XHip--Please finish your synth!!
- KVRAF
- 12615 posts since 7 Dec, 2004
http://xhip.cjb.net/xhip/releases/v0/b6 ... 6.14.4.dll
- fixed sine amp modulation problem
- fiddled with pcm initialization, might fix the crashing.
about the problem with using the trigger for filter oscillation, i understand that you need to use it to initially get the thing to oscillate, but then you want it not to click. the solution would be to trigger the filter only when the voice state == cut on a new note-on, i guess. that is a really unnatural behaviour though and it would be better if you could manage by using some noise or other low level input to get it going.
http://xhip.cjb.net/temp/public/badsine-bad.adxi - random filter thing patch using sine-amp modulation showing that i (i hope) fixed the problem.
http://xhip.cjb.net/temp/public/nopopy.adxi - ramp used at same frequency to get it going. the harmonics are low enough level that you really cant tell the ramp is there at all. remember that if you tune a resonance to the same frequency as an input you can get a >1 strength and this will produce an ever increasing amplitude. in the filter of course it only increases until you hit the clipping point, then it is stable. using a higher input level waveform can increase the strength meaning that you get a faster exponential attack, however it then becomes more audible at the same time.
- fixed sine amp modulation problem
- fiddled with pcm initialization, might fix the crashing.
about the problem with using the trigger for filter oscillation, i understand that you need to use it to initially get the thing to oscillate, but then you want it not to click. the solution would be to trigger the filter only when the voice state == cut on a new note-on, i guess. that is a really unnatural behaviour though and it would be better if you could manage by using some noise or other low level input to get it going.
http://xhip.cjb.net/temp/public/badsine-bad.adxi - random filter thing patch using sine-amp modulation showing that i (i hope) fixed the problem.
http://xhip.cjb.net/temp/public/nopopy.adxi - ramp used at same frequency to get it going. the harmonics are low enough level that you really cant tell the ramp is there at all. remember that if you tune a resonance to the same frequency as an input you can get a >1 strength and this will produce an ever increasing amplitude. in the filter of course it only increases until you hit the clipping point, then it is stable. using a higher input level waveform can increase the strength meaning that you get a faster exponential attack, however it then becomes more audible at the same time.
-
- KVRian
- 673 posts since 15 Nov, 2004 from Montevideo, Uruguay
Ok, that solves it. 
A question regarding the GUI. Do the osc invert controls make more sense attached to the mixer or the oscillator itself?
Yep, I meant that. Just open Xhip, set FM mode to "A!->B" and change to program #2. The box still displays "A!->B". Happens with glide and sync mode too.aciddose wrote:if you mean it doesnt change sometimes for any of the other boxes.. i've never noticed that and i cant reproduce it here. can you reproduce it?
A question regarding the GUI. Do the osc invert controls make more sense attached to the mixer or the oscillator itself?
- KVRAF
- 12615 posts since 7 Dec, 2004
they make more sense in the mixer. what they do is they just invert the value output by the gain sliders. i could change it back to percentages and have only the gain sliders with center = 0db, left = -48db inverted and right = -48db normal, just like the mod depth sliders. that may not make sense to some people though.
i tested the drop box thing and yes that is very strange. it seems like there is something going on which is related to the value of the parameter. it only does this when you're switching to a patch where the parameter's value is == 0. when you switch back and forth again however it works fine. it's probably a simple bug somewhere here.. i'll try to fix that too then.
i tested the drop box thing and yes that is very strange. it seems like there is something going on which is related to the value of the parameter. it only does this when you're switching to a patch where the parameter's value is == 0. when you switch back and forth again however it works fine. it's probably a simple bug somewhere here.. i'll try to fix that too then.
- KVRAF
- 12615 posts since 7 Dec, 2004
i did some more messing around and really when i did the drop box thing i didnt go all the way. it still isnt ready for prime-time and i only meant to leave it on the gui for a single revision. i forgot and somehow figured it was in there and worked "ok" so i'd keep it.
anyway i've restructured the code a bit and i'm almost done with the proper implementation of the drop boxes. since i'm not done however i quickly implemented something people have asked for since the beginning: i finally have the code in place which has enabled me to do this in about two minutes
.
anyway i've always said "there is no point adding stuff to the gui since i'll have to rewrite it ten times.. i have to implement things correctly first".. well, now they're implemented correctly.
http://xhip.cjb.net/xhip/releases/v0/b6 ... 6.14.5.dll

i might put the boxes back when they're done. probably at that point i'll be able to really think about implementing more guis though rather than having just this one. another exciting thing to think about in the future might be to implement some kind of textual format for creating the gui controls. that way it would become a lot more skin-able. i would need to get all the widgets working with identical abstract behaviours so that they can be interchanged seamlessly. that is in the next while though and at the moment i really need to follow the todo list
(slightly sad note: the pcm selection obviously is forced back to the pulse-width slider once again, for now.)
anyway i've restructured the code a bit and i'm almost done with the proper implementation of the drop boxes. since i'm not done however i quickly implemented something people have asked for since the beginning: i finally have the code in place which has enabled me to do this in about two minutes
anyway i've always said "there is no point adding stuff to the gui since i'll have to rewrite it ten times.. i have to implement things correctly first".. well, now they're implemented correctly.
http://xhip.cjb.net/xhip/releases/v0/b6 ... 6.14.5.dll

i might put the boxes back when they're done. probably at that point i'll be able to really think about implementing more guis though rather than having just this one. another exciting thing to think about in the future might be to implement some kind of textual format for creating the gui controls. that way it would become a lot more skin-able. i would need to get all the widgets working with identical abstract behaviours so that they can be interchanged seamlessly. that is in the next while though and at the moment i really need to follow the todo list
(slightly sad note: the pcm selection obviously is forced back to the pulse-width slider once again, for now.)
-
- The Teach
- 8273 posts since 23 Jul, 2002 from flatness
just getting myself back in another loop
aciddose - what happened (or didnt happen) regarding the tentative linux port - i have a version from the end of january but did it get any further than that ?
slainte
rob
aciddose - what happened (or didnt happen) regarding the tentative linux port - i have a version from the end of january but did it get any further than that ?
slainte
-
- KVRian
- 1028 posts since 20 Jun, 2007
can anyone point me in the direction of the preset banks that everyone made?
- KVRAF
- 12615 posts since 7 Dec, 2004
"what happened (or didnt happen) regarding the tentative linux port"
there is no point in spending time on a port until i have a stable windows version. the port will happen either just before or soon after beta 7 is released.
there is no point in spending time on a port until i have a stable windows version. the port will happen either just before or soon after beta 7 is released.
-
- The Teach
- 8273 posts since 23 Jul, 2002 from flatness
coolaciddose wrote:"what happened (or didnt happen) regarding the tentative linux port"
there is no point in spending time on a port until i have a stable windows version. the port will happen either just before or soon after beta 7 is released.
sounds more than fair enough to me
like i said ive been away for a bit and just wondered what the state of play was - no crit implied
slainte
- KVRAF
- 12615 posts since 7 Dec, 2004
on the other hand if somebody wanted to volunteer to get a linux version working and maintain that for me the code is here and available to you. it is just there are not all that many people around with the necessary skills. the only things that need to be done at the moment are some minor changes to the files i use as replacements for the "vst sdk" files and a few changes here and there to make sure there isnt any reference to windows.h/etc in the code. actually i'll give it another go i suppose. the last time i did this "port" i spend about 30 minutes adjusting the code and makefiles to work on linux and windows without changes (ifdefs, etc) and afterward i accidentally over-wrote the modified code with the windows code.
the problem is i need to make the changes and update the main source without doing anything else and the last time i fixed some bugs on the windows side without first updating the main source - so when i copied the source to get the bug fixes onto the linux side i over-wrote the modified makefiles and so on. i've just been pissed off ever since that so i've wanted to get the code into a fixed state while i do it.
i'll definitely have another go at it now though since there isnt all that much i'm working on with the windows side at the moment.
the problem is i need to make the changes and update the main source without doing anything else and the last time i fixed some bugs on the windows side without first updating the main source - so when i copied the source to get the bug fixes onto the linux side i over-wrote the modified makefiles and so on. i've just been pissed off ever since that so i've wanted to get the code into a fixed state while i do it.
i'll definitely have another go at it now though since there isnt all that much i'm working on with the windows side at the moment.
- KVRAF
- 12615 posts since 7 Dec, 2004
http://xhip.cjb.net/xhip/releases/v0/b6 ... .6.14.5.so
well, i suppose two hours isnt all that much to have been pissed off about. you od get pissed off quite a lot when your work is trashed by one very stupid mistake though, i've just remained bitter
unfortunately it doesnt work.. this will take more than two hours now. there is a problem with the gui causing a crash. i'm not sure at all what the problem is. at least you have an up to date guiless version for linux now.
well, i suppose two hours isnt all that much to have been pissed off about. you od get pissed off quite a lot when your work is trashed by one very stupid mistake though, i've just remained bitter
unfortunately it doesnt work.. this will take more than two hours now. there is a problem with the gui causing a crash. i'm not sure at all what the problem is. at least you have an up to date guiless version for linux now.
- KVRAF
- 12615 posts since 7 Dec, 2004
anybody running linux can try this one out:
http://xhip.cjb.net/temp/public/3de0.bin
there are major problems with keyboard/mouse input and timing. the timing isnt used in xhip but now i'm starting to remember that i had to fix the keyboard/mouse and windowing code as well as make minor changes to the graphics code the last time around. i'll have to repeat those actions now - this at least shows however that the major parts of the code work just fine.
http://xhip.cjb.net/temp/public/3de0.bin
there are major problems with keyboard/mouse input and timing. the timing isnt used in xhip but now i'm starting to remember that i had to fix the keyboard/mouse and windowing code as well as make minor changes to the graphics code the last time around. i'll have to repeat those actions now - this at least shows however that the major parts of the code work just fine.
-
- KVRian
- 795 posts since 2 Apr, 2006
I'm going to try out the guiless version tonight. The previous linux version crached when opening the gui.
- KVRAF
- 12615 posts since 7 Dec, 2004
http://xhip.cjb.net/xhip/releases/v0/b6 ... 6.14.5b.so
this is a test release (there are known serious bugs!) of the gui working on linux. there is definitely a segfault which occurs when you try to close the editor, other than that i suspect there may be other serious issues.
i think i can cure the segfault with a little more effort. in this version the dropboxes are also not behaving correctly due to some core changes i made while implementing this layout test: http://xhip.cjb.net/hardware/adsx1.dll http://xhip.cjb.net/hardware/layout.v3.png
so, i'll need to fix both of those issues before i can make another release including a linux version.
the linux version also lacks the ability to save or load patches and pcms due to the fact i use windows specific functions for that on the windows side. i'll need to write an abstract version of those dialogs that will be fully portable like the rest of xhip.
so, remember this version is _useless_ because it dies when you close the gui. i suppose on a high resolution you can fit it into a corner, but just do not expect to get too much use out of this until i fix the remaining bugs.
oh, and the bug with the drop boxes: if you release the mouse button without selecting an option, an option of -1 to the current option or zero is selected. afterward you will not be able to click on any items in the dropbox until you click on the header again to close it a second time. i have not worked out the logic here.. something weird is going on and i just need to think through all of the logic to find the problem.
this is a test release (there are known serious bugs!) of the gui working on linux. there is definitely a segfault which occurs when you try to close the editor, other than that i suspect there may be other serious issues.
i think i can cure the segfault with a little more effort. in this version the dropboxes are also not behaving correctly due to some core changes i made while implementing this layout test: http://xhip.cjb.net/hardware/adsx1.dll http://xhip.cjb.net/hardware/layout.v3.png
so, i'll need to fix both of those issues before i can make another release including a linux version.
the linux version also lacks the ability to save or load patches and pcms due to the fact i use windows specific functions for that on the windows side. i'll need to write an abstract version of those dialogs that will be fully portable like the rest of xhip.
so, remember this version is _useless_ because it dies when you close the gui. i suppose on a high resolution you can fit it into a corner, but just do not expect to get too much use out of this until i fix the remaining bugs.
oh, and the bug with the drop boxes: if you release the mouse button without selecting an option, an option of -1 to the current option or zero is selected. afterward you will not be able to click on any items in the dropbox until you click on the header again to close it a second time. i have not worked out the logic here.. something weird is going on and i just need to think through all of the logic to find the problem.
-
- KVRAF
- 10815 posts since 26 Nov, 2004 from UK
is that dll supposed to make noise or is it just a teaseaciddose wrote:some core changes i made while implementing this layout test: http://xhip.cjb.net/hardware/adsx1.dll
looks kinds kool though
sorry still no Linux here
will try to test them on a live distro later
Subz
