XHip--Please finish your synth!!
-
- KVRian
- 889 posts since 10 Jul, 2004
thx dude sounds great but even at 3x (playing chords)it uses up all my cpu ...the atom 1.6 is just the equivalent of a pentium 800-900mhz.gsoto wrote:This patch sounds decent with 3x unison:_starcraft_ wrote:on my portable i cant really use xhips unison (too heavy)...so i'll have to just resort to a chorus effect or dust off synth1 for choirs.
gsoto_synth_ahhs_2.adxi
gsoto_synth_ahhs_2.mp3
xhip should have unison per oscillator like in sylenth1, zebra, surge etc (and also like in your mockup GUI
so maybe its in the plans 
...that way it would use singnificantly less cpu i believe. (as it only multiplies the oscillator and not everything else too). global unison is really a bitch. i tried sylenth1 this weekend and made a huge choir there with 2 half squares (8x voice unison/detune on each)...and it not even use half of the cpu. might even join the groupbuy
-
- KVRAF
- 2029 posts since 21 Jul, 2004
Dude, you don't have to use full unison, if you are dealing with saw waves. The pulsewidth control functions as a "supersaw" feature when you have ramp selected as an oscillator._starcraft_ wrote:thx dude sounds great but even at 3x (playing chords)it uses up all my cpu ...the atom 1.6 is just the equivalent of a pentium 800-900mhz.gsoto wrote:This patch sounds decent with 3x unison:_starcraft_ wrote:on my portable i cant really use xhips unison (too heavy)...so i'll have to just resort to a chorus effect or dust off synth1 for choirs.
gsoto_synth_ahhs_2.adxi
gsoto_synth_ahhs_2.mp3
xhip should have unison per oscillator like in sylenth1, zebra, surge etc (and also like in your mockup GUI
i tried sylenth1 this weekend and made a huge choir there with 2 half squares (8x voice unison/detune on each)...and it not even use half of the cpu. might even join the groupbuy
Do not lick the fablanky
-
- KVRian
- 889 posts since 10 Jul, 2004
funkadil wrote:Dude, you don't have to use full unison, if you are dealing with saw waves. The pulsewidth control functions as a "supersaw" feature when you have ramp selected as an oscillator._starcraft_ wrote:thx dude sounds great but even at 3x (playing chords)it uses up all my cpu ...the atom 1.6 is just the equivalent of a pentium 800-900mhz.gsoto wrote:This patch sounds decent with 3x unison:_starcraft_ wrote:on my portable i cant really use xhips unison (too heavy)...so i'll have to just resort to a chorus effect or dust off synth1 for choirs.
gsoto_synth_ahhs_2.adxi
gsoto_synth_ahhs_2.mp3
xhip should have unison per oscillator like in sylenth1, zebra, surge etc (and also like in your mockup GUI
i tried sylenth1 this weekend and made a huge choir there with 2 half squares (8x voice unison/detune on each)...and it not even use half of the cpu. might even join the groupbuy
cant u guys even read? i'm using PULSES(and using the pw as pw!) not saws.
we are talkin about making choirs not supersaws .
-
- KVRian
- 889 posts since 10 Jul, 2004
Mutant wrote:xhip should be xhip
full unison is what it makes it so nice
imho ofc
you are clueless then if thats what u really believe.
global unison is used in many other synths (oatmeal,synth1,korgs legacy,twin etc etc).
when it was implemented (and guess what it was my request) it was the standard way of doing it (just multiply all ).....but as new synths have shown....its more efficient to work at oscillator level.
AD was probably first to notice this when he introduced that quick supersaw detuned 16x stack on ramp.(which uses a lot less cpu than it would take to make the same sound in global unison) unfortuntely unison is not useful only for stacking saws....
imo the unison in xhip at the momemnt is the weakest point : first cos it don't even get saved with the rest of the patch, 2nd cos its too cpu intesive (as its global), 3rd cos it misses stereo spread, 4th cos it has those controls like distribution and randomization which ..hmm..i would remove :p
-
- KVRAF
- 2910 posts since 26 Jul, 2005 from dun unda
That's true. It's a common misconception that unison is only good for supersaws, simply because it's been attached to such sound so many times in review media. At times, it does so much of a better job than using chorus on sounds. (Though Oatmeal's chorus effect was pretty good for a unison substitute, until Fuzzpilz implemented "hacky" chorus, which by the way might seem global but is not global unison since it screws up with Sync/FM OSC modes. (ie: unwanted results.)_starcraft_ wrote:Mutant wrote:xhip should be xhip
full unison is what it makes it so nice
imho ofc
you are clueless then if thats what u really believe.
global unison is used in many other synths (oatmeal,synth1,korgs legacy,twin etc etc).
when it was implemented (and guess what it was my request) it was the standard way of doing it (just multiply all ).....but as new synths have shown....its more efficient to work at oscillator level.
AD was probably first to notice this when he introduced that quick supersaw detuned 16x stack on ramp.(which uses a lot less cpu than it would take to make the same sound in global unison) unfortuntely unison is not useful only for stacking saws....
That said, synth1 uses two types of unison if I recall; either one uses only a pinch of CPU.
-
- KVRAF
- 3499 posts since 9 Oct, 2004 from Poland
I stand by what i said._starcraft_ wrote:Mutant wrote:xhip should be xhip
full unison is what it makes it so nice
imho ofc
you are clueless then if thats what u really believe.
global unison is used in many other synths (oatmeal,synth1,korgs legacy,twin etc etc).
when it was implemented (and guess what it was my request) it was the standard way of doing it (just multiply all ).....but as new synths have shown....its more efficient to work at oscillator level.
AD was probably first to notice this when he introduced that quick supersaw detuned 16x stack on ramp.(which uses a lot less cpu than it would take to make the same sound in global unison) unfortuntely unison is not useful only for stacking saws....
imo the unison in xhip at the momemnt is the weakest point : first cos it don't even get saved with the rest of the patch, 2nd cos its too cpu intesive (as its global), 3rd cos it misses stereo spread, 4th cos it has those controls like distribution and randomization which ..hmm..i would remove :p
I want Xhip to stay this way.
About that "clueless"...
I know exactly what is the difference between these 2 types of unison.
If i want Synth1 CPU efficiency i go to Synth1, if i want Xhip superior sound i go to Xhip.
EOT for me.
If you want to argue you can argue with someone else...
[====[\\\\\\\\]>------,
Ay caramba !
Ay caramba !
-
- KVRian
- 889 posts since 10 Jul, 2004
synth1 is not the right example as it doesnt have voice stacking at the oscillators.(sylenth1 does!) so its clear you are completely clueless.Mutant wrote:
I stand by what i said.
I want Xhip to stay this way.
About that "clueless"...
I know exactly what is the difference between these 2 types of unison.
If i want Synth1 CPU efficiency i go to Synth1, if i want Xhip superior sound i go to Xhip.
EOT for me.
If you want to argue you can argue with someone else...
unison at oscillator level would not affect the sound or quality of xhip. (unison if implemented corrctly is just sum of the parts....) and wasting cycles on 16 enevlopes and 16 filters when u stacking 16 oscillators is just that....a waste of cycles..cos having 16 envelopes giving the same value to each voice is just the same as having 1 enevelope at the end of 16 oscillators.
only difference with global unison and unison at oscillator level is that in the second case u have to adjust voices and detune for each oscillator....which takes a little more time...but significantly less cpu ...and also more control over sound.
does the xhip unison for ramp (through pw)at oscillator level sound any worse than doing the same thing in the global unison section? only difference is that in global unison it would take double more cpu.
now i'm done explaining things to u like u were a 5 year old....
-
- KVRAF
- 1669 posts since 4 Nov, 2007
no really, what you exlained is not entirely rigth.
i will wait sg to correct you cause im a lazy ass.
cheers.
i will wait sg to correct you cause im a lazy ass.
cheers.
-
- KVRAF
- 12615 posts since 7 Dec, 2004
the oscillators cant possibly have a unison feature implemented without adding controls for that to the gui.
in xhip2 i'll make it dynamic with a list of oscillators so you can "add" oscillators and routings. just take note that if you wanted 16 pwm waveforms you'd need 16 lfo routings and a bunch of other stuff, it's nearly the same as global unison.
if the filter and other effects are turned OFF they actually do not process - compare for example a plain ramp on global unison (16 voices) to the 'supersaw', you should see only about 30% savings, and that only applies because the 'supersaw' is the optimal configuration for saving processing. with an oscillator unison of pulses there will be more like 5% savings. triangle and sine will have absolutely no savings.
there is about 50x more processing applied to scaling modulators (as for the pwm, oscillator tuning and etc) than there is in the vca, panning and so on. the filter for example uses about 30% in it's core and 70% due to modulation scaling. one way to avoid this is to block-process, but then xhip would sound just as shit as other software synths.
in xhip2 i'll make it dynamic with a list of oscillators so you can "add" oscillators and routings. just take note that if you wanted 16 pwm waveforms you'd need 16 lfo routings and a bunch of other stuff, it's nearly the same as global unison.
if the filter and other effects are turned OFF they actually do not process - compare for example a plain ramp on global unison (16 voices) to the 'supersaw', you should see only about 30% savings, and that only applies because the 'supersaw' is the optimal configuration for saving processing. with an oscillator unison of pulses there will be more like 5% savings. triangle and sine will have absolutely no savings.
there is about 50x more processing applied to scaling modulators (as for the pwm, oscillator tuning and etc) than there is in the vca, panning and so on. the filter for example uses about 30% in it's core and 70% due to modulation scaling. one way to avoid this is to block-process, but then xhip would sound just as shit as other software synths.
-
- KVRian
- 889 posts since 10 Jul, 2004
yes not pushing to have all that implemented any time soon.
its probably just me and few other masochists running shit mobile cpus for audio.....but i'm sure cpu will return to be an issue for the vast majority when these touch smart phones inprove their hardware and start supporting major operationg systems....then we will see ppl making music on them.
at present i just wish to see the GUI finished : fix of dropdown list boxes, left-right movement added for knobs, neutral reset positions, left click +ctrl to reset parameter etc......then to me it feel like a 1.0 release.
btw the site is downn ...i hope u not have bandwith problems....i actually wanted to check if u set up a paypal for donations yet.
ps: i also think its time to set up a subforum
...this thread is really awwwful.
its probably just me and few other masochists running shit mobile cpus for audio.....but i'm sure cpu will return to be an issue for the vast majority when these touch smart phones inprove their hardware and start supporting major operationg systems....then we will see ppl making music on them.
at present i just wish to see the GUI finished : fix of dropdown list boxes, left-right movement added for knobs, neutral reset positions, left click +ctrl to reset parameter etc......then to me it feel like a 1.0 release.
btw the site is downn ...i hope u not have bandwith problems....i actually wanted to check if u set up a paypal for donations yet.
ps: i also think its time to set up a subforum
...this thread is really awwwful.-
- KVRAF
- 12615 posts since 7 Dec, 2004
the things you want done are no problem, that's sort of a five minute job. it's just the version of the code currently doesn't save patches so it can't be "released" even here in the thread. i can make those changes and send you the binary. i'll make the changes actually just to get it out of the way. before i can release anything i need to make it save proper float format patches in the new patch format - it'll be a chunk format like 'riff' or '3ds' so that things can easily be added or removed with less assumptions being made in the code.
(currently a bad patch can crash xhip easily..even bad input from the host can!)
the site is down because the power went out, and i was fiddling with a modprobe config file at that moment.. so now i'm trying to figure out how to restore the files to the way they were, i think some script is running in init i just don't know how to remove it. it might be a day or two, or maybe i'll ask in a debian channel for some help on this one.
i was gone for the last weekend and came back to find it like this.. it was a bad idea to leave the server half-configured like that. it can't be booted now and i don't have any backup, but i do have access to the whole system so there is no real problem. if it were the worst case i could copy the site across to a new installation in an hour, i just want to learn how to fix this instead. (coder masochism)
(currently a bad patch can crash xhip easily..even bad input from the host can!)
the site is down because the power went out, and i was fiddling with a modprobe config file at that moment.. so now i'm trying to figure out how to restore the files to the way they were, i think some script is running in init i just don't know how to remove it. it might be a day or two, or maybe i'll ask in a debian channel for some help on this one.
i was gone for the last weekend and came back to find it like this.. it was a bad idea to leave the server half-configured like that. it can't be booted now and i don't have any backup, but i do have access to the whole system so there is no real problem. if it were the worst case i could copy the site across to a new installation in an hour, i just want to learn how to fix this instead. (coder masochism)
-
- KVRAF
- 3928 posts since 23 Oct, 2005 from vassalboro, maine