To be honest, this has always irked me. Even the cheapest toy keyboard from Walmart has near-zero latency. They must load those things up with massive amounts of RAM and multi-core i7 processors, right?
Meanwhile, the average DAW struggles whenever you load up any sort of track count, or have active plugins, or whatnot.
(Yes, I fully realize that DimPro vs Walmart toy is not a fair comparison - but this is just a harmless rant so don't overthink it). The way I see it - it's now 2015. We're living in the future. We ought to be able run softsynths without bringing a DAW to its knees. I'd be happy if there was a lo-res playback engine that we could use while tracking. Or better yet, dedicate processor cores and memory for the softsynth with its own buffer/latency settings.
How much more powerful do our machines need to be for developers to start working on this anyway? Are they just biding their time, hoping the softsynth craze fades away and the kids go back to playing guitars and drums?
As far as I'm concerned, making adjustments to the buffer/driver will always be a "work around", not a "solution".
That's the end of my rant. I am hereby subscribed, and eagerly await more useful feedback from the more educated among us.
For now, this is how I deal with it. First, I should say that my rant is completely unwarranted, and I only make it more difficult for myself by being so stubborn. I
never adjust drivers/buffers, I just set 'em high and forget about it. Likewise, I
never disable plugins or freeze tracks/stems. And to make matters worse, I stack up ridiculous tracks counts with loads of plugins (mostly audio tracks, with a few soft synths). So yeah, I get it. Boo-hoo for me, I'm hardly a sympathetic character with regard to latency.
Anyway, here's how I deal with keyboard latency...
1) I record MIDI data into my DAW through my keyboard/controller's USB output.
2) I disable input monitoring in Sonar, and set the MIDI track output so it goes OUT through the multi-pin midi connection on my interface.
3) From the multi-pin connection on my interface, I route it through an old 90's era Alesis NanoSynth module (
which must have also been loaded massive amounts of RAM and multi-core processors, lol).
4) I monitor the outputs of the NanoSynth so I can play with real-time monitoring while recording and upon playback.
5) Once my performance is captured, I toggle the MIDI track's output back to the soft synth. Then I tweak my sounds in DimPro or another softsynth, and I move on to the next track.
Is that less annoying than adjusting the buffer? Only slightly.
Is it less annoying than freezing stems, disabling plugins, and then unfreezing and re-enabling plugins when I'm done? You're darn right.