a year or so ago, the two main areas for mock-ups which were problematical in my view were vocal renderings and solo strings. Then we had SynthesizerV for vocals and the even more impressive, if perennially buggy, Cantai. Well Dreamtonics are back in town with almost a complete orchestra created with their AI synth. Initial tests do not suggest that they will be about to replace sample libraries for orchestral works but in my opinion,their solo strings break new ground in musicality and expressiveness. Some time ago some may remember I posted my final movement from my 12th quartet - dolce e molto espressivo and now you can hear this with the new technology.
Would be interested in getting reactions to this one – as with SynthV this works via a MusicXML export from Dorico (plus MIDI required if there are gradual tempo changes) and because it works as a plug-in can also be used together with other instruments though in the case of a string quartet, it’s probably easiest just to do all the editing in its own interface.
by that do you mean you are worried you might have to spend some money on them? The solo string quartet collection is around €230 (plus another 30 or so if you want double bass) – not cheap but not bad compared to offerings from the big vendors. And of course you can get any size of string section as well, though I don’t think the larger ones work.
that too. But “AI” synthesis is something to really keep an eye on both for the good it can bring and all the damage it can do to the entire sample ecosystem. I mean, VSL, Eastwest etc, really should be taking notice.
I understand. Sample libraries still have life in them but probably eventually they will start to fade away. Of course something like Instrumnet X can’t actually be played and part of the appeal for DAW composers is how they respond to live playing. But for a mock-up from notation software like Dorico which is what concerns us here, I suspect AI synth will become increasingly prominent.
ok, I’m trying to understand something here…
does this “AI” use mean that you have to upload your file and then some AI on some external server does the heavy lifting and tries to “interpret” your score?
not really – you export a MusicXML file which already has all the dynamics and articulations for the software to work on. All it really then tries to do is to make music out of it. All processing is done on the local computer and you can play the score as normal in Dorico but using instrumentX as a plug-in. In the plug-in you can edit anything you don’t like. The only thing is you can’t actually compose in real time using it as it requires an exported score as the base. So it’s a good idea to have the score really complete before you do the export, otherwise you’ll need to re-export for the two to remain in synch.
For the strings all standard playing techniques are supported including sul pont, harmonics, con sord etc. Only one or two really modern techniques such as you might find in the most comprehensive libraries like XSample are not (so far) present.
what do you mean exactly? Are the score dynamics not being imported? I find that import of dynamics from the score somewhat approximate – sometimes the software has its own interpretation but in general the rendering follows the spirit of the score and of course you can easily adjust the dynamics in the UI where it gets it wrong. This is different from SynthV which completely ignored score dynamics.
Hi my friend composers. I tried the ACE alternative to X instruments few months ago and submitted my audio to Spotify and it has been tagged AI Generated, event if the score was human made. So to publish such audio rendering to streaming platforms you need to provide proofs to these platforms that it is an AI fair use for each song submitted. I cancelled immediatly my subscription to ACE Audio. Wondering if we have the same issue with the X Instruments : did any of you tried to submit an audio mock up of your original work using X instruments to a Streaming Platform ? Is it tagged AI generated ?
well certainly I had no problems with Reelcrafter. But anyway, I don’t understand how with Instrument X, a project could be tagged as everything is locally rendered and indeed you can simply export audio direct from Dorico using the library if desired. It works completely differently from ACE.
I suppose audio from ACE is watermarked. I do not have a clue how Tunecore achieved this (the distributor I use To Spotify). But AI is AI, on your machine or on the cloud. Good to know X instruments are not tagged generative AI.
it works just the same way that SynthV works. In other words, after you’ve exported the MusicXML file to Instrument X, you can play back through Dorico if you simply assign the instrument to an InstrX channel.
I just finished a (by my standards relatively short) string quartet which had Instrument X in mind from the outset, though of course you have to use a different library first (in my case Synchron solo strings) initially while composing. You get a reasonable variety of textures in this rather restless and enigmatic work.
beginning to run out of things I still want to convert to Instrument X but perhaps as a final bash is my Piano Quintet no. 1 because one or two seem to like it most among my chamber works – at any rate it’s among the strangest, reticent and most “modern” with some influence of Schnittke in the background. Particularly in the bleak third movement, it seems to bring a new dimension to the atmosphere although a good real ensemble would be certainly better still.
Can be found here in the “other chamber works” section.
The solo string quartet collection is around €230 (plus another 30 or so if you want double bass) – not cheap but not bad compared to offerings from the big vendors. And of course you can get any size of string section as well, though I don’t think the larger ones work.