Glenn Stanton

the separation files are generated and then removed on render. however a couple of factors can play into it. "rarely" in my lingo means "sometimes" and this could result in those files ? and as mettelus mentions - "caching" beyond memory allocation. lastly, i've seen issues with cloud storage where if the cache is in a cloud service folder (like Documents) then it's possible OneDrive (e.g.) backs them up and restores them (you know, trying to keep your files safe and all) ?

depends on the soloist ? i find i have maybe 4 or 5 "people" on the RT side who consistently produce usable solos. but the chord progression and the backing style can impact things as well as tempo. i find myself sometimes going 1/2 speed and making the progressions 1/2 as long and get better results than if i'm using full tempo (even if the tempo and changes are roughly equiv @ half speed). then when in doubt, i'll try some other tools like EZ Keys (MIDI & re-instrument), Scaler, and sometimes, gasp, just play it myself and then edit (and edit).

thanks. i don't use it myself (tried it, found it lacking). i am simple and either use the CbB Matrix for snippets or Kontakt.

maybe share some tips for BIAB - there are quite a few folks on the forum (myself quite a lot) who would like to see how other people are using it beside the DAW plugin. for myself - i'm either exporting stems of Real Tracks performance (solos mainly) or purely the MIDI (also usually solos). depending on the quality of the RT performance, i might use Melodyne to extract the MIDI and re-instrument it.

i keep it on my "quick access" list (along with a few other products which cache renderings in the documents folder - sometime mine, sometimes the "public" user...) so i am reminded to periodically clean them up. ensuring you render the Melodyne region when you're done, the clean up is mostly done for you. if you don't render it and close the project or delete the track w/o rendering it, you get leftover separations.

maybe use seperate tracks? MIDI for the, erm, MIDI source, and an audio one for, ah, the audio (which may then be exposed as a sidechain option?)

it's possible that the couple of subdirectories for the Karzog plugins are confusing things . not sure. i see several of my vst3 plugins have the directory and seem to work fine. those plugins are instruments though, so maybe there is an issue with FX vs VI? @Noel Borthwick

yeah the Karzog support says lot's of people doing it now. so just remove the .vst3 from the folder name and it should be good. the only issue might be when content is stored in the same directory and the path is hardcoded...

no, it's supposedly Steinberg's fault. according to this response from Karzog support: of course, this seems weird as i have at least 800 VST3 and none of them do this poorly thought-out folder naming. @Noel Borthwick @Jesse Jost any thoughts on this?

i know windows can be sensitive to adding known program extensions to folders. not sure why developers do it given the "standard" is Common Files -> VST3 ? but sometimes people get too clever for their own good i guess. for conspiracy minded folks - maybe someone in a competing DAW noticed this defect in CbB and decided to tell all their friends to add the ".vst3" to their folders for "cleaner" housekeeping (wink wink) and then knowing they don't support "free" DAW, that CbB users will be frustrated... ?

couple of thoughts - use Ripple Edit (turn on when needed, turn off immediately after), and use the project > insert time/measures as i think there are more options there. remember to turn off Ripple Edit after you make the move otherwise you can really skew the project... as to why "slide" is misbehaving, no idea, never use it.

so, just like a number of 70's Kung Fu films (as well as any number of PBS translated TV series)... ?

the latest versions of the Karzog plugins use a ".vst3" extension on the folder which causes the scanning of those plugins to pass, but not actually reach the real plugin DLL themselves. i've notified the Karzog support folks. in the meantime, if you remove the ".vst3" from the folder they will work. the VST 2 are not affected as they don't have that extension.

are you using a mixer to record the vocal and listen to the recorded material? if so, you're likely mixing them there and that is being fed into the vocal track. if you have an IO unit with a separate monitoring output - maybe your monitoring is too loud and that is getting pickup by the mic. another option is you have some software mixer with your audio card which is mixing them. and lastly you have something routing into your vocal track from the instrument track that is perhaps your monitoring source. i.e. you want to hear both vocals and instruments, and have routed the instruments and vocals together (like a patch point for example or aux track).

pretty sure using sidechain compression has been around for at least 40 years and longer to duck things like: bass from kick drums, instruments from vocals and solos, reverbs/delays from vocals and solos, etc. in combination with delays on reverb and echoes for either clarity or "slap" purposes. and one suspects this was common across genres...

make sure everything being exported is output to the "master buss" OR the speakers. if any of your tracks are going only to the speakers and the export is happening only on the master buss, then you will be missing parts.

+1 on Dell - until just this past year, i only used Dell computers (and since like 1990) and my previous laptop was a secondhand-me-down Dell which once cleaned, new SSD drives (replaced DVD drive with SSD OS disk and used higher performing internal connect for the project drive), and OS (W10), no option to expand memory on this version beyond the 8GB it had, it worked perfectly for several years as a DAW machine. the Dell is still alive and sits under my HP at the moment. i only really updated to the HP because a) was on sale, b) i needed it more for my studio design work (vector CAD and rendering) (and it brought my processing down into the seconds and minutes vs hour or more rendering - a CPU-intensive CAD and a GPU-based rendering product).

so try this: ALWAYS connect your bluetooth headset BEFORE opening this particular project ? or create a new clean project as it seems like it's corrupted for some reason (based your comment that ALL OTHER projects WORK). I doubt it's a "bug" since it's only the one project, and only when you don't connect your headset before opening it. to me, that says something is wrong in the project.

for composing and arranging outside of the DAW, i use Hookpad, Musescore, Scaler 2, and depending on granularity, the NI Session instrument patterns, EZ products (generate and grid edit), and the Ample "riff" mechanisms to get instrument specific "subsections". then some additional bits like Arpeggiators, Chordz, and other note to chord type translators to add elements. for the EZ Keys, Bass, Drums, i'll re-create the arrangement chords in EZK, get it the way i like it, then take it over the EZB and EZD (if i think the drums could be more interesting), and get those to where i think they'll play well. usually, if the arrangements in MIDI are good, then importing and applying those tracks in my recording template, are fairly fast. some selection of instruments, and some edits, but if there are significant arrangement changes that i want to make, i simply go back to the MIDI source(s) and regenerate, re-import into the project with the already selected instrumentation. this keeps things sync'd consistently for only some moderate extra work. most of the rendering is then done in the DAW and i can export the tracks (as well as intermediate WAV and MP3) for import into the mix template (which has all the routing, typical effects and settings etc). which at this point is effectively the same as if getting client files to mix.

i only really use the MIDI take lanes when i'm recording someone who want to do multiple takes over a given loop, but most times its much simpler - do the take, ok? try again, ok? repeat... ? because in those situations i'm recording audio of the instrument (keys, guitar, MIDI horn etc) and sometimes their vocals as well. for myself, in general, i do all the MIDI composition outside of the DAW.

it usually a workspace setting which is preventing the auto-scrolling.

i should have mentioned the unmasking EQ stuff is usually very small (~1db) type moves across several instruments (if needed) and some clip gain adjustments to elevate slightly where needed as well (like that errant tambourine hit - i might reduce that one or two to get it lower - 1db or so and maybe boost 1db on that instrument being masked). so 99% of the time is small moves - for most songs - i depend heavily on arrangement to get a good mix...

usually just #1 but if i have a dense mix i'll use my Izotope tools to find and unmask neutron/nectar. ultimately though i also check with MP3 listens (either via the Ozone Codec or an exported MP3) since the MP3 algorithm (even on its highest setting of 320K) will make choices on masking and even small mix changes can have significant results in the MP3 output (one example comes to mind - a tambourine hit which is just barely there in the mix suddenly becomes the next loudest thing in an MP3 - why? - the algorithm felt it was the strongest signal that the particular frequency and then suppressed other). so, i would say checking your target output file type is critical. and if you're streaming, recheck there as well - some streaming services (in an effort to make things "sound better") seem to do some EQ'ing and loudness adjustments (perhaps they call it "mastering") even if you don't want it. for example, Reverbnation - seems to do some slight bandwidth reduction (stuff seems slightly crunchy compared to my source file) cause a very slight darkening on the material, otoh, Broadjam seems to add some additional HF and loudness making things brighter and louder (regardless of the LUFS i submit - whether -12 or -14). Reverbnation target artists who want to sell their stuff, gigs, etc; Broadjam is about getting paid to help find deals for your music so tries to make it "sound better" to "assist". maybe i'm incorrect but that's my observations using those two services.

MIDI take lanes ?

as always - check your buffers settings - low latency with plugin hogs can yield unsatisfactory behaviour ? as a general rule - set the buffers to as high as possible that you are comfortable with. for me that is 2048. and most times even with Nectar (major hog) and Abbey Roads Chambers - seldom over 10% CPU.