix than a sampled EP or Lounge Lizard.

    I'll have to revisit the SWAM stuff.
     
  10. fiction

    fiction Audiosexual

    Joined:
    Jun 21, 2011
    Messages:
    1,889
    Likes Received:
    687
    I'd love to hear one single authentic funky slap bass line coming out of MODO.
    To me it's immediately noticeable that it's "not the real thing".

    Maybe I've missed the tricks to get there?

    Other than that, PM serves me better for creating new phantasy instruments than emulating real instruments.
    Pianoteq, Chromaphone, Ableton Suite and Drambo (iOS) can do a great job here.
     
  11. damian9

    damian9 Kapellmeister

    Joined:
    Apr 26, 2021
    Messages:
    65
    Likes Received:
    44
    How has no one mentioned Reaktor's Prism? Incredible synth, it has a steep learning curve but it's worth the frustration. It has a very unique organic and just plain beautiful sound. It's not modeling anything in particular but it really excels at wind/flute type sounds. I even bought a mod called Animator for Prism and I was blown away at how far it can go. You could easily make a whole album with just that synth. You can even make evolving soundscapes with it.

    Another one is Plasmonic, it's a newer synth from the creator of Absynth. I haven't really spent enough time with it to go into much detail but from what I heard it sounded great.
     
    Last edited: May 7, 2022
  12. BEAT16

    BEAT16 Audiosexual

    Joined:
    May 24, 2012
    Messages:
    9,082
    Likes Received:
    6,990
    I have not tested it - but it looks and sounds interesting.

    Hybrid Acoustic - Imagine VSTi (physical modeling synthesizer) 10th Sept.2021 - 139 €


    It has been a real pleasure to collaborate with Expressive E in the creation of Imagine. At the forefront of a new generation of controllers that provide multidimensional expression, their expertise proved to be a perfect match for our physical modeling technology. Clearly the best of both worlds.

    Marc-Pierre Verge CEO Applied Acoustics Systems

    www.expressivee.com/63-imagine
     
    Last edited: May 7, 2022
  13. orbitbooster

    orbitbooster Audiosexual

    Joined:
    Jan 8, 2018
    Messages:
    957
    Likes Received:
    550
    On that I agree, a lot of coding effort and years of experimenting since the VL1, but results can justify that.
    Who the hell are you to spit such dumb suggestions?
    YOU asked an opinion about physmod, I told mine, I still think that while it's not perfect it's a (long) way to promote, I still use (not many) SAMPLE libraries and I used them a lot more in past until I felt more comfortable with physmod.
    I don't judge musical talents by GBytes installed on Pcs, and...
    You don't know a jack shit about my job or hobbies or my resources or my instruments or me, so don't you dare to say that again, ok?
    By the way your avatar tits are fake. Liked it?
    Well not easy to asnwer, 42dB are not a lot considering that even analog cassettes have 60dB, I understand, but let's say we must enjoy it for the moment. I can say that I don't find it too shabby, even convincing at times.
    MIDI 2.0 is not yet very implemented, in future you'll have 16 bits of velocity resolution, plus 32 bits for CCs, cross fingers, but it's a long road.
    Yes, but as I said you need special controllers or tricks like MIDI joysticks on virtual pads.
    However if you're not live you can record with at least expression and then add vibrato and all params you want in DAW post, it works pretty well.
    I did it, but in other threads. It's exceptional, I agree.
     
  14. orbitbooster

    orbitbooster Audiosexual

    Joined:
    Jan 8, 2018
    Messages:
    957
    Likes Received:
    550
    It resembles Chromaphone?
     
  15. BEAT16

    BEAT16 Audiosexual

    Joined:
    May 24, 2012
    Messages:
    9,082
    Likes Received:
    6,990
    Correct, because that's also where AAS's expertise is.

    It has been a real pleasure to collaborate with Expressive E in the creation of Imagine. At the forefront of a new generation of controllers that provide multidimensional expression, their expertise proved to be a perfect match for our physical modeling technology. Clearly the best of both worlds. - Marc-Pierre Verge CEO Applied Acoustics Systems
     
  16. Genoveva Bernhard

    Genoveva Bernhard Producer

    Joined:
    Jan 31, 2022
    Messages:
    135
    Likes Received:
    124
    • Like Like x 1
    • Useful Useful x 1
    • List
  17. orbitbooster

    orbitbooster Audiosexual

    Joined:
    Jan 8, 2018
    Messages:
    957
    Likes Received:
    550
    I read it, that's why I feared a restyling:wink:
    A friend of mine bought a vertical AvantGrand, velocity resolution is something we discussed, but we didn't get a clear picture: does it have high resolution within internal sound engine (let's say 1024) and then is scaled down for MIDI 1?
    It seem so for new generation from here:
    http://forum.pianoworld.com/ubbthre...antgrand-midi-dialect-same-as-disklavier.html
    I have an idea on how to test it but we're far away and we rarely see, so...
    Duh.
    However he's satisfied with it, and is a pro pianist, I trust him.
     
  18. Xeraser

    Xeraser Producer

    Joined:
    Dec 9, 2020
    Messages:
    342
    Likes Received:
    127
    Location:
    Japari Park
    I'm Xeraser, pleased to meet you. Consider investing in some storage. :wink:

    My original comment was 100% serious and not bait but considering how much that upset you it's pretty hard to resist taking another jab -on purpose this time- :rofl:

    I'm better than that though. :yes:
     
  19. Dblurgh

    Dblurgh Ultrasonic

    Joined:
    Jan 27, 2022
    Messages:
    62
    Likes Received:
    27
    Okay, to be fair, I absolutely despise slap bass so I've never tried it out in MODO... You might be right on that one.
     
  20. genophyte

    genophyte Producer

    Joined:
    Mar 13, 2015
    Messages:
    248
    Likes Received:
    112
    i use orchestral , before anything else piano,drum,horn, string related...but a note is mandatory for instruments and settings used cause that thing hates fl studio and savestates

    ancient cringe track of mine.
    orchestral , saxlab bari. motown drums soundfont , too much drum compression..


    honestly ive found some swam do better in certain era mix standards.
    sometimes its ok not to have pristine fidelity
     
    Last edited: May 7, 2022
Loading...
Loading...