| *** edogawa has quit IRC | 00:01 | |
| *** ricardocrudo has quit IRC | 00:20 | |
| *** ColaEuphoria has quit IRC | 00:27 | |
| *** Magnus_RM has quit IRC | 00:40 | |
| *** rncbc has quit IRC | 00:42 | |
| *** Magnus_RM has joined #lv2 | 02:22 | |
| *** drobilla has quit IRC | 03:27 | |
| *** Magnus_RM has quit IRC | 03:34 | |
| *** falktx` has joined #lv2 | 04:23 | |
| *** falktx has quit IRC | 04:27 | |
| *** artfwo has quit IRC | 05:45 | |
| rgareus | why is there http://lv2plug.in/ns/ext/buf-size/#fixedBlockLength ? | 09:57 |
|---|---|---|
| rgareus | it's equivalent to bufsz:minBlockLength == bufsz:maxBlockLength | 09:57 |
| rgareus | and even requires the latter | 09:58 |
| *** edogawa has joined #lv2 | 10:13 | |
| *** ricardocrudo has joined #lv2 | 10:31 | |
| *** ricardocrudo has quit IRC | 11:44 | |
| *** rncbc has joined #lv2 | 11:56 | |
| *** ricardocrudo has joined #lv2 | 11:57 | |
| falktx` | that's a feature, not a property | 12:31 |
| falktx` | same with boundedBlockLength and powerOf2 | 12:32 |
| rgareus | it's reduncant | 12:45 |
| rgareus | fixedBlockLength says a plugin must set minBlockLength == bufsz:maxBlockLength. | 12:46 |
| rgareus | what extra information does fixedBlockLength provide if minBlockLength is already equal to maxBlockLength ? | 12:46 |
| rgareus | I'm about to outline optionalFeature "NoSampleAccurateControl" -> inform the host: "don't split the process-cycle on my account (to provide me with sample-accurate automation)" | 12:48 |
| rgareus | IOW. the plugin *prefers* a fixed buffer. | 12:48 |
| falktx` | fixedBlockLength is for plugins that want a fixed number of frames | 12:52 |
| falktx` | if that is not set as a required feature, the host can do as he pleases | 12:52 |
| falktx` | for plugins that set fixedBlockLength, we need way to ask if it support changes (like jack buf size callbacks) | 12:53 |
| falktx` | that's the thing that's missing right now | 12:53 |
| rgareus | falktx`: what's the difference between a plugin in the .ttl specifying fixedBlockLength vs plugin in the ttl specifying minBlockLength = number maxBlockLength = same number? | 13:28 |
| rgareus | http://lv2plug.in/ns/ext/buf-size/#fixedBlockLength says "This length MUST be provided as the value of both the bufsz:minBlockLength and bufsz:maxBlockLength options." | 13:29 |
| rgareus | is ther a case where bufsz:minBlockLength has the same value as bufsz:maxBlockLength that's different from having fixedBlockLength ? | 13:29 |
| falktx` | ? | 13:30 |
| rgareus | falktx`: fixedBlockLength, is always a fixed number. we don't need a way to change it. | 13:30 |
| falktx` | I don't want to discuss this with you today | 13:30 |
| rgareus | that'd break it by design | 13:30 |
| falktx` | I thought this was all pretty obvious | 13:31 |
| falktx` | you always get crazy about semantics :/ | 13:31 |
| falktx` | discuss this in the mailing list please | 13:31 |
| * rgareus was interested why there's fixedBlockLength at all. it's redundant | 13:31 | |
| falktx` | I think the spec is incomplete | 13:31 |
| falktx` | there's missing bits to make it all clear | 13:31 |
| rgareus | I don't wanto get rid of it. I was just curious. | 13:31 |
| falktx` | anyway, got other stuff to do. need to finish a new plugin.. | 13:32 |
| rgareus | and if it's indeed redudant.. we should consider to deprecate it | 13:32 |
| rgareus | falktx`: new plugin? | 13:32 |
| falktx` | fluidplug | 13:33 |
| falktx` | loading sf2 files directly as lv2 plugin | 13:33 |
| rgareus | falktx`: ooh. a synth. | 13:33 |
| rgareus | cool | 13:33 |
| rgareus | falktx`: disk-streaming or all in RAM? | 13:33 |
| falktx` | it's fluidsynth. so all in ram | 13:33 |
| falktx` | linuxsampler is the one that does disk streaming | 13:33 |
| rgareus | falktx`: how does it differ from calf fluidsynth? | 13:33 |
| falktx` | no UI to load a file. the plugin itself already includes the sf2 in the bundle | 13:34 |
| falktx` | works as-is. just load it | 13:34 |
| rgareus | nice | 13:34 |
| rgareus | oh wait - included in bundle... why is that? | 13:34 |
| rgareus | can't you just use existing LV2 patch:set to specify a file? | 13:34 |
| falktx` | well, could be a symlink | 13:34 |
| falktx` | I want a 1:1 sf2 to plugin thing | 13:35 |
| falktx` | it's made specially for mod | 13:35 |
| rgareus | falktx`: can factory-presets set the file? (e.g you can include multiple sf2 in the bundle, factory preset selects) | 13:35 |
| falktx` | you'll see soon | 13:36 |
| falktx` | but as I have it right now the programs get exported as a lv2 control port | 13:36 |
| rgareus | sometimes I wish the calf plugins could just be fixed and "work as-is" reliably. | 13:37 |
| falktx` | they at least are working to get rid of instance-data | 13:38 |
| *** bgola has joined #lv2 | 13:41 | |
| falktx` | this will be the first time I try fluidsynth on mod. let's see how it goes.. | 13:42 |
| *** bgola_ has quit IRC | 13:43 | |
| *** deva has joined #lv2 | 15:30 | |
| *** nordstrom has quit IRC | 15:56 | |
| *** nordstrom has joined #lv2 | 16:14 | |
| *** Magnus_RM has joined #lv2 | 16:51 | |
| *** coderkevin has quit IRC | 17:19 | |
| *** coderkevin has joined #lv2 | 17:20 | |
| *** ventosus has joined #lv2 | 17:38 | |
| *** falktx|work has joined #lv2 | 17:54 | |
| *** deva has quit IRC | 18:01 | |
| falktx|work | rgareus: I like the proposed noSampleAccurateCtrl, but not sure if it's the right thing for lv2 | 18:28 |
| falktx|work | I still want the fixed-block-size plugin callback information. just because the size is 'fixed' (a better word would be static) doesn't mean it can't change | 18:29 |
| falktx|work | I think 'fixed' is a bad word to describe it. 'static' is better but still doesn't sound right | 18:30 |
| falktx|work | perhaps constant | 18:30 |
| rgareus | a fixed-block-size which can change without re-instantiating the plugin is different from "noSampleAccurateCtrl" and harder. | 19:17 |
| rgareus | "noSampleAccurateCtrl" is useful for current gx-convolver. | 19:18 |
| rgareus | and convo.lv2 | 19:18 |
| rgareus | if the block-size changes those plugins re-initialize their convolution engine. | 19:19 |
| rgareus | they can cope with variable blocksize (no pre-allocation) | 19:20 |
| rgareus | but I'm still to test a bit more if it makes sense.. "noSampleAccurateCtrl" is still a prototype | 19:21 |
| falktx|work | I'd use it in carla too | 19:48 |
| falktx|work | it matches perfectly btw | 19:48 |
| falktx|work | current code just checks for fixed bufsize https://github.com/falkTX/Carla/blob/master/source/backend/plugin/CarlaPluginLV2.cpp#L3013 | 19:49 |
| rgareus | falktx|work: does carla have automation? | 19:52 |
| falktx|work | via midi cc and osc | 19:53 |
| falktx|work | midi cc events are sample accurate | 19:53 |
| rgareus | ie jack-bufsize = 64, automation-data at sample 37. ardour does run(37); change_port_values(); run(27); | 19:53 |
| falktx|work | midi cc mapped to parameters can cause the process to be splitted | 19:53 |
| rgareus | OK | 19:53 |
| falktx|work | same | 19:53 |
| rgareus | now we only have to get drobilla to accept it as official URL and write documentation for it :) | 19:55 |
| rgareus | using LV2_CORE_PREFIX "NoSampleAccurateCtrl" was not nice of me. | 19:55 |
| rgareus | falktx|work: what namespace do you think it would fit better than lv2: ? | 19:57 |
| rgareus | I think it's in line with "inPlaceBroken" "isLive" "hardRTCapable" and fits into http://lv2plug.in/ns/lv2core/ but there may be a better place. | 19:59 |
| falktx|work | it should be a feature | 19:59 |
| falktx|work | noSampleAccurateControls | 19:59 |
| falktx|work | big names is not a issue afaik, it's only checked once during init | 19:59 |
| rgareus | it is a Feature. (I envisaged optionalFeature) | 19:59 |
| falktx|work | yes | 19:59 |
| rgareus | "...Controls" instead of "...Ctrl" is fine with me. | 20:00 |
| rgareus | falktx|work: I'll write to the ML. | 20:02 |
| rgareus | sent | 20:15 |
| falktx|work | rgareus: btw, drobilla accepted my patches directly as github pull requests before. perhaps this is something you can do as well | 20:17 |
| falktx|work | drobilla might feel lazy and just click 'merge' ;) | 20:17 |
| rgareus | good call. | 20:19 |
| * rgareus loves git send-email | 20:29 | |
| rgareus | falktx|work: would you know if "a lv2:optionalFeature ;" in my patch is correct? | 21:02 |
| rgareus | others use "a lv2:Feature" | 21:03 |
| rgareus | mmh sord_validate actually flags it :( | 21:03 |
| *** ventosus has left #lv2 | 21:07 | |
| falktx|work | yeah, it needs to be lv2:Feature | 21:11 |
| rgareus | that would allow using it as requiredFeature | 21:14 |
| falktx|work | the same applies to hardrtcapable | 21:14 |
| falktx|work | having it as requiredFeature means what? | 21:15 |
| rgareus | if the host does not support it, the plugin must fail to instantiate | 21:15 |
| rgareus | or the host must refuse to load the plugin. | 21:15 |
| falktx|work | case 1 doesn't make sense | 21:16 |
| falktx|work | even if the host is non-rt, it can still load fullrt plugins | 21:17 |
| falktx|work | for the plugin it makes no difference | 21:17 |
| falktx|work | host refusing to load the plugin yes, that sounds ok | 21:17 |
| rgareus | indeed. | 21:17 |
| rgareus | well it makes no sense to *require* noSampleAccurateControl in the first place since it's just a hint. | 21:18 |
| rgareus | anyway, lv2:Feature it is. | 21:19 |
| * rgareus waits for drobilla to return from his week-end trip.. | 21:20 | |
| * falktx|work likes http://dev.drobilla.net/changeset/59601f4e1cf94c0062e6e3d343822142ec5ffecb/lilv | 21:21 | |
| * rgareus too | 21:23 | |
| rgareus | it'll probably break a few plugins at first. | 21:23 |
| rgareus | falktx|work: ite even prints the file that breaks (see in http://pastebin.com/rKi7e0pW) | 21:27 |
| rgareus | though as drobilla said: no line number | 21:28 |
| *** ricardocrudo has quit IRC | 21:37 | |
| *** Magnus_RM has quit IRC | 21:40 | |
| *** falktx|work has quit IRC | 22:11 | |
| *** edogawa has quit IRC | 22:47 | |
| *** rncbc has quit IRC | 23:15 | |
Generated by irclog2html.py 2.13.0 by Marius Gedminas - find it at mg.pov.lt!