I have a bunch of GSt nits I would like to see fixed in the upcoming versions, hopefully by putting them out here they will get some attention.
1) The level meters are terrible, they are just not responsive or useful. It seems like they\'re optimized to minimize CPU usage, which is fine, but as long as they\'re not going to be updating frequently enough, they should at least really show you the peak that passed through the channel during the update interval. If I put a transient sample on a channel like closed hats, it rarely shows on the meter at all. It looks like an instantaneous single measure of the level, which doesn\'t cut it for short sounds.
2) The faders jump when you touch them, they don\'t retain the offset from the mouse position to the center of the fader.
3) If you unlink a stereo channel in the dsp station view, move 1 fader, then relink them and move the fader, it retains the offset only if you don\'t move a fader to the top or bottom, if you do that the offset is lost. As long as the offset is allowed, it shouldn\'t be able to be lost by moving a fader.
4) In the DSP view, if you mute a channel and touch the fader, the mute is lost.
5) In the DSP view, if you disable the \'dsp\' processing (ie inserts and auxes) and save the .gsp performance, then reload it, the disabled state of the dsp is not retained.
6) I wish there were a decibel calibration in the midi mixer view like there is in the dsp mixer. What does a 6 db movement look like here? I wish I could label mixer channels similarly to dsp channels, I have trouble differentiating the channels here (what with the poor meters, cluttered close channels and all). It feels like the dsp area is a better place to mix (labels, fader calibration, space) but it\'s not (no pan, goofy mute and linking).
On my hardware mixer, I really differentiate channels with a bunch of artists tape. GSt would be much clearer if it only had dark/light alternating channel pairs and better meters.
7) I wish you could save quicksound keywords in .gsp files like you could in .gig\'s.
8) Mapped network drive support is poor, and this is not good given that the best way to run GSt is on a separate machine networked to your DAW. I don\'t care if I can run gigs over a network drive (I heard it was supposed to work but doesn\'t here, anyway I won\'t ever do that over my 100mb network). I just want good support for .gsp performances saved on the network with my other project files. This sort of works, but not well: Quicksound will not search mapped drives for .gsp files, though it will find .gsp files that GSt saves IF you turn on \'monitor file activity\' and it won\'t list newly saved ones until you restart GSt. If you explicitly put a mapped network drive in the list of places to search, GSt will crash building the index.
9) This isn\'t UI, but I\'d like to see improvements to the voice stealing algorithm. Some of my other gear lets me limit the number of voices that can be applied to a single note, but with GS, I can burn all the polyphony on a single note repeated with the sustain pedal down, and then when it does start stealing, it\'s about the worst sounding stealing of all my gear. Seems like other gear makes a better choice about what to steal and what happens when they do it.
Whew! That\'s it for now, I\'m sure I\'m forgetting some... This isn\'t meant as a rant, GS is an indispensible tool here, it\'s a great sounding engine with great features, for me totally solid in normal use, tight timing, best libraries, etc, but there\'s still room for improvement.