Re: [LAD] Software sound
On Sat, Aug 30, 2014 at 10:13:19AM +0100, W.Boeke wrote: Yoshimi is a neat repacking of ZynAddSubFX. It offers all kinds of interesting sound, but I call it musician-hostile. What you see is a bunch of controls that you have to manipulate with the mouse while playing, trying to read and understand the very small labels (on a modern high-DPI monitor) and try to get the sound, attack etc. that you want. It is next to impossible to perform this, so you will only try some of the 510 build-in patches, but you can't modify them easily. And modifying the sound while playing is essential for a good performance. In other words: Yoshimi is way too complicated. It would be more useful to supply only 10 patches that are easy to modify. There are efforts under way to separate yoshimi's DSP from the GUI so that alternative GUIs are possible. A long term goal, so don't expect this to happen all too soon. Also hovering around is the idea (and a branch[1]) to add MIDI-learn functionality to the knobs to replace/expand the existing hard-coded MIDI bindings[2] [1] https://github.com/licnep/yoshimi [2] http://zynaddsubfx.sourceforge.net/doc_3.html Take for instance an ADSR envelope generator. In Yoshimi (and in most other software synths) there are 4 knobs that you must control. This is replicated from old hardware synths. Why not show a small diagram of the envelope, that you can modify directly with the mouse or with your fingers? In both zyn and yoshimi, the buttons labelled 'E' next to the envelope controls will open a window containing a mouse-editable diagram of the envelopes. ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
On Sat, 2014-08-30 at 10:13 +0100, W.Boeke wrote: [snip] I'm able to edit existing Yoshimi sounds, but that's not my point. My point is to add a vector control that is similar to the one provided by the Yamaha TG33, then you exactly would get a musician friendly control to edit sounds. People often can't imagine how effectively different sounds can be produced by recording a mixing sequence with a joystick, even if it only records volume and fine tune. There is no easier effective sound editing than to record a volume/pitch mix with a vector control. The joystick could be provided virtually by a GUI, as done by Alchemy http://www.camelaudio.com/images/ios/1-1-35/iphone/alchemy-mobile-iphone-1-1-35-xy-pads.png . Instead of using Yoshimi simply providing a MIDI filter that can record vector movements and store them and sends it to 4 other synth. The vector simply could send CC data for volume. IOW the MIDI filter, perhaps using the same CC as the TG33, IOW CC 16 vector x-axis and CC 17 vector y-axis are translated to CC 7 volume for 4 different synths and then is send to them. The MIDI filter will record the sequence of the vector control and MIDI channels of the 4 synth in presets. JFTR have you ever used a vector controlled synth such as the TG33? ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
[W.Boeke] Compared to the available technical possibilities of the past, software designers nowadays have a much easier life. A computer and a MIDI keyboard is all you need, you can try all kinds of sound creation, so why should you stick trying to reproduce the sounds of yore? I definitely agree with the sentiment but it's not an easy task to purely digitally create timbres as rich, complex and pleasing as produced by analog, let alone physical (non-electronic) instruments. For example, analog op-amp or diode saturation is quite simple to realise and capable of producing anything from very smooth harmonic extension to screaming distortion. The digital equivalent needs to be oversampled, complicating things greatly. Another example is the stretching of the partials of a string with non-negligible stiffness. Wavetables need to be very large to create a similar spectrum, while waveguide oscillators need a very high order of allpass filters, again complicating things greatly. The numerous imperfections and nonlinearities in real instruments and the way they play all contribute similar effects, and while they may often be inadvertent and/or subtle, they are vital to sonic interest. You can of course eschew these and set out to find features unique to the digital realm (and there certainly are a few that have some appeal). Yet I think that there is a deep agreement between the way we perceive sound and the fundamental characteristics of physical oscillation, and that is of great importance for me personally (YMMV of course). Cheers, Tim ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
On Sun, 2014-08-31 at 12:21 +0200, Tim Goetze wrote: [W.Boeke] Compared to the available technical possibilities of the past, software designers nowadays have a much easier life. A computer and a MIDI keyboard is all you need, you can try all kinds of sound creation, so why should you stick trying to reproduce the sounds of yore? I definitely agree with the sentiment but it's not an easy task to purely digitally create timbres as rich, complex and pleasing as produced by analog, let alone physical (non-electronic) instruments. For example, analog op-amp or diode saturation is quite simple to realise and capable of producing anything from very smooth harmonic extension to screaming distortion. The digital equivalent needs to be oversampled, complicating things greatly. [snip] Wait a moment. Building an anlog synth by electronic circuits could be harder or easier to do, it depends to the skills of the developer. For software designers in the past hard real-time MIDI was easier to provide, than it is to provide nowadays. There are many factors important. Several times even galvanic isolation for the MIDI interfaces provided by old gear and by modern USB devices was mentioned. What ever in theory could be programmed, even the different output amplifiers of different real synth already could provide differences for the sound, one computer, using the same sound card for several virtual instruments only would provide, if somebody would care about it. Not to mention CEM/Curtis filters etc.. Understanding an ADSR with or without a graphic display shouldn't be an issue for musicians, it might be hard for some beginners. Anyway, that is provided since decades, even the faders of a simple ADSR already show the envelope by the fader positions. I suspect the OP has got less experiences with professional gear, it abilities and what professional musicians do need. It can't be repeated often enough. The way that was provided already in the 80s is a vector control. Everybody can create sounds by a vector control, absolutely no knowledge is needed. For using an ADSR there's the need to understand very much, less how the envelope does look like, more important is what is controlled by the envelope and that the envelop might variate to the dynamic of the musicians playing, by the played note number etc.. ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
PS: So to provide a new electric instrument even nowadays not only a computer and a keyboard is needed. The virtual instrument might work on some distros and on other distros it might not work. It might work with some computers and sound cards and MIDI interfaces, but not work with other hardware. If somebody want's to provide the complete solution it's not just having a computer and keyboard. Even when a software developer just cares for the studio in the box and maintains it for several releases of just one distro, but does not care about using it with external gear, some algorithm (user friendly or not) might need much computer resources, could cause unwanted latency etc. pp. In the end building a synth by a circuit layout and some semiconductor could be easier to do, than to provide software with the guarantee that the software will do what it should do, since a wrong MIDI interface already could cause ground loops, assumed somebody want's to connect the virtual synth with what external gear ever. Making new sounds for a synth always needs a learning curve. A good GUI isn't the solution, if a user e.g. doesn't understand how a filter does work, the best GUI for a filter can't help. What ever synthesis is used, there is the need to have knowledge, no new synthesis will be a solution for this issue. The only way are workarounds such as the vector control, or anything comparable to it. Computers can't solve higher-ranking issues caused by nature. ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
Hi Tim, As an example of attempts to recreate digitally a famous component from the past I would take the Moog 3-stage lowpass filter. This was build using diodes where their non-linearity was exploited in a careful way, yielding a warm, rich sound. Numerous attempts have been made to do this digitally, resulting in very complicated and CPU-intensive algorithms. I have the feeling that this is not the way to go. One better can search for methods that can easily be implemented in software and that still sound good. While experimenting I found that slow, small changes in the waveforms are essential, as well as changes resulting from the excitation level. At that, the kind and value of distortion is important. Last but not least the basic waveform is determining, a simple pulse or sawtooth isn't enough. Believe me or not, but I found sounds that I enjoy very much, I can listen to a single note or some simple chords with great satisfaction. And it would be impossible to create them mechanically or electronically! The sound of a good piano can be marvelous, but it is always the same sound, you cannot modify it. In contrast, if you get bored by a digital sound, you modify it, or choose another. Vivat software sounds! Wouter On 08/31/2014 11:21 AM, Tim Goetze wrote: [W.Boeke] Compared to the available technical possibilities of the past, software designers nowadays have a much easier life. A computer and a MIDI keyboard is all you need, you can try all kinds of sound creation, so why should you stick trying to reproduce the sounds of yore? I definitely agree with the sentiment but it's not an easy task to purely digitally create timbres as rich, complex and pleasing as produced by analog, let alone physical (non-electronic) instruments. ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
On Sun, Aug 31, 2014 at 01:34:59PM +0100, W.Boeke wrote: As an example of attempts to recreate digitally a famous component from the past I would take the Moog 3-stage lowpass filter. This was build using diodes where their non-linearity was exploited in a careful way, yielding a warm, rich sound. Numerous attempts have been made to do this digitally, resulting in very complicated and CPU-intensive algorithms. The famous Moog 'ladder filter' has 4, not 3 stages, and doesn't use diodes but transistor pairs as a non-linear impedance. The result is similar to a tanh(x) function in each stage. It doesn't take very complicated code nor excessive CPU. Ciao, -- FA A world of exhaustive, reliable metadata would be an utopia. It's also a pipe-dream, founded on self-delusion, nerd hubris and hysterically inflated market opportunities. (Cory Doctorow) ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
On Sun, Aug 31, 2014 at 08:13:01PM +0100, W.Boeke wrote: Sorry Fons, if you google moog filter schematic you will see 3 stages, where a stage is a capacitor in series with 2 transistor emitters (which have a diode V/I characteristic). All versions I've every seen have four stages, and with three the feedback wouldn't have the right phase. If you find a schematic with three stages please post a link. Each stage is a capacitor in parallel with the series connection of two emitter impedances, and driven by a current source, the collector currents of the previous stage. About that complicated code: I saw implentations with a lot of code lines, and preventing oscillations originating from the non-linear tanh functions needed a lot of care. I analysed this thing to dead around ten years ago, when I wrote the MCP plugins. There is no instability problem resulting from the non-linearity. The main problem with a straightforward digital implementation is that it becomes a bad approximation at higher frequencies, this requires some attention to get right. The easy solution is to run it at a higher sample rate. Ciao, -- FA A world of exhaustive, reliable metadata would be an utopia. It's also a pipe-dream, founded on self-delusion, nerd hubris and hysterically inflated market opportunities. (Cory Doctorow) ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
On Sun, Aug 31, 2014 at 07:42:48PM +, Fons Adriaensen wrote: On Sun, Aug 31, 2014 at 08:13:01PM +0100, W.Boeke wrote: Sorry Fons, if you google moog filter schematic you will see 3 stages, where a stage is a capacitor in series with 2 transistor emitters (which have a diode V/I characteristic). All versions I've every seen have four stages, and with three the feedback wouldn't have the right phase. If you find a schematic with three stages please post a link. Each stage is a capacitor in parallel with the series connection of two emitter impedances, and driven by a current source, the collector currents of the previous stage. Correct. Even the 18dB/octave Roland TB303 ladder filter has four stages. Why isn't it 24dB/octave? Because they're more honest about the real-world performance than Moog :-D Once you go hanging weirdass source and load impedances off your ladder the response can vary wildly. Tim Stinchcombe has an incredible analysis which shows all sorts of odd bumps and peaks in what you'd typically expect to be flat passband. About that complicated code: I saw implentations with a lot of code lines, and preventing oscillations originating from the non-linear tanh functions needed a lot of care. I analysed this thing to dead around ten years ago, when I wrote the MCP plugins. There is no instability problem resulting from the non-linearity. The main problem with a straightforward digital implementation is that it becomes a bad approximation at higher frequencies, this requires some attention to get right. The easy solution is to run it at a higher sample rate. Ah, the fons-filter. Extremely stable and a nice acid-y squelch. Since a tanh function tends to unity at either end, I can't see how that *would* go unstable. You'd need to have greater than unity gain for it to really go off. Fons - ever considered turning your analytical skill to the MS20 and Steiner Synthacon diode-chain Sallen-Key filter? -- Gordonjcp MM0YEQ ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
On Sun, Aug 31, 2014 at 09:00:08PM +0100, gordon...@gjcp.net wrote: Correct. Even the 18dB/octave Roland TB303 ladder filter has four stages. Why isn't it 24dB/octave? Because they're more honest about the real-world performance than Moog :-D Or they may be some IP related issues. Once you go hanging weirdass source and load impedances off your ladder the response can vary wildly. Tim Stinchcombe has an incredible analysis which shows all sorts of odd bumps and peaks in what you'd typically expect to be flat passband. Yes. In a first approximation the filter isn't to difficult - each stage is in fact nicely isolated from the following once (since emitter current will determine collector current, and collector voltage won't affect that very much since the bases are almost grounded for AC. Fons - ever considered turning your analytical skill to the MS20 and Steiner Synthacon diode-chain Sallen-Key filter? If you can provide a schematic I may be tempted... Ciao, -- FA A world of exhaustive, reliable metadata would be an utopia. It's also a pipe-dream, founded on self-delusion, nerd hubris and hysterically inflated market opportunities. (Cory Doctorow) ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
On Sun, Aug 31, 2014 at 08:26:18PM +, Fons Adriaensen wrote: Fons - ever considered turning your analytical skill to the MS20 and Steiner Synthacon diode-chain Sallen-Key filter? If you can provide a schematic I may be tempted... http://yusynth.net/Modular/EN/STEINERVCF/ :-D -- Gordonjcp MM0YEQ ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
On 08/31/2014 08:42 PM, Fons Adriaensen wrote: On Sun, Aug 31, 2014 at 08:13:01PM +0100, W.Boeke wrote: Sorry Fons, if you google moog filter schematic you will see 3 stages, where a stage is a capacitor in series with 2 transistor emitters (which have a diode V/I characteristic). All versions I've every seen have four stages, and with three the feedback wouldn't have the right phase. If you find a schematic with three stages please post a link. You are right: it's 4 stages. W,Boeke ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
On Sun, Aug 31, 2014 at 09:43:36PM +0100, gordon...@gjcp.net wrote: On Sun, Aug 31, 2014 at 08:26:18PM +, Fons Adriaensen wrote: Fons - ever considered turning your analytical skill to the MS20 and Steiner Synthacon diode-chain Sallen-Key filter? If you can provide a schematic I may be tempted... http://yusynth.net/Modular/EN/STEINERVCF/ :-D :-D That's going to be horrible. Everything coupled to everything else, defeats any analysis except a quite low amplitudes. Reminds me of a simulation I had to develop last year: aircraft landing gear. Non-linear forces all over: the struts, friction, sticktion, and whatever you can imagine. Had to run at a few kHz sampling rate to get any credible results. Ciao, -- FA A world of exhaustive, reliable metadata would be an utopia. It's also a pipe-dream, founded on self-delusion, nerd hubris and hysterically inflated market opportunities. (Cory Doctorow) ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev
Re: [LAD] Software sound
On Sunday, August 31, 2014, Fons Adriaensen f...@linuxaudio.org wrote: On Sun, Aug 31, 2014 at 09:43:36PM +0100, gordon...@gjcp.net javascript:; wrote: On Sun, Aug 31, 2014 at 08:26:18PM +, Fons Adriaensen wrote: Fons - ever considered turning your analytical skill to the MS20 and Steiner Synthacon diode-chain Sallen-Key filter? If you can provide a schematic I may be tempted... http://yusynth.net/Modular/EN/STEINERVCF/ :-D Original Steiner article http://user.xmission.com/~dingebre/N-Steiner-VCF-1974.pdf -- -- A C Censi (on the road) acce...@gmail.com acce...@montreal.com.br ___ Linux-audio-dev mailing list Linux-audio-dev@lists.linuxaudio.org http://lists.linuxaudio.org/listinfo/linux-audio-dev