In some ways that may be true. The thing I’ve observed, though, with BiaB is that productions I’ve heard people make with it always feel kind of “flat” to me on the arrangement front. I used to use BiaB many years back to experiment with styles when considering where to go in making demos of my songs. It could be helpful for that, but the actual results weren’t very usable, even if exported to MIDI and brought into a DAW to assign better quality instruments.
Once they added RealTracks, that could improve things for sounds that weren’t good with MIDI, and I was very interested in that (especially for instruments like pedal steel). Somehow, though, I was never able to figure out how to get that to work in my workflow, which really depended on the DAW. I know they came out with a plugin to try and help with that, but I never managed to get it to work in practice.
The most recent version I had was a number of years old, and I pretty much never ended up using it in any useful way, so I didn’t even bother to install it on my new Windows 11 system, nor have I been inspired to upgrade to a newer version.
In some ways, I might be inclined to consider Tonalic more like a competitor for the higher end Suno subscriptions, albeit a potentially more ethical possibility. However, it doesn’t do some of the things Suno does. For example, it doesn’t do vocals, and my impression thus far is it isn’t likely to do lead parts in general.
However, for what I might be inclined to consider Suno for, Tonalic seems more controllable. One of my songwriting collaborators started using Suno to record and release versions of songs he’s written, and he approached me about some of our songs. There is one in particular that I’ve never been able to do because it is a decidedly female (mother/daughter) song, that simply couldn’t fit a male singer (at least not without a totally different lyric – it’s not just a matter of switching pronouns and such). He used my original piano/vocal demo of the song (from way back in 1998!) as input and had Suno generate a few versions of it. I gave him some feedback, and he adjusted prompts and generated more versions, and this went on for a few rounds. The thing is, each version got some things right and some things very wrong, and it wasn’t a case of continuous improvement – e.g. things that were good in one version might be bad in a later one – and it was ultimately clear that it would probably never get it completely right, despite having the same piano/vocal input for all the iterations, where the melody should have been clear.
I ended up having him get stems from multiple versions, then I put those into Cubase to try and come up with one version that would work all the way through. The arrangements were different from version to version, so it wouldn’t have be practical to try and combine different versions on the instrumental front, so I just picked the version that had the best instrumental as a starting point. Then I comped vocals from all the versions, shifting them around as needed to match the chosen arrangement. That required a fair bit of work, but, aside from one phrase near the end of the song, it at least got the right melody and phrasing, and the comping and editing made it feel like one performance. For the one phrase at the end, I had to use Melodyne to change some of the notes because none of the versions got both the right melody and the right phrasing.
Of course, Tonalic doesn’t do vocals, but if using it to do instrumental tracks on a recording, first there is the thing that you’re specifying the chord changes and structure in your DAW, so it’s not as random as Suno making creative decisions that don’t correspond with your input and desires. But then, if there is something you don’t quite like, and you have the high-end version of it, you can tweak it at a detailed level (with its Melodyne-like capabilities).
What I mostly do on rhythm guitar parts now is use products like UJAM’s Virtual Guitarist series or NI’s Session Guitarist series. In some ways, it is similar in concept to what Tonalic appears to do, where you supply the chords and the instruments adjust their strumming or picking patterns to suit those chords. But it can be a bit of a crapshoot to find patterns that match your song and what you’re looking for in that context (and that is something that is a big question mark with Tonalic, too – i.e. is there enough variation to find what is needed in a specific context). But let’s say you get close but want to tweak further (this can especially be true with transitions and endings). That can be a real challenge with the products I’m currently using, and it can take a fair bit of work and compromise in working around limitations. This is the area where it seems to me the high-end version of Tonalic could excel.
The net is I’m excited about the promise of Tonalic based on what I’ve seen in the video demos, but how that maps to reality in terms of workflow and needs for any given project remains to be seen. There certainly have been times in the past where I’ve been wowed by a demo but ran up against practical limitations for my specific needs that made the product far less useful than I’d expected. I am looking forward to trying Tonalic out sometime in the not-too-distant future.
It’ll never catch on …