Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2010-04-25

00:01:26 Quit DerPapst (Quit: Leaving.)
00:01:48 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
00:03:43 Join archivator [0] (
00:04:49 Quit anewuser (Quit: What do you know...THE WORLD'S first NTRQ (that's for NES/FAMICOM) tracking compo. Have powerpak? Try it out! Otherwise ROM IMAGE.)
00:08:30 Quit xavieran (Ping timeout: 258 seconds)
00:21:35 Quit merbanan (Ping timeout: 240 seconds)
00:28:08 Quit pamaury (Quit: Page closed)
00:32:50 Quit archivator (Quit: Leaving)
00:33:07 Quit ved (Quit: Changing server)
00:44:52 Join __arbingordon [0] (
00:45:39 Join ved [0] (
00:46:04 Nick __arbingordon is now known as arbingordon (
00:46:11 Quit arbingordon (Changing host)
00:46:11 Join arbingordon [0] (~w@unaffiliated/arbingordon)
00:47:05 Quit _arbingordon (Ping timeout: 240 seconds)
00:54:32 Quit n17ikh (Ping timeout: 245 seconds)
01:00:36 Join n17ikh [0] (
01:13:10***Saving seen data "./dancer.seen"
01:15:39 Quit ender` (Quit: I don't see what C++ has to do with keeping people from shooting themselves in the foot. C++ will happily load the gun, offer you a drink to steady your nerves, and help you aim. -- Peter da Silva)
01:28:56 Quit antaw (Quit: Verlassend)
01:37:20 Quit mirak_ (Quit: Ex-Chat)
01:38:05 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
01:51:56 Join saratoga [0] (~463f90ed@gateway/web/freenode/x-dcnfolmzumtgomru)
01:53:01 Quit DataGhost (Ping timeout: 264 seconds)
01:54:29saratogaUnhelpful: (for the logs) you might be interested in FS #11216
01:55:20 Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
01:55:54saratogagreat timing
02:06:42saratogaUnhelpful: i skimmed FS #11216 but don't really know enough to have an opinion
02:20:39 Quit bertrik (Quit: De groeten)
02:25:52*Strife89 will test on his devices.
02:35:01Strife89Doesn't look like it was made right, though.
02:35:10 Quit kadoban (Remote host closed the connection)
02:35:14Strife89That or a recent change messed up a file location.
02:45:42 Join kugel [0] (~kugel@rockbox/developer/kugel)
02:46:21Strife89My Fuze's battery is dead. :/
02:50:24kugelsaratoga: it's not the right way if you ask me#
02:51:07kugelI mean, I only looked at the patch quickly (not throughoutly) but it seems it could be done better
02:51:41Unhelpfulholy crow, what did they do to my loader?! (and i've only looked at the diffstat thus far)
02:52:49Unhelpfulwhat's this stuff in codec_get_buffer? why should anything in codec_thread need to be touched?
02:53:46kugelUnhelpful: please bear with him, at times some of your optimisations are strange to look at for others
02:53:48saratogai think he puts the image on the codec buffer
02:54:20Unhelpfulkugel: i'm just not clear on why this should need *huge* changes in the loader.
02:54:40Unhelpfulyou should be able to do this with a very *small* change by hooking the getc implementation.
02:54:42kugelright, that's most definitely happening, which seems strange as the actual image is already loaded as part of buffering the whole file
02:55:03kugelsaratoga: ^
02:55:50saratogaisn't because of the stupid desynchronization stuff in ID3v2?
02:55:58kugelI also find the memcmp's strange (i.e. I would use str*cmp) but that's probably minor given the scope of the patch
02:56:32kugelunless id3 tags aren't \0-terminated
02:57:21Unhelpfulkugel: the one i'm looking at appears to be against *part* of a string... is there a str*cmp for substrings?
02:58:06 Join anewuser [0] (anewuser@unaffiliated/anewuser)
02:58:25kugelI think we should have a talk with uchida about the patch, once we get to catch him in irc...
02:59:53kugelas of now, I'm *pretty* happy with how rb handles albumart which is perfectly reasonable for embedded devices as long as suitable tools exist, so I'm not really happy with accepting a hacked together solution
03:00:32kugelotoh I cannot really determine how hacky it really is, but it seems hacky in some places (especially the codec buffer thing)
03:01:03Unhelpfulsaratoga: my idea for handling tag images was something along the lines of passing the loader an alternate getc implementation - this could read in the data and do whatever transformation needs to be done, seeking past non-image data, desync, base64 decode (for some tag types)
03:01:49saratogathis would load it into the audio buffer?
03:01:53Unhelpfulor perhaps we could make this more optimal by hooking *below* getc with a read function... make the buffer big enough, and the cost of the call-via-pointer on read gets tiny.
03:02:17Unhelpfulimo it ought to load as-needed into a small chunk of buffer, as the current reader does.
03:02:35kugelthe call-via-pointer thing is really overestimated
03:03:21kugelwe really should focus on clean and understandable code instead, otherwise we'll dig our own graveyard
03:03:25Unhelpfulkugel: probably is, there is a *huge* amount of stuff going on here besides getc
03:04:49kugeleven if your function consists of 10 or more instructions, the pro- and epilog is already neglible
03:05:00kugelmost of the time
03:05:32kugelthe other option is increasing our bus factor
03:06:05kugel(which is high enough in some areas if you ask me)
03:06:18Unhelpfuli'm also *pretty* sure that getc is inlined right now. and that a read hook would not need to be so much worse. it would take a uintptr_t opaque argument in place of the file handle, and otherwise look like a "real" read.
03:07:47Unhelpfulbut the hook version would need to fill the buffer, and then transform it as needed. to my knowledge the transforms done on image data in tags are all expanding, so you should always be able to fit the decoded image data into the same space
03:09:34kugelI'm pretty sure we can optimize later
03:10:34 Join Strife1989 [0] (
03:10:56 Quit Strife89 (Disconnected by services)
03:10:57kugelwe always can
03:11:01 Nick Strife1989 is now known as Strife89 (
03:11:46Unhelpfulkugel: i also think this might be easier to implement. hooking at getc means that the hook also has to implement stuff that's in getc now - checking available buffer and advancing the pointer if not empty. also hooking read means that we don't have to worry at all about how getc changes impact ungetc.
03:13:13***Saving seen data "./dancer.seen"
03:13:31kugelwhat about the general idea of reserving space for image loaders? it strikes me a bit that png is not supported
03:15:33kugeli.e. handle them like codecs
03:15:40Unhelpfulkugel: right now all image-loader code is hardcoded. i'm not sure how many image loaders we need before a loader loader becomes a good idea?
03:16:15kugelwell, there are two plugin image viewers not supported in the core
03:16:48Unhelpfulpng and ppm? i wouldn't worry about supporting ppm in core ;)
03:17:01kugelI think it would be a better idea to invest a bit of ram usage for supporting them in the core, loading the decoders when needed
03:17:13Unhelpful*but* i can see some nice bonuses for loadable image codecs
03:18:15kugelpng is major enough to justify that idea IMO, but there's also gif potentially (or svg in an imagination)
03:18:30Unhelpfulcon: only one image load (via the core loader) can run at a time, image load always needs a disk spinup (shouldn't be a problem now, core only loads from disk)
03:19:01Unhelpfulpros: image loader can have dedicated buffers, including iram on targets where that matters - big win on coldfire, probably
03:19:15kugelI also cannot understand why we still only support bmp for .wps/.sbs, png would be a so much better choice
03:19:56kugelUnhelpful: if you load an image, the disk spins up anyway, the actual decoder is neglible in that case
03:20:21LloreanWouldn't png drastically increase the WPS load time?
03:20:32LloreanEspecially on slower targets?
03:21:11kugelI'm not sure about that. the major time for loading skins is spent in the storage drivers IIRC
03:21:32Unhelpfulkugel: i addressed that, didn't i? ;)
03:22:44kugelpng decodes pretty fast, judging by the plugin. I'm sure that if we remove the (up/down)scaling it would be even faster for skins
03:23:22LloreanI'm just disputing the "so much better choice" aspect of it.
03:23:29LloreanIt seems subjective, at best.
03:24:00kugeluntil we messure it, it's pretty objective as I see it
03:24:07LloreanNo, that's not how it works.
03:24:14LloreanUntil we measure it, it's definitely subjective.
03:24:24LloreanMeasurement is what creates an objective outcome.
03:24:29*Unhelpful can't really imagine that png would save so much for wps.
03:25:18Unhelpfulissues with png for AA: antialiased PNG can't be scaled (reasonably) without decoding the whole image into memory
03:25:27Unhelpfuls/anti//g ;)
03:25:29kugeluntil we messure it we can only assume it's worse or the same w.r.t. to loading times. but we can only *assume*. the disk space saving is a fact though
03:26:17Lloreankugel: it's also fact that *some* more time will be spent decoding the images, the question is how much.
03:26:22LloreanAnd how badly it will affect what targets.
03:26:48kugelbut we cannot determine that until we messured
03:26:49LloreanSo, as I said, currently anything is subjectivity.
03:26:59Unhelpfulurgh. actually i mean, um, interlaced. i fail at language tonight. ;)
03:27:45kugel"some" could very well mean *free* because the time is spend in the storage driver
03:27:47Lloreankugel: This may be a language issue - "objective" means based only on observable phenomenon - as long as you're assuming that the time won't be noticeable but haven't measured it yet, it's still subjective.
03:27:59 Quit StickleBricks (Quit: This now concludes our broadcast day.)
03:29:49Unhelpfulanyway, if we support PNG loading, we have to also think about how often interlaced PNGs are going to come up, how hard it's going to be to explain to users why their AA won't load and how to fix it, and how much they'll still complain even though they can now load *some* PNGs instead of *none*.
03:30:06LloreanUnhelpful: Don't we have that problem with .jpg now anyway?
03:30:12kugeldon't we already support only 1 type of jpeg?
03:30:44LloreanOr are interlaced pngs more common?
03:31:07kugeleither we say only non-interlaced "for jpeg" or non-interlaced "for all", it doesn't seem to matter
03:31:34kugelright, unless the standard png is interlaced
03:32:03Unhelpfulkugel: that's correct. the question is how *common* the unloadable files are going to be by format. conveniently progressive jpeg is useless with IE. ;)
03:33:32kugelI still think saying "we generally support png for AA but the same limitations apply as for jpeg" is better than "we don't actually support png at all"
03:34:04LloreanWe don't seem to get nearly as many calls for .png album art as we did for .jpeg.
03:34:43LloreanIs .png album art all that common in the first place?
03:34:58Strife89Might anyone know what's up with the file paths in that patch?
03:35:25Strife89I'd rather not manually point to the files again when I patch my VM sources.
03:35:26*kugel guesses git, and hence patch -p1
03:36:19Strife89Anyway, at the moment I have four DAPs to test it on.
03:36:26Strife89Two of them are redundant, though.
03:36:27kugelStrife89: make a copy of your sources, and anyway "svn revert" does a good job in most cases
03:36:47 Join TillW [0] (
03:38:23 Join Rondom [0] (
03:38:33 Join n1s [0] (~n1s@rockbox/developer/n1s)
03:43:55 Quit Unhelpful (Remote host closed the connection)
03:45:36 Join bgs100 [0] (57o9@unaffiliated/bgs100)
03:53:41 Join Unhelpful [0] (~quassel@
03:53:47 Quit Unhelpful (Changing host)
03:53:47 Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
03:57:38 Quit GeekShadow (Quit: The cake is a lie !)
03:58:46saratogaPNG is quite rare for album art I think
04:03:56 Nick fxb is now known as fxb__ (
04:09:07*Unhelpful doesn't, and won't, use embedded AA... but doesn't *terribly* mind tossing some pointers re: implementation, and perhaps dropping that patch into a branch in his git and looking at what could improve.
04:10:31Unhelpfulif we're supporting compressed formats *to save space* why would we ever want a copy of the AA inside each file? :/
04:11:23LloreanI think embedded jpeg might be added to improve compatibility, since it's apparently bloody common. But I really would see an alternative of making a tool available on a wiki that'll go through a collection with embedded AA and extract it to Rockbox usable AA.
04:12:16 Join Barahir_ [0] (
04:14:34UnhelpfulLlorean: bonus points if it strips the embedded AA from the tracks, too, ;)
04:15:31 Quit Barahir (Ping timeout: 245 seconds)
04:15:40LloreanUnhelpful: Extra ones if it compares all album art in a folder, and if they're identical, uses cover.jpeg instead of individual track jpegs.
04:17:58Unhelpfulhrm, right now the buffering code figures out whether to call the jpeg or the bmp loader itself, right? what if we wrapped individual filetype loaders in a frontend that *also* understands embedded AA (and does what is needed to get it). if the metadata code is then capable of reporting whether a file contains embedded art and how to load it, this also makes it easy on other things that want to use the current AA (i'm thinking of sliding
04:17:58Unhelpful puzzle here)
04:18:32Unhelpfulie, metadata reports that the track has AA in <trackfilename>, you ask the loader to load the image in <trackfilename>, it figures out what it is and loads it.
04:19:32 Join xavieran [0] (
04:20:23n1ssounds clever
04:20:36n1smaybe i should start using AA sometime :)
04:23:15Unhelpfuland when somebody wants to make image loaders pluggable, they have a nice *start* in that all code loading images is calling the same function.
04:24:40 Quit Rondom (Disconnected by services)
04:24:56 Join Rondom_ [0] (
04:30:19n1si think pluggable image loaders is a bit overkill since we are not likely to ever have more than bmp,jpeg and possibly png
04:32:00*Unhelpful simply doesn't want to write the image loader loader, or any link scripts for anything at all
04:37:18JdGordonreading the log abit... I also thought the codec buffer stuff was wierd
04:37:38JdGordonwhy isnt it decoded stright into a new buffer like disk AA?
04:39:51JdGordonjhMikeS: well, its not doing anything here...
04:44:54UnhelpfulJdGordon: it might have to do with the problem of embedded AA almost never being "raw" image file data
04:45:32JdGordonisnt the aa blog jpeg data?
04:46:40Unhelpfulin id3, it may or may not have been filtered first to remove anything that looks like an mp3 sync string - this needs to be reversed to get the jpeg data
04:47:47Unhelpfulvorbis comments are required by the spec to be valid utf-8, iirc, and so image data needs to be encoded, probably with base64, to be a valid comment
04:52:46JdGordonso embedded AA is a p[retty nasty hack?
04:54:17 Quit n1s (Ping timeout: 240 seconds)
04:54:41Unhelpfuldepends on the format. vorbis comments were *always* meant to be text-only. id3v2 has specific tags for images, and others for other binary data, but is designed to allow that it might be played by a player that knows mp3 but not id3
04:55:58 Quit TheSeven (Ping timeout: 246 seconds)
04:56:59 Join Rob2222 [0] (
05:00:15 Quit n17ikh (Read error: Connection reset by peer)
05:00:32 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
05:00:52 Quit Rob2223 (Ping timeout: 265 seconds)
05:08:19 Join n17ikh [0] (
05:08:38 Quit n17ikh (Remote host closed the connection)
05:10:12 Quit saratoga (Quit: Page closed)
05:10:56 Join n17ikh [0] (
05:13:14***Saving seen data "./dancer.seen"
05:19:16soapIs there a reason bmp album art is still supported? Won't save much to remove it, but wasn't it always a compromise?
05:20:47LloreanNo artifacts, and it's nearly free?
05:21:11soapWould bmp albumart be _added_ if jpg album art existed first?
05:21:23LloreanWe also don't generally remove existing features very often.
05:21:39LloreanThere'd be a lot of people with broken AA if .bmp were removed.
05:21:55soapboth of those are poor reasons IMHO.
05:22:36*Llorean doesn't like the idea of adding things "to be removed later"
05:22:39soapI think my question is a more valid way of looking at the situation than either of those.
05:23:21Lloreansoap: I often ask when a feature is going in "would we be adding this if (future feature) was done" so I understand the sentiment.
05:23:49LloreanBut I really think it's a bad idea to start stripping features that users may be using unless we *need* to reclaim the space.
05:23:55LloreanWe should just be very careful about adding things in the first place.
05:24:20LloreanDo we gain something *right now* from removing .bmp that's worth making users convert their images? If not, we should wait until we need that space back.
05:24:39soapWhoa, where's the battery life argument? ;)
05:25:37soapThe people who receive the gift of broken album art are the people who were forced to kludge their (presumably) jpg album art into bitmaps already. ;)
05:25:51LloreanYes, so we force them to repeat the conversion process again?
05:26:28soapNot to be pedantic, but it wouldn't be a repeated conversion, it would be a sync.
05:26:29LloreanAs for the battery life argument, I don't *know* by any means, but I'd imagine removal of bitmap album art would be just about the smallest binary size difference possible.
05:26:59Lloreansoap: Depends entirely on whether they have the original jpeg art around. For example, I acquired art specifically for my player, so the only existence of it in my collection is .bmp
05:27:08soapFor I think it is safe to assume few, if any, individuals deleted their jpegs when creating bmp copies.
05:27:15*Llorean never had jpegs.
05:27:29LloreanBut the point still stands - it invalidates previously supported files on their player.
05:28:25LloreanI'm also worried about removing features in general - "we can remove it later" shouldn't really be an argument for adding things.
05:28:57soapThat appears to be "all or nothing" vs evolution.
05:29:22LloreanEvolution isn't the same, really.
05:29:34LloreanA feature can improve and expand, but it shouldn't go in in the first place if it can't stand on its own.
05:29:42soapThe album art feature as evolved into the common practice.
05:30:10LloreanRemoval is an option, I just think it should, generally, be one that is considered very, very carefully.
05:30:12soapSo any existing feature gets a free pass and is considered "standing on its own" by virtue of existing?
05:30:22LloreanSomething should be removed if it has a measurable negative impact, and no effective gain.
05:30:52soapThat's a very high bar, but nonetheless different than what you said before.
05:31:02LloreanNo, it's not different from what I said before.
05:31:07LloreanI didn't deny you had reasons to remove the feature.
05:31:30LloreanI just stated my opinion that they don't outweigh the impact it would have on the users, given that I think it seems likely the cost for the feature is about as minimal as costs for features get.
05:34:52LloreanI think if we had jpeg only right now, and someone said "Hey, it's basically free to enable .bmp album art" it would basically come down to seeing how much it actually cost in binsize.
05:35:02LloreanSo we should see how much that is.
05:39:17Unhelpfulprobably a KB or two for the code for deciding which loader to call for album art, and for the code to search for both formats. and the scaler could possibly be more tightly bound to the loader if it was only going to be used by one loader... maybe it could even be inlined and the compiler could reduce/specialize it bit.
06:00:30 Nick bgs100 is now known as bgs000 (57o9@unaffiliated/bgs100)
06:04:59 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
06:05:52 Quit Strife89|PalmTX (Quit: Bed.)
06:11:17 Quit bluebrother (Disconnected by services)
06:11:21 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
06:42:17 Quit anewuser (Quit: What do you know...THE WORLD'S first NTRQ (that's for NES/FAMICOM) tracking compo. Have powerpak? Try it out! Otherwise ROM IMAGE.)
06:43:09 Join FlynDice [0] (
07:00:13 Part S_a_i_n_t_
07:00:41 Join S_a_i_n_t [0] (S_a_i_n_t@
07:06:45 Join Strife1989 [0] (
07:10:24 Quit Strife89 (Ping timeout: 260 seconds)
07:13:16***Saving seen data "./dancer.seen"
07:21:41JdGordonsoap: I dont see any point in removing .bmp AA... it is bassically free
07:21:52JdGordonI would like to change the search order though
07:38:04 Join phanboy4 [0] (
07:43:55 Quit Horschti (Ping timeout: 240 seconds)
07:48:35 Join ucchan [0] (
07:49:09 Join Horschti [0] (~Horscht2@xbmc/user/horscht)
07:50:23ucchanThere are various comments in IRC for FS #11216, and I reading now.
07:52:46ucchanIt is not possible to comment in real time always, please comments in this task.
07:55:15 Quit CGL (Quit: Saliendo)
07:56:16ucchanWhen having done for a while, I quit IRC sorry. Then I will write the answer in this task.
07:57:33 Quit ucchan (Quit: Leaving...)
08:07:54 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
08:30:21CIA-5New commit by amiconn (r25708): Charcell: Fix spurious letter appearing between board and score display.
08:39:59 Quit liar (Ping timeout: 258 seconds)
08:55:57S_a_i_n_tFS #11216 <−−Errrr...nice ;)
08:56:25*S_a_i_n_t puts this on his list of things to test in the near future.
09:01:35 Join stoffel [0] (
09:13:17***Saving seen data "./dancer.seen"
09:13:56amiconn[04:10:31] <Unhelpful> if we're supporting compressed formats *to save space* why would we ever want a copy of the AA inside each file? :/
09:14:33amiconnI once saw an .mp3 with the very same AA image attached 100 times
09:14:51amiconnThe total size of these exceeded the size of the actual mp3 data...
09:15:12S_a_i_n_tPersonally, I don;t really think saving space is a particularly good argument...but ease of use certainly is.
09:15:24S_a_i_n_tThe current system is a bit of an ass when it comes to AA
09:15:44amiconnImo separate files for AA are much more convenient than embedded aa
09:16:45S_a_i_n_t"new users" would appreciate this...the "mess-around" (or percieved mess-around) with AA is offputting at best.
09:16:54amiconnJust taking a cover.jpg and throwing it into the folder is certainly less work than firing up the tagging software and attaching it to every track
09:17:47amiconn*and* it needs less space at the same time
09:17:49S_a_i_n_tSome/most of the time, whichever media player they use has probably added the AA in embedded form anyway though.
09:18:47S_a_i_n_tI *do* agree with you...but being able to read embedded AA had been *begged* for in the ZForums
09:18:51S_a_i_n_tMANY times.
09:19:45S_a_i_n_tThe fact that the current system doesn't "just work" seems to get a few new users a little steamed.
09:19:59*amiconn wouldn't use a media player that touches his files' tags without him knowing
09:20:18S_a_i_n_tWell, that's you...many people do however.
09:20:48S_a_i_n_tWMP is a pretty good example of that.
09:21:08amiconnA media player should play media files, as flexibly as possible, but nothing else
09:21:18S_a_i_n_t*should* ;)
09:21:37amiconnWMP doesn't touch tags unless you let it
09:22:48S_a_i_n_tI've actually met a few people that *thought* RB did embedded AA because they used WMP and didn;t realise that it saves a hidden/system "Filder.jpg" in the albums directory...and they'd just been drag & dropping from the WMP library
09:24:21 Join wodz [0] (
09:24:58 Join robin0800 [0] (
09:25:31wodzWas there any disscussion about hw supported DSP functions vs. software DSP and how to organize UI for that?
09:47:57 Join n1s [0] (~n1s@rockbox/developer/n1s)
09:48:16 Join ender` [0] (
09:52:05 Quit stoffel (Read error: Operation timed out)
09:52:57 Join stoffel [0] (
09:54:02pixelmasoap: I much prefer folder album art over embedded too
09:59:00S_a_i_n_tThat's from the POV of someone with a sanely managed media library though.
09:59:18S_a_i_n_t"Mr.Everyday User" may not agree with that...
09:59:58S_a_i_n_tI prefer the current AA system, but I see the merits of reading embedded AA also
10:00:30*amiconn thinks that embedded "album" art is even named wrong
10:00:49amiconnIt should be called track art
10:00:50S_a_i_n_tIt is?
10:00:59 Quit phanboy4 (Ping timeout: 265 seconds)
10:01:32S_a_i_n_t"Album Art" very rarely differs (if at all...) between different tracks on an album though
10:01:59S_a_i_n_tSome albums have a different image for each disk, but not each track.
10:02:10S_a_i_n_t*not that I've seen anyway
10:02:11pixelmayes, that's why there is no need to put it on every track in the album
10:02:32n1swodz: i think the only suggestion is for a separate "Hardware EQ" screen that lets users adjust a hw eq as flexibly as possible while keeping the regular sw eq screen, and having settings for any other dsp present in the dac
10:02:42amiconnI'm not referring to the actual content, but to the fact that it is attached to the track
10:02:45S_a_i_n_tIf you want it to "just work" though...without any extra pissing around.
10:02:56S_a_i_n_t+then its a huge bonus
10:04:00n1swhile it might confuse the users a bit, i like the flexibility
10:04:31S_a_i_n_tTo someone switching from an OF that manages the AA itself, switching to Rockbox and having to create an image foe every album in the library (although this can be automated) is a daunting task.
10:04:50S_a_i_n_tAnd if it could be avoided, I'm sure a LOT of people would appreciate it.
10:05:17n1sI'd also like to insert into the earlier discussion that i personally don't think that that a feature is already in is a good reason to keep it, (not saying we should drop bmp aa)
10:08:55 Quit Lynx_ (Ping timeout: 276 seconds)
10:09:04wodzn1s: but there is more DSP settings then simple EQ in some of the codecs
10:09:53n1si think just introducing settings for them is fine, we have that for a lot of the dsp stuff in the mas codecs
10:10:27wodzn1s: Yes but there is not distinguish between hw DSP and software DSP now
10:10:35 Join Lynx_ [0] (~bayer@
10:11:46n1strue, but i don't think that matters a lot, except for the eq, since we'd have two eq's then
10:12:26wodzit matters from the clearity point of view
10:12:47n1swe have bass/treble in hw on some targets and sw on others
10:13:19wodzin my opinion we should provide all features of the particular codec in HW menu and standard rockbox DSP in other
10:13:59 Quit bluebroth3r (Read error: Connection reset by peer)
10:14:35wodzcurrent mix is missleading
10:14:48 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
10:15:09kugelit is?
10:15:43 Join Lynx__ [0] (~bayer@
10:15:59wodzkugel: setting EQ do You know if this is hardware supported or not?
10:16:03kugelI see no point of offering a feature in SW if the hardware can do it on it's own
10:16:06 Join bimbel [0] (~Miranda@unaffiliated/bmbl)
10:16:07 Quit bluebrother (Read error: Connection reset by peer)
10:16:14 Quit Lynx_ (Ping timeout: 240 seconds)
10:16:25 Quit bmbl (Disconnected by services)
10:16:29 Nick bimbel is now known as bmbl (~Miranda@unaffiliated/bmbl)
10:16:35kugelin usage, it doesn't matter where it comes from
10:16:57wodzkugel: it matter - software DSP reduce runtime
10:17:15kugelis it important that you can see whether the hardware supports it in the settings menus?
10:18:28wodzI preffer to see if particular setting will have impact on runtime
10:18:59wodzs/impact/negative impact/
10:19:15 Join merbanan [0] (
10:19:50 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
10:20:02 Quit Lynx__ (Ping timeout: 246 seconds)
10:20:45 Join Lynx_ [0] (~bayer@
10:21:20kugelthe dsp functions supported in HW are very basic (until now?) and very cheap to do in SW, I don't think the impact on battery life is messurable or even existent
10:21:45n1swodz: if you'd like to change this i suggest discussing it on the malinglist, i personally disagree though
10:21:45 Quit bluebrother (Read error: Connection reset by peer)
10:22:28n1skugel: the beast's dac can do a 5 band eq that is a bit less flexible than the sw one in hw
10:22:55n1siirc wodz's target has some 3d dsp thing
10:22:57kugelwe don't use it, do we?
10:24:11wodz3d thing in wm8750 is very similart to stereo width setting in rb but can only increase channel separation
10:24:19n1snot yet, but we want to and i think having the option to use either one (or both) is best
10:25:24wodzn1s: thats my point also to provide option to use both hw and sw dsp
10:25:40wodzbasically the flow sw dsp -> hw dsp
10:26:15wodzpreprocess signal in sw if requested and than process in hw if requested
10:27:35kugelwodz: whatÄs the difference between the rb setting and the hw 3d thing exactly?
10:27:51kugelI thought the setting also only increases the separation
10:27:58 Quit Lynx_ (Ping timeout: 245 seconds)
10:28:11n1swodz: but having f.e. bass/treble both in sw and hw makes little sense to me since they do the same thing and the sw is only there if the hw can't do it
10:28:40 Quit merbanan (Ping timeout: 252 seconds)
10:28:43pixelmakugel: you can also reduce stereo width (seperation)
10:28:43hobbswodz: you mean a single interface, but use the best one for a given effect behind the scenes?
10:29:32 Join pamaury [0] (~c2c7a50a@rockbox/developer/pamaury)
10:29:35kugelhobbs: no, that's what's being currently
10:29:40wodzhobbs: This is arbitrary decision which is the best isn't it?
10:29:56kugelnot really
10:31:08n1shey, hobbs, you got a real name so i can commit your png fix?
10:31:26hobbsn1s: sure. Andrew Rodland <>
10:32:17wodzI have no obejctions to use hw instead sw if both provide the same functionality but how to treate situation where hw support is not exactly the same as sw?
10:32:44pixelmahobbs: this channel is logged, I wouldn't put the plaintext email address here
10:32:58S_a_i_n_tbit late for that...
10:33:01hobbsyou're a bit late ;)
10:33:36hobbsoh well, I've got some high-quality spam filtering ;)
10:34:00hobbsbut I appreciate the thought
10:34:01pixelmaif you are lucky someone who has access to it could edit it (maybe Bagder)
10:35:32CIA-5New commit by nls (r25709): Fix for some bogus code in png by Andrew Rodland, fixing a warning when compiling with gcc 4.5
10:35:52pixelmaas a note to the png discussion - the png viewer is still colour screens only. And even if there was a greyscale "port" it could use the greyscale library in a plugin, not in the core (bmp in the core vs. the plugin does the same btw.). And I also can't see the benefit of png for e.g WPS graphics, the file size difference is probably not big for most of them with the exception of the backdrop maybe and bmp decoding is super simple and png is not
10:36:21pixelmain comparison
10:36:49wodzfor WPS graphics much more interesting would be svg support
10:37:00n1spixelma: i agree about the wps graphics
10:37:05pixelmaI don't think so
10:37:36hobbswodz: making SVGs that render nicely at really low res is way harder than it sounds
10:37:42hobbsit's easier just to remake the pixel art ;)
10:38:15 Nick fxb__ is now known as fxb (
10:38:46pamaurydoes SVG really makes sense for such low resolution ? There would probably be no gain for complicated backdrops
10:39:56wodzProbbably - I'am not graphic expert. My point is only that reusing scalable graphics will be much easier between targets with different screen sizes.
10:41:38pamauryyes that's true. My point is that a the SVG file will also be big relatively the bmp ones because the resolution is quite low. And so does it makes sense to keep in SVG rather than generating a bmp from it for each target ? I don't know the answer either :)
10:44:08wodzDoes the size of theme file really matter much? We are speaking about size in range of tenth to hundreds of kB max
10:44:22pixelmaan image that has to be scaled very much for the slow resolutions and colour depths won't look good anyways, I'm pretty sure
10:44:33pixelmaerr... low resolutions
10:45:10hobbsyeah, that's what I was saying
10:45:20pixelmawhat hobbs said (should have read compl...)
10:45:42*kugel wonders why r25706 is for SWCODEC only
10:46:30 Quit scorche (Disconnected by services)
10:46:36wodzso maybe ask themers to include vector version of theme as a separate package?
10:46:41 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
10:47:16pamauryIf they have one !
10:48:07wodzSuch ask may promote habbit to have one
10:48:38S_a_i_n_tThat would over-complicate the thems IMO
10:49:46wodzS_a_i_n_t, I am not speaking to include this in 'uploadable to the device' theme package. I would see this as a sort of 'source code' for theme rather
10:50:14pixelmaI think some do
10:50:26S_a_i_n_tHmmm, I see your point. But is there any *need* to switch from .bmp?
10:50:35S_a_i_n_t.bmp is fast...and easy.
10:51:55 Join JdGordon_ [0] (~jd@rockbox/developer/JdGordon)
10:51:58wodzS_a_i_n_t: As I said on device bmp is fine but providing vector version as a sort of 'source code' will make porting theme to other screen resolution easier
10:52:02hobbsmy "original source" isn't vector anyway, it's a GIMP file :)
10:52:21wodzok I giveup
10:53:12 Join M3DLG [0] (~M3DLG@
10:54:08kugelS_a_i_n_t: nobody suggested removing .bmpt
10:59:00 Nick scorche` is now known as scorche (~scorche@rockbox/administrator/scorche)
11:01:20 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
11:08:43 Join Highlander [0] (
11:13:18***Saving seen data "./dancer.seen"
11:17:55 Join DerPapst [0] (
11:20:48 Join kifo [0] (
11:20:58 Quit mc2739 (Ping timeout: 258 seconds)
11:23:49 Quit JdGordon_ (Quit: Bye)
11:25:10 Join Luca_s [0] (
11:25:48 Quit kifo (Quit: CGI:IRC (Ping timeout))
11:29:17 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
11:32:35 Quit Highlander (Quit: Quitte)
11:36:04 Join mikroflops_ [0] (
11:37:23 Join Highlander [0] (
11:39:39 Quit mikroflops (Ping timeout: 260 seconds)
11:47:23 Nick fxb is now known as fxb__ (
11:52:00 Join dfkt [0] (dfkt@unaffiliated/dfkt)
11:52:14 Quit Strife1989 (Quit: Thunderstorm.)
12:04:26 Join ender [0] (
12:06:55 Quit ender` (Ping timeout: 264 seconds)
12:07:27 Join flydutch [0] (
12:09:18 Join ender` [0] (
12:10:41 Join jnss_ [0] (janes@gateway/shell/
12:11:57 Quit ender (Ping timeout: 258 seconds)
12:12:11 Quit jnss (Ping timeout: 240 seconds)
12:19:25 Quit n1s (Ping timeout: 246 seconds)
12:23:49 Join liar [0] (
12:25:09 Quit stoffel (Remote host closed the connection)
12:25:35 Join {phoenix} [0] (
12:31:53 Quit kugel (Ping timeout: 240 seconds)
12:32:36 Join n1s [0] (~n1s@rockbox/developer/n1s)
12:33:10 Join kugel [0] (~kugel@rockbox/developer/kugel)
12:45:17 Join efyx [0] (
12:46:35 Join stoffel [0] (
12:48:30 Quit S_a_i_n_t ()
12:49:11 Join S_a_i_n_t [0] (S_a_i_n_t@
13:13:23***Saving seen data "./dancer.seen"
13:14:27 Quit stoffel (Ping timeout: 276 seconds)
13:14:54JdGordonpixelma: I havnt heard otherwise so I'm assuing you arnt having issues with %pv inside conditionals anym,ore?
13:20:07 Quit Luca_s (Quit: CGI:IRC (EOF))
13:21:25kugeln1s: how's gcc 4.5?
13:24:46 Quit S_a_i_n_t (Ping timeout: 240 seconds)
13:26:00 Join S_a_i_n_t [0] (S_a_i_n_t@
13:27:38JdGordonwhy the hell isnt my battery level bar working :'(
13:31:51 Join Jaykay [0] (
13:32:19 Quit chaos (Ping timeout: 260 seconds)
13:32:39linuxstbwodz: I don't know how important it is, but docs/COMMITTERS was in alphabetical order before you added yourself.
13:32:53CIA-5New commit by jdgordon (r25710): Make battery level (%bl) work with the bmp bars drawer like %pb and %pv
13:40:25 Join Xerion_ [0] (~xerion@
13:41:57 Quit Xerion (Ping timeout: 264 seconds)
13:41:58 Nick Xerion_ is now known as Xerion (~xerion@
13:46:24wodzlinuxstb: ups I didn't noticed this. I'll fix this later
13:48:18 Quit M3DLG (Ping timeout: 264 seconds)
13:49:07 Join M3DLG [0] (~M3DLG@
13:56:58CIA-5New commit by wodz (r25711): Fix alphabetical order of COMMITERS file
13:57:18 Join Xerion_ [0] (~xerion@
13:59:18 Quit Xerion (Ping timeout: 276 seconds)
13:59:18 Nick Xerion_ is now known as Xerion (~xerion@
14:01:12 Quit {phoenix} (Remote host closed the connection)
14:01:27CIA-5New commit by jdgordon (r25712): Make the buffering screen use the multiscreen API (FS #11143)
14:01:34 Quit liar (Ping timeout: 258 seconds)
14:02:11 Join Xerion_ [0] (~xerion@
14:02:50kugelis it known/intended that bass/treble settings don't actually do anything in the sim?
14:04:11 Quit Xerion (Ping timeout: 276 seconds)
14:04:11 Nick Xerion_ is now known as Xerion (~xerion@
14:05:02kugelI'm currently doing an alsa-lib audio driver for the sim and have to notice that none of the dsp stuff actually works
14:05:59S_a_i_n_tre: Bass/Treble...I had noticed it, but whether it's intended or not is an entirely different story I guess.
14:07:14kugelI can't imagine it's intended, I think that simply nobody has done the needed work yet
14:07:57*JdGordon wonders why that yellow didnt hit every target without a remote lcd
14:11:36kugelthe others probably have LCD_HEIGHT > 80, so the || part isn't even evaluated
14:12:47JdGordonthat makes sense
14:14:15JdGordonalso would help if I got the define right :p
14:14:30CIA-5New commit by jdgordon (r25713): fix yellow
14:16:06 Join w1ll14m [0] (
14:19:06 Join Xerion_ [0] (~xerion@
14:20:02 Quit M3DLG (Ping timeout: 245 seconds)
14:20:51 Quit Xerion (Ping timeout: 260 seconds)
14:20:51 Nick Xerion_ is now known as Xerion (~xerion@
14:23:31 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
14:28:10 Quit Highlander (Quit: Quitte)
14:30:26 Nick bgs000 is now known as bgs100 (57o9@unaffiliated/bgs100)
14:37:18 Quit pamaury (Ping timeout: 252 seconds)
14:41:37kugelhrm, with minimal changes bass&treble works for the sim
14:45:02 Join Xerion_ [0] (~xerion@
14:45:18 Join M3DLG [0] (~M3DLG@
14:45:58 Quit Xerion (Read error: Operation timed out)
14:45:58 Nick Xerion_ is now known as Xerion (~xerion@
14:51:18 Quit M3DLG (Ping timeout: 276 seconds)
14:51:55 Join M3DLG [0] (~M3DLG@
14:55:18 Quit n17ikh (Ping timeout: 265 seconds)
14:56:06 Quit linuxstb (Ping timeout: 264 seconds)
14:56:10 Quit M3DLG (Ping timeout: 246 seconds)
14:59:40pixelmaJdGordon: yes, your last fix made my WPS work again. Is %bl| in a conditional also safe with the latest addition?
15:00:31 Quit mt_ (Ping timeout: 240 seconds)
15:02:34 Quit antil33t (Read error: Connection reset by peer)
15:02:40 Join antil33t [0] (
15:04:29 Quit jordan` (Ping timeout: 265 seconds)
15:10:41 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
15:13:26***Saving seen data "./dancer.seen"
15:23:46 Quit kugel (Remote host closed the connection)
15:26:43 Join panni_ [0] (
15:34:47n1skugel: a test compile with gcc4.5 for coldfire went fine, with the only error being the nsf codec not fitting in iram, and a handful of new warnings, haven't been able to try the buid yet thoug so can't say anyting about performance or if it even works :)
15:35:29n1salso i would say that at least dsp stuff that is usually done in software should work in the sim
15:35:56amiconnTreble/bass is hardware on many targets though
15:36:32 Quit FlynDice (Remote host closed the connection)
15:37:05n1syeah, but imo it makes sense to have them in sw in the sim so they consistently work
15:39:51 Quit ps-auxw (Ping timeout: 248 seconds)
15:40:08 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
15:43:08 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
15:51:36 Join liar [0] (
15:53:33 Quit mt (Read error: Connection reset by peer)
15:58:50 Join stoffel [0] (
15:59:25 Join lpereira [0] (
16:02:16 Join mt [0] (~mtee@rockbox/developer/mt)
16:04:41 Quit n1s (Ping timeout: 260 seconds)
16:06:01 Join Strife1989 [0] (
16:06:16 Join phanboy4 [0] (
16:09:11 Quit mt (Ping timeout: 240 seconds)
16:09:59soapiTunes and e-music both ship files with embedded album art. For people who don't do as I do and only deal with complete albums embedded appears to not only be common, but to make some sense. Regardless, embedded is /common/,
16:12:52 Join adnyxo [0] (
16:15:35 Join Antibuddha [0] (
16:15:57Antibuddhaim using h120 rockboxed, what volume do i set it at for the line-out to be normal volume? i think i remember reading something like -14dB
16:18:23 Join n1s [0] (~n1s@rockbox/developer/n1s)
16:18:43linuxstbAntibuddha: 0dB
16:18:56Antibuddhaare you sure?
16:19:08Antibuddhawhy is it that the iriver h120 has variable line-out?
16:19:57 Join jordan` [0] (
16:20:28Antibuddhai heard people say they damaged their h120 by going too high volume on lineout
16:21:15Antibuddhaim about to get a portable headphone amplifier and want to know the best volume to set the lineout at
16:24:58 Quit lpereira (Quit: Leaving.)
16:25:32 Quit DerPapst (Quit: Leaving.)
16:32:54soap0 = line-out on my iPod's headphone jack and my H340's headphone AND line-out jack.
16:33:29soapShould a line-out be fixed? Yes, It is on much hardware, but not on my H3x0, and therefore I assume not on your H1x0.
16:34:35linuxstbIt's a "feature" of the h1x0 that the volume changes the line-out volume - Rockbox can't prevent it.
16:35:53 Quit bertrik (Read error: Connection timed out)
16:36:25Antibuddhawould i be avoiding distortion by choosing a particular volume for lineout?
16:36:50soaphigher and you will potentially clip, lower and you decrease the SNR.
16:37:03Antibuddhai heard people say that they damaged their h120 or reached distortion by going very low volume on lineout
16:37:30Antibuddhais it because they tried driving headphones?
16:38:44soap /though/ FWIW my cassette tape adapter, not Rockbox, distorts @ anything louder than -5. But only that adapter. All my line-in's do fine @ zero.
16:39:25Antibuddhaso if i set the lineout of h120 to -20dB, it will have much less SNR?
16:39:48soapGoing very low volume would expose the native noise of the system due to low signal levels, but anyone who said that causes damage needs to really explain how and why.
16:40:09soap20db less. You're lowering the signal while noise remains constant.
16:40:20Antibuddhasoap: i believe they damaged only by using lineout to drive headphones
16:40:35Antibuddhanot by using it as normal lineout
16:41:20soapWouldn't lower volumes = less voltage and /decreased/ likelyhood of problems?
16:42:48Antibuddhawhat i mean is people damaged it at around -7dB if i recall
16:43:06Antibuddhausing lineout as opposed to headphoneout
16:44:45 Join pamaury [0] (~c2c7a50a@rockbox/developer/pamaury)
16:46:31 Quit Antibuddha (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
16:47:25 Quit phanboy4 (Ping timeout: 265 seconds)
16:53:06soapI don't know enough to say said people are wacked in the head - but it is not uncommon practice for H3x0 users who have broken their headphone-out to use their line-out as a headphone out.
16:53:46 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
17:04:30 Quit wodz (Quit: Leaving)
17:04:32 Quit liar (Ping timeout: 276 seconds)
17:11:07 Quit bertrik (Read error: Connection timed out)
17:11:40 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
17:13:19amiconnn1s: Imo it should depend on the target
17:13:29***Saving seen data "./dancer.seen"
17:20:17n1sso on targets with hw bass/treble they do nothing in the sim?
17:23:18 Nick fxb__ is now known as fxb (
17:24:11n1ssoap: i did that, untill the line-out broke :/
17:24:21 Nick fxb is now known as fxb__ (
17:43:47 Quit stoffel (Ping timeout: 246 seconds)
17:46:24 Quit bertrik (Read error: Connection timed out)
18:00:10 Quit adnyxo (Ping timeout: 252 seconds)
18:00:46 Join chaos [0] (~chaos@gentoo/user/ch4os)
18:01:09 Join kkurbjun [0] (
18:01:18 Quit kkurbjun (Changing host)
18:01:18 Join kkurbjun [0] (~kkurbjun@rockbox/developer/kkurbjun)
18:01:23 Join adnyxo [0] (
18:02:51 Join anewuser [0] (anewuser@unaffiliated/anewuser)
18:04:36 Join liar [0] (
18:06:55 Quit Topy44 (Ping timeout: 276 seconds)
18:10:22CIA-5New commit by kkurbjun (r25714): M:Robe 500 USB: Go back to OF timings for upcoming clock changes
18:11:51CIA-5New commit by kkurbjun (r25715): M:Robe 500 Power: Update battery Curve and simplify calculations.
18:14:03CIA-5New commit by kkurbjun (r25716): DM320 UART: Enable clock in UART module since it will not be done in system initialization.
18:14:09 Join stoffel [0] (
18:15:01 Join Topy44 [0] (
18:19:09 Quit tchan (Quit: WeeChat
18:19:55 Quit Topy44 (Ping timeout: 276 seconds)
18:20:50 Nick jnss_ is now known as jnss (janes@gateway/shell/
18:23:16 Quit linuxstb (Ping timeout: 260 seconds)
18:26:51 Join Topy44 [0] (
18:28:28 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
18:29:53 Join bertrik [0] (
18:30:19 Quit bertrik (Changing host)
18:30:19 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:31:37 Quit Topy44 (Ping timeout: 276 seconds)
18:37:58 Join n17ikh [0] (
18:38:21 Quit n17ikh (Client Quit)
18:38:36 Quit n1s (Quit: Lämnar)
18:38:52 Join n17ikh [0] (
18:39:22 Join Topy44 [0] (
18:40:59 Join Highlander [0] (
18:44:37 Quit Topy44 (Ping timeout: 276 seconds)
18:46:15 Quit adnyxo (Ping timeout: 276 seconds)
18:48:37 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
18:51:04 Join mirak [0] (
18:52:59 Join adnyxo [0] (
18:59:08ollebei have an e10 simulator now that seems to work reasonably well
18:59:30 Join merbanan [0] (
19:00:04ollebethink i'll upload the patch for the build target today :)
19:01:13ollebeplayback sounds quite good as well :)
19:01:59 Join phanboy4 [0] (
19:05:12ollebeshould the BMP pic of the player used for the simulator be actual size?
19:05:20ollebeit isn't.
19:05:52LloreanI don't see how you can practically have an "actual size" since the displayed size depends on the monitor pixel size.
19:06:13 Join Topy44 [0] (
19:06:31ollebethat's true. but should the area used for the LCD have the same resolution as the actual LCD? i think so.
19:06:59ollebeas of now the simulator only uses part of the "LCD" on the screen
19:07:44LloreanYes, it should 'fit' around the virtual LCD.
19:08:15ollebeis there a way to include a binary file, such as a BMP, in a patch?
19:09:37 Join japper [0] (
19:10:53 Quit japper (Client Quit)
19:11:55 Quit Topy44 (Ping timeout: 276 seconds)
19:11:55 Join CGL [0] (~CGL@
19:13:31***Saving seen data "./dancer.seen"
19:15:21 Quit bertrik (Read error: Connection timed out)
19:16:13 Join bertrik [0] (
19:35:55 Join Topy44 [0] (
19:36:17ollebepatch is uploaded, now it's just a matter of patience i guess
19:41:10 Quit Topy44 (Ping timeout: 276 seconds)
19:42:43 Join RoC_MasterMind [0] (
19:44:01linuxstbollebe: First question - where did you get the simulator background image from? Do you own the copyright?
19:44:38ollebelinuxstb: sorry about that. no, i don't own the copyright.
19:44:51ollebei'll add a comment about excluding the image.
19:45:16 Quit mirak (Quit: Ex-Chat)
19:45:43 Quit Unhelpful (Remote host closed the connection)
19:47:54 Join saratoga [0] (~463f90ed@gateway/web/freenode/x-dfbqriqwgwmobxas)
19:48:24saratogawhats the storage like on the e10? another tcc NAND translation layer?
19:48:42linuxstbollebe: In config/irivere20.h you should choose an unused MODEL_NUMBER value (look in the other config files in that directory to see what is used). Also
19:49:05linuxstbIf I'm understanding the patch, it's a standard ATA interface.
19:49:35ollebesaratoga: IIRC it is a seagate 6 GB HDD with PATA to a PATA interface in the SoC
19:49:47saratogaoh so ATA?
19:50:13saratogashould make things a lot easier
19:50:16ollebeok linuxstb, i'll fix the model number.
19:51:03ollebeyeah it simplifies. i think ATA support is pretty much finished already. LCD driver, power handling, etc is missing.
19:51:50linuxstbAh, a 128x128x16 LCD - same as the H10. Maybe it's the same LCD controller?
19:51:54saratogahow do load code onto the e10?
19:52:19*linuxstb was about to ask that ;)
19:53:51*amiconn tries to single-step the elio rom in skyeye
19:53:57ollebei don't know that, but someone will figure something out :)
19:54:15amiconnlinuxstb: Btw, I *guess* the lcd controller in the elio is a TL1769
19:55:02linuxstbamiconn: I never managed to find anything that I could identify as the LCD driver... I assume you haven't made it work yet though?
19:56:47 Quit RoC_MasterMind (Quit: Leaving)
19:57:05 Quit antil33t (Read error: Connection reset by peer)
19:57:13 Join antil33t [0] (
19:57:51saratogaollebe: maybe figuring out how to run code is more important then the Sim? usually people do that first
19:58:46ollebesaratoga: well, the simulator wasn't such a challenge so I did that first
19:59:38linuxstbollebe: So you can't run anything on your E10?
19:59:54 Join Xerion_ [0] (~xerion@
20:00:05ollebelinuxstb: nope.
20:00:43linuxstbHmm, then I'm not sure if there is a point in committing anything. Finding out how to run your own code is the first thing to be done - if you can't do that, the rest is pointless.
20:01:21ollebethere's a possibility that we could use irde.c to encrypt the image and do a FW upgrade, but E10's encryption are not consistent with the other ifp-like images.
20:01:43 Quit Xerion (Ping timeout: 252 seconds)
20:01:43 Nick Xerion_ is now known as Xerion (~xerion@
20:02:07linuxstbHave you managed to decrypt the original firmware to disassemble it?
20:03:21ollebelinuxstb: I think the OF is decrypted by ifp_decode or by irde.c, yes. but since what's know about the checksums don't seem to apply to E10, I'm not sure.
20:03:31ollebeI have written a bit about that on the wiki
20:04:25ollebeit is being decrypted, but what was thought to be checksum fields are not checksums. applies to a large portion of the OF.
20:04:42linuxstbSo the firmware upgrade process is similar to the older irivers? i.e. copy a .hex file to the device and upgrade via the original firmware?
20:05:55 Join mozetti [0] (~60ffd62b@gateway/web/freenode/x-pxbyheadcggapzvs)
20:05:59ollebethere is also a USB emergency mode, that might be a DFU mode, but it's not clear how to enter.
20:06:02 Quit liar (Quit: Verlassend)
20:06:19ollebeIIRC i entered that mode when i messed up the formatting of the HDD.
20:07:25linuxstbWhere is the info on the wiki? I can just find IriverE10Info which doesn't mention the firmware format.
20:07:34 Nick Strife1989 is now known as Strife89 (
20:07:48mozettiAfternoon - my HDD on my iRiver H340 died. I'm going to do the CF-Mod but my main use will be recording concerts (that permit taping). Should I be overly concerned about the CF card write speeds? Or will they all generally be fast enough for audio recording? Thanks
20:08:08ollebethere's also some kind of hardware error that happened to some users. might be the connection to the HDD that fails, and it enters emergency mode.
20:08:09kkurbjunollebe: I was just looking at the board scans and they clearly silkscreened the JTAG pins so you should be able to get hardware access to the player through JTAG to test new images and recover.
20:08:38ollebei'll add that information to the wiki linuxstb
20:08:41kkurbjun(TMS, TDO, TCK, TDI)
20:08:53 Join liar [0] (
20:08:57 Join kugel [0] (~kugel@rockbox/developer/kugel)
20:09:00ollebethanks kkurbjun
20:09:53kkurbjun(TRST) - looks like there's also some UART pins for debugging too (TX0, RX0)
20:09:57kkurbjunno problem
20:10:39kkurbjunthey also have the ISP pins pulled out for that atmel chip if you want to re-program that too on the other side of the board :-D
20:11:08ollebeyes i know about the Atmega :)
20:11:50linuxstbIs there any NOR? I can't see it in the list of chips, or is it embedded in the SoC?
20:11:59ollebeit might be for the remote control and button handling, which works even when the system is in off mode.
20:12:07 Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
20:12:29ollebelinuxstb: IIRC there's a 50 kB flash in the SoC
20:12:40kugelamiconn: what do you mean by it should be target dependant?
20:12:59linuxstbollebe: IIRC, that's the boot ROM. I'm wondering where the ".hex" file is written to.
20:13:06kugelif a target does bass/treble, no matter of in hw or sw, it should be simulated, IMO
20:13:24 Quit bertrik (Read error: Connection timed out)
20:14:11 Join bertrik [0] (
20:14:33linuxstbollebe: Have you tried removing your hard disk and reading it in a PC?
20:14:45ollebelinuxstb: it might be to some hidden area on the HDD. excellent!
20:15:36linuxstbollebe: How do you know that the "checksum fields" are not checksums?
20:16:07ollebelinuxstb: i explained that on a wiki page, one second
20:17:08ollebe However I ran into problems when i tried running irde on E10's firmware. It seems on the positions in the blocks where there should be checksums (offset x*51 + 50, and offset 255), large parts of the E10 firmware has no checksums but something else entirely. Even if the contents of the blocks varies, the "checksums" stays the same, so they're not checksums.
20:17:11 Quit Unhelpful (Remote host closed the connection)
20:17:32 Join Topy44 [0] (
20:18:31 Join Luca_s [0] (
20:19:51saratogahow big is the firmware file? on a color target large sections may be graphics which don't need to be check summed
20:20:06saratogai think the older ifp targets weren't color and so would have had few graphics
20:22:19ollebegood point.
20:22:46 Quit Topy44 (Ping timeout: 276 seconds)
20:22:59ollebee10 uses SWF flash for the user interface. that might have something to do with it as well.
20:23:15ollebeit's 6 MB.
20:23:26ollebemuch larger than IFP's.
20:25:48 Join stripwax [0] (
20:25:58ollebei can confirm that the SoC has only one small RAM and one small ROM
20:26:51 Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
20:27:34 Join funman [0] (~fun@rockbox/developer/funman)
20:31:25funmanranma: nice findings, do you understand what the as3514 bits you modified mean ?
20:32:24amiconnkugel: Targets with software treble/bass should have it working in the sim, targets wih hardware treble/bass shouldn't (unless sdl allows setting the soundcard's tone filters)
20:32:41amiconnlinuxstb: I found code initing structures with lcd addresses
20:33:11amiconn...but no code using literal addresses
20:33:31amiconnThat's why I'm trying the emulator approach instead
20:34:09kugelthat doesn't make sense to me
20:34:12amiconnThe elio rom uses a lot of thumb code
20:34:17kugelit's a simulator afterall
20:34:29kugelnot an emulator
20:34:56pixelmathen MAS playback should be simulated as well ;)
20:35:29kugelwe also simulate backlight without really dimming the brightness of the display
20:36:44 Join Topy44 [0] (
20:38:37funmankugel: it makes sense but the SW controls wouldn't run on the target, so they shouldn't run in the sim either
20:39:31kugelsurely simulating in another way is better than not simulating at all?
20:40:01 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
20:41:30LloreanIsn't it a "UI Simulator"? Isn't the key presenting an accurate visual representation of the target, not necessarily anything else?
20:41:36 Quit Topy44 (Ping timeout: 276 seconds)
20:42:06linuxstbWell, the more it simulates, the more useful it is for developers.
20:42:06funmanLlorean: no it can be used to debug code as well
20:42:32funmanI wonder if running this code when simulating a target where it wouldn't run, would harm debugging audio code
20:43:26pixelmait's probably a question how representative it is (and if it adds much code complication) if code goes a different code path for a specific target's sim and the real target
20:44:36 Quit droidcore (Ping timeout: 252 seconds)
20:45:07linuxstbollebe: Are you able to remove the disk and access it directly?
20:45:30ollebelinuxstb: probably.
20:46:15 Quit ender` (Ping timeout: 276 seconds)
20:46:15 Quit mapi (Ping timeout: 252 seconds)
20:46:21ollebei'm looking at the HDD through the usual interface right now
20:46:55linuxstb"usual" interface? You mean with it still in the E10?
20:46:59ollebei'm expecting the OF to "censor" some hidden area at the beginning or the end of the disk.
20:47:01saratogaLlorean: it also runs the whole SWCODEC and buffering systems too, so I use it a lot to debug and develop for those parts of rockbox too
20:47:04 Quit Unhelpful (Remote host closed the connection)
20:47:13 Join DerPapst [0] (
20:47:17pixelmakugel: but something different. I tested the simplify PLA patch on my Ondio and it works fairly well on my Ondio, with the one of metronome. It's still better than SVN but I *believe* an earlier patch worked better. Unfortunately it's long ago that I tested earlier version and I only remember being impressed before, I'm not a 100% sure though if it also was the Ondio etc.
20:47:33 Join mischasworld [0] (
20:47:37linuxstbYes, the OF will probably not expose the first X MB of the disk via USB - so you'll need to physically remove it to see what's there.
20:48:20pixelmakugel: one exception I mean
20:48:21ollebehowever i have found a 1.5 MB area that I'll look at.
20:48:34kugelpixelma: can you tell what worked and now doesn't anymore?
20:49:54pixelmano, hence my "I'm not sure". The problem is tapping and start/stop "playing" all on the same button
20:50:38pixelmamaybe I can find out by looking at the different patches
20:50:58kugelI guess you can still try out the older patches by applying it to an older svn rev
20:51:59pixelmayeah - but updating builds on the Ondio is no fun :\ maybe the sim can help
20:52:27 Quit flydutch (Quit: /* empty */)
20:52:36kugelpixelma: I can imagine what changed
20:53:00kugeltry changing the "#define METRONOME_TAP PLA_SELECT" to "#define METRONOME_TAP PLA_SELECT_REL"
20:54:19Luca_sfunman: FuzeV2 - I did the battery bench without CPU frequency scaling. lame_128.mp3 played in loop for 13hrs 30minutes. The difference is about 4hrs, so it's actually noticeable.
20:54:30 Nick Luca_s is now known as Luca_S (
20:55:27funmanyep I saw that, it was to be expected
20:56:02 Join MethoS- [0] (~clemens@
20:56:11pixelmakugel: ok, I'll try. I have to run off for a while in a bit though
20:56:20 Join Topy44 [0] (
20:56:57 Join ender` [0] (
20:57:52 Join Lear [0] (chatzilla@rockbox/developer/lear)
21:01:06 Quit Topy44 (Ping timeout: 276 seconds)
21:02:34 Join mapi [0] (
21:02:37 Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
21:03:26 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
21:04:33 Join mirak [0] (
21:09:41 Join archivator [0] (
21:10:29 Quit GeekShadow (Read error: Connection reset by peer)
21:10:45 Quit Unhelpful (Remote host closed the connection)
21:11:06mozettiQuestion: When doing the CFMod (H340), should I be concerned about CF Card write speeds if I plan to use the H340 to record live concerts? Or should any decent card be fine for recording audio?
21:12:21 Join Topy44 [0] (
21:13:00gevaertsmozetti: I suspect the forums might be a better place for that question. My impression is that there are more people there who do recording
21:13:29 Quit stripwax (Quit:
21:13:33***Saving seen data "./dancer.seen"
21:13:42archivatorIs there a non-callback recording API? I'd like to add recording to FFT but would like to keep the current 2-thread system.
21:15:34mozettigevaerts: thank you. I knew the devs spent time here and thought it might be more of a hardware capability/technical kind of question. I'll try the forums.
21:16:57gevaertsmozetti: only one or two of the devs really care much about recording, and they don't seem to be online right now
21:17:21 Quit Topy44 (Ping timeout: 276 seconds)
21:17:26 Join stripwax [0] (
21:17:52 Join dfkt [0] (dfkt@unaffiliated/dfkt)
21:18:07linuxstbmozetti: Even uncompressed PCM is only about 176KB/s, so I wouldn't worry too much. But yes, it wouldn't hurt to ask in the forums.
21:18:32mozettithanks everyone
21:21:55 Join Unhelpful [0] (~quassel@
21:21:55 Quit Unhelpful (Changing host)
21:21:55 Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
21:24:22 Quit mischasworld (Remote host closed the connection)
21:28:10 Quit funman (Quit: free(random());)
21:28:21 Quit Lear (Quit: ChatZilla 0.9.86 [Firefox 3.6.4/20100413172113])
21:31:04Lloreanmozetti: Just a question - why wav instead of lossless wavpack? If you're worried about write speeds for some reason (as linuxstb mentioned, it shouldn't be a big concern) that would give you even more elbow room.
21:31:09 Quit esperegu (Read error: Connection reset by peer)
21:32:17mozettiLlorean: familiarity - never used wavpack before. Just .wav and the encode it to FLAC/MP3/OGG/etc on my PC
21:32:25 Join esperegu [0] (~quassel@
21:32:38mozetti*then encode
21:36:38 Join Topy44 [0] (
21:36:40 Quit saratoga (Quit: Page closed)
21:36:54 Quit Rob2222 (Remote host closed the connection)
21:37:09 Join Rob2222 [0] (
21:37:45 Quit pixelma (Disconnected by services)
21:37:46 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
21:38:08 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
21:38:20 Quit amiconn (Disconnected by services)
21:38:22 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
21:38:31 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
21:42:03 Quit Topy44 (Ping timeout: 276 seconds)
21:44:28 Quit Jaykay (Ping timeout: 240 seconds)
21:45:06 Quit pamaury (Quit: Page closed)
21:48:02 Join Topy44 [0] (
21:48:44 Join Topy [0] (
21:50:27 Quit Strife89 (Quit: See ya.)
21:54:18 Join Kitr88 [0] (
21:56:49 Quit anewuser (Quit: What do you know...THE WORLD'S first NTRQ (that's for NES/FAMICOM) tracking compo. Have powerpak? Try it out! Otherwise ROM IMAGE.)
21:57:05 Quit Kitar|st (Ping timeout: 276 seconds)
21:59:07 Quit stoffel (Remote host closed the connection)
22:00:23 Join mischasworld [0] (
22:00:26 Quit mischasworld (Remote host closed the connection)
22:02:03 Quit merbanan (Ping timeout: 264 seconds)
22:02:19 Nick YPSY is now known as Ypsy (
22:03:58 Quit stripwax (Ping timeout: 265 seconds)
22:04:50CIA-5New commit by jethead71 (r25717): Optimized DSP sample out functions for armv6. (For stereo output) ~9% faster than SVN asm and about 4% faster than SVN asm rearranged to observe ...
22:06:18 Quit Kitr88 ()
22:13:20 Join Kitar|st [0] (
22:15:54 Join stripwax [0] (
22:26:01 Nick fxb__ is now known as fxb (
22:27:37 Nick Ypsy is now known as YPSY (
22:30:35 Join krazykit [0] (
22:30:53 Quit stripwax (Quit: Miranda IM! Smaller, Faster, Easier.
22:31:46 Join stripwax [0] (
22:32:03 Quit krazykit` (Read error: Connection reset by peer)
22:32:05 Quit mozetti (Ping timeout: 252 seconds)
22:42:22 Join Watermark [0] (
22:42:57 Join emrecelikten [0] (~anubis@
22:43:18Watermarkhello linuxstb, user gave me other info to share :)
22:43:26WatermarkJan 10, 2008 SPECIFICATION TCC79XX
22:43:29Watermark29-4 Preliminary
22:43:31WatermarkSDRAM Configuration Register (SDCFG) 0xF1000000
22:43:32Watermark31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16
22:43:34WatermarkCL BW CW RC RCD WR 0 0
22:43:35Watermark15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
22:43:37WatermarkRP RW 0 0 0 0 0 0 AM 0 PPD SR
22:43:38WatermarkField Name RW Reset Description
22:43:40Watermark0 SR RW 0
22:43:41WatermarkSelf refresh enable register
22:43:43Watermark0 : Disabled(exit from self-refresh state)
22:43:44Watermark1 : Enabled(enter self-refresh state)
22:43:46WatermarkBefore entering self-refresh state, the instruction and
22:43:47Watermarkdata in the SDRAM should be written back.
22:43:49Watermark1 PPD RW 0
22:43:50WatermarkPrecharge power-down mode enable
22:43:52Watermark0 : Disabled, 1 : Enabled
22:43:53Watermark3 AM RW 0
22:43:55WatermarkAddress matching option
22:43:56Watermark0 : BANK-RAS-CAS address order
22:43:58Watermark1 : RAS-BANK-CAS address order
22:44:00WatermarkThe address bit width for SDRAM banks is two.
22:44:01Watermark11 ~ 10 RW RW 3
22:44:03WatermarkRAS address width configuration register
22:44:04Watermark{0} : 12 bits RAS address width
22:44:06Watermark{2} : 13 bits RAS address width
22:44:07Watermark{1, 3} : 11 bits RAS address width
22:44:09Watermark15 ~ 12 RP RW 4 Precharge-to-Refresh cycle parameter
22:44:10Watermark19 ~ 18 WR RW 2 Write-to-Read wait cycle parameter
22:44:12Watermark23 ~ 20 RCD RW 4 RAS to CAS cycle parameter
22:44:14Watermark27 ~ 23 RC RW 10 Refresh-to-Idle cycle parameter
22:44:16Watermark29 ~ 28 CW RW 2
22:44:17WatermarkCAS address width
22:44:19Watermark{0,1} : 8 bits CAS address width
22:44:20Watermark{2} : 9 bits CAS address width
22:44:22ender`wouldn't pastebin be more appropriate for this?
22:44:22Watermark{3} : 10 bits CAS address width
22:44:23Watermark30 BW RW 1
22:44:25WatermarkSDRAM Bus width
22:44:26Watermark0 : 32 bits data bus
22:44:28Watermark1 : 16 bits data bus
22:44:29Watermark31 CL 0
22:44:31WatermarkCAS latency
22:44:32Watermark0 : 2 Cycles
22:44:34Watermark1 : 3 Cycles
22:44:35Watermark0x6A484C00 is the reset default vaule
22:45:10gevaertsWatermark: please don't paste lots of lines here
22:45:17stripwax*PLEASE* use pastebin or some other sharing resource for this kind of information next time
22:50:00 Join saratoga [0] (~9803c6dd@gateway/web/freenode/x-uxoetmnexcpttmhh)
22:50:21saratogawhen did the bot stop banning people for channel flooding?
22:52:22 Quit Watermark (Remote host closed the connection)
22:52:42 Join Watermark [0] (
22:53:07Watermarkhope this helps
22:57:13 Quit tchan (Quit: WeeChat 0.3.3-dev)
22:57:53 Join droidcore [0] (
23:01:00 Join Spaceghost [0] (irssi@unaffiliated/spaceghost)
23:01:25Spaceghostexist something for run Rockbox in Android?
23:02:16ollebeno. but it has been said that "the future for rockbox is an app"
23:02:20Spaceghostis possible now run android on iPhone 2G, and is near for the iTouch, so, can I use the rockbox's interface on it? insiode android?
23:02:42ollebeso maybe in the future, but not now, i guess.
23:02:47Spaceghostauch, *inside
23:03:12Spaceghostand what about can run rockbox in the iTouch?
23:03:31Spaceghostif will be possible run android, will be possible run rockbox on it?
23:03:34Farthenrockbox can not run as an app, yet
23:03:50Farthenmaybe in the future
23:03:57Spaceghostand have three option in the boot? would be cool
23:04:05ollebeSpaceghost: NO
23:04:15Spaceghostok, sorry
23:05:26 Join FlynDice [0] (
23:06:11Spaceghostwho is the bot that is loggin?
23:07:20Farthenhmm, scrolling through the list of users I clearly see a nick called "logbot". :-P
23:08:01bluebrotherSpaceghost: if you want to read the logs just check the topic
23:12:27 Quit bmbl (Quit: Bye!)
23:13:36***Saving seen data "./dancer.seen"
23:16:53SpaceghostI read
23:17:29Spaceghostlogbot, impossible more clearer name
23:19:58Watermark@linuxstb, what could you make out of the Link i posted
23:22:38CIA-5New commit by bluebrother (r25718): Log failures in mountpoint resolving, log filesystem formats in mountpoints().
23:23:44 Part Farthen
23:24:43 Join Farthen [0] (
23:25:36 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
23:26:13 Quit tchan (Client Quit)
23:26:44 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
23:31:47bluebrotherdomonoky: around?
23:32:01domonokyjup, but not long :-)
23:33:24bluebrotherI've added a repeat detection to the system trace logging function. I'm wondering if that's a good idea at all, as it will consume all repeated lines and only output a duplicate count when the next different line gets logged
23:34:01bluebrotherany thoughts on this? The idea came from some parts logging too much (like the Info tab ... logged the same string > 200 times for me :o )
23:34:16linuxstbWatermark: I know nothing about SDRAM configuration...
23:35:09 Quit bertrik (Ping timeout: 246 seconds)
23:36:48 Quit Highlander (Quit: Quitte)
23:37:40amiconnkugel: Well, an emulator would emulate a target, and run a target build on top of that
23:38:17amiconnThe simulator should still simulate the target, by means of running (almost) the same code compiled for the host cpu
23:39:30amiconnThat said, we could simulate tone controls for targets with hardware tone controls, *but* if we do that this should either use the host's hardware (if possible), or if it has to be done in software, it should be done in the simulator layer, not in the app layer
23:41:11domonokybluebrother: hm, the idea is nice. But it may get problems if we do something twize because of a bug.
23:41:56domonokyif for example talkfile generation speaks a string multiple times due to a bug, we wont see it.
23:42:54bluebrotherwell, I put a "(last message repeated n times)" in the log. The problem is that this output is delayed as I only know the number on the first different log message
23:44:24domonokyah, then it shouldnt be a problem. maybe add this log entry also when opening/saveing the systrace ?
23:45:52Watermarkcould you try that value out that was posted? 0x6A484C00
23:46:10bluebrotherthat notice is added to the log itself, so it appears everywhere −− on stdout for debug builds, the log viewer and when saving
23:46:10domonokyie end the dublicaton detection when getTrace() gets called :-)
23:46:37bluebrotherit's hooked into SysTrace::debug()
23:47:05linuxstbWatermark: That's the one you gave me a few days ago.
23:47:12domonokyyes, but if the log ends whith dublicates, you wont see this line.
23:48:02domonokywe could just log the getTrace() call to always get this line, even when a log ends in dublicates :-)
23:48:03bluebrotherah, right. Good point.
23:48:34bluebrothercurrent state:
23:48:55 Join mikroflops [0] (
23:51:16 Quit robin0800 (Remote host closed the connection)
23:51:21 Quit Watermark (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
23:52:00domonokybluebrother: looks good. just add a qdebug() to getTrace() and commit it :-)
23:52:08 Quit mikroflops_ (Ping timeout: 252 seconds)
23:54:04bluebrotherok :)
23:54:13domonokybluebrother: if you find time, can you test the paralell tts patch from archivator on Mac with TTSCarbon ? (i had problems with sapi, so we should make sure it runs fine with the other non-paralel tts before commiting)
23:55:47bluebrotherok, will try to find some time the next days.
23:56:50pixelmakugel: the release action really works better (before you sometimes triggered another tap (and messing up the bpm) when you used the "Long Mode" (Select repeat, I guess) to start "playing"

Previous day | Next day