2011/2/23 Alexandre Prokoudine <alexandre.prokoudine(a)gmail.com>:
> On 2/22/11, David Robillard wrote:
>
>> I have a working plugin (called "dirg") that provides a UI by hosting a
>> web server which you access in the browser. It provides a grid UI either
>> via a Novation Launchpad, or in the browser if you don't have a
>> Launchpad. Web UIs definitely have a ton of wins (think tablets, remote
>> control (i.e. network transparency), etc.)
>>
>> I also have a complete LV2 "message" system based on Atoms which is
>> compatible with / based on the event extension. Â Atoms, and thus
>> messages, can be serialised to/from JSON (among other things,
>> particularly Turtle).
>
> Any of them available to have a look at?
>
>> Currently dirg provides the web server on its own with no host
>> involvement, but every plugin doing this obviously doesn't scale, so
>> some day we should figure this out... first we need an appropriately
>> high-level/powerful communication protocol within LV2 land (hence the
>> messages stuff).
>
> Where do you stand with priorities now? That sounds like something
> very much worth investing time in.
>
> You see, one thing I'm puzzled about is that you have beginnings of
> what could be significant part of a potentially successful cloud
> computing audio app, and then you talk about how donations don't even
> pay your rent :)
Before I totally forget about it... I think it might be a very clever
thing to do to have some web-based thing (wiki or whatever, ideally a
social network kind of thing) were LAD people can notify of what they
are working on and what are their plans, so that it's easier to: a.
know about it and b. start cooperations, etc.
For example, Dave is doing lots of stuff that I plan to reuse, but I
only know it because I happen to lurk on #lv2 on freenode from time to
time, and the same goes for lots of stuff I'm seeing coming out
lately.
If it is a problem for me to keep up to date with this stuff, I can
only imagine what it would be like for a newcomer.
I can't comment on this more now, but please somebody consider the idea.
Stefano
Hi everyone,
I am currently spending a lot of time working on Android, and on the andraudio
mailing list [1] we are discussing about possible improvements to the internal
Android audio system. Currently, latencies are very high, over 100ms, and we're
looking for ways to improve the situation.
In my opinion this can't be achieved on Linux without realtime scheduling. On
Android, there's something called audioflinger which acts as a sound server, and
apps act as clients of this server. The server and clients run in distinct
processes. What I'm thinking about is having a realtime thread within the
server, as well as another realtime thread in the (each) client.
The one thing about Android is that it has a strict security model. Every app is
considered potentially harmful and is thus "sandboxed". Here, this for example
means that apps can lower their threads priority, but not increase it. And of
course they can only use non-realtime scheduling.
On desktops, for instance using JACK, apart from a few multimedia distributions,
realtime permissions are not granted by default to normal users. And when one
enables it, security is usually not a primary concern AFAIK. If a piece of
software happens to crash the system when running in realtime, the user may just
uninstall the buggy software, etc..
But on phones, this is critical, for example if the system crashes while you're
waiting for a call. So, on Android, the security policies are strict. But this
could certainly be necessary to plenty of other Linux usages.
Now my question is: how to allow user-space apps to use realtime scheduling for
one of their threads without compromising the overall security?
For example, in man sched_setcheduler() I see SCHED_RR, and "Everything
described above for SCHED_FIFO also applies to SCHED_RR, except that each
process is only allowed to run for a maximum time quantum".
Would this help and be sufficient? Would there need to have some kind of
watchdog/monitor running with SCHED_FIFO scheduling to prevent realtime client
threads from consuming too much resources?
Or is there some other ways to achieve this? Some kernel patch maybe?
Thanks in advance
[1] http://music.columbia.edu/mailman/listinfo/andraudio
Olivier
Hi, I've written a small program (Leevi is its name) to drive Lexicon
MX300, similar console what Lexicon ships for Windows operating system
along with their devices. Leevi supports Linux and BSD, and needs
either libusb-0 or libusb-1. It does not require any user interface
libraries, as everything is built on libX11, so if you can boot to
runlevel 5, you can run Leevi.
Please note: As Lexicon wasn't very keen to tell me how to talk with
their devices, USB protocol is reverse engineered by me and there are
myriad of things that I haven't yet revealed. Although Leevi is proven
to work without breaking anything, there is always a small change that
something goes wrong. Therefore I don't take any responsibility, you
use Leevi at your own risk. Check BSD License in Leevi's homepage.
Leevi is yet in development stage, changing effect in stereo mode and
changing routing to/from stereo does not yet work. All other features
should be fully functional.
http://leevi.sourceforge.net
- Jani Salonen
Hi list,
i want to write a jack2 network backend, like netjack2, for my bachelor
thesis.
There is a new network standard comming up specialized for audio and
video transmission, called AVB (Audio Video Bridging -> IEEE 802.1AS,
802.1Qat, 802.1Qav,...).
I want do integrate this standard into jack2.
I want to choose AVB (not ALSA or NET) from the driver dropdown box.
does anyone know where i can find this implementation in the source code?
On Thursday 24 February 2011 12:09:19 David Robillard wrote:
> This is, of course, a big problem in terms of our greater mission
> to provide software that caters to the needs of precisely nobody while
> irritating everybody else.
>
> To resolve this situation, we now have an exciting new Clippy inspired
> assistant that hops around your screen begging you to add MIDI tracks
> constantly.
>
> If, after 20 minutes, you have still not created a MIDI track, Ardour
> will overwrite every .waf file found in your home directory with a 4/4
> electronic kick drum loop, then shut down.
>
> -dr
Dude!
This is just what I have been wanting but am too much of a newb to have even
thought of undertaking such a complex project on my own!
Hence my starting my soundwall project which I newbily thought was going to be
a simple little app.
I have some code inspired by some "anon" silence that could perhaps be
included in this project. wwnnsnmsnm.
cage-ily,
drew
On Tue, Feb 22, 2011 at 3:49 PM, Philipp Ãœberbacher
<hollunder(a)lavabit.com> wrote:
> The rest sounds nice, and it might well be that X has become old, but I
> don't see the big improvement coming up. Windows are called surfaces
> now, can have different shapes and are more flexible, compositing,
> transformations, I got that bit, but I don't see the UI improvement.
> I've seen the demos with shapes flying around the desktop, I've seen
> the conventional compositing window managers and wayland will probably
> do all that and more, but I don't see the improvement in User
> Interfaces.
what its going to do,i think, is two-fold:
1) promote more and more toolkit design that makes everything just a
compositing stack. GTK has already moved significantly in this
direction, but could go a lot further. Qt is in a similar position.
the more this happens, the easier it is to reason and create new GUI
widgets that do cool things, easily and simply, because its all part
of a very simple model: you draw to your surface, it will be
composited onto the screen in ways that you don't have to worry about.
sounds a bit like X ... except that X is explicitly *not* a
compositing model. for a simpler explanation of the kind of thing i
mean, consider the difference in ardour between the main "tracks" area
of the editing window and all the widgets around it. its fundamentally
impossible to implement the tracks with widgets - it uses a "canvas"
object instead which embodies idea like z-axis stacking, transparency
and so forth. but likewise at present it would be a lot of work to
implement all the widgets as canvas "items". now fast forward a few
years, and find a spot where the drawing model for the canvas, the
button widgets, the tree/listviews, for everything *inside* the
program is the same as the model for everything *outside* the program.
drawing a particular "thing" on any other thing becomes identical,
whether the other thing is a "window", a "button", a cell of a
listview, etc, etc.
2) more and more apps able to take advantage of v-blank sync to reduce
computational load due to unnecessary redraws. instead, the whole
system will be a lot like a video-framebuffer version of JACK: the
vblank interrupt arrives. everything with a surface gets a chance to
redraw if it needs to, the surfaces are composited together, and boom,
its on the display. no more guessing how often to redraw stuff, no
more wierd ass hacks to get smooth animation, etc. if you think this
sounds like special effects, i suggest a few minutes playing with a
relevant iPod/iPhone/iPad app where these smooth transformations of
what is on the screen is a central metaphor in how the UI's work.
Oh, well, since GUIs are this week's flavor in the list, I might as well
throw FLAM's alpha in the mix.
I am preparing FLAM's (Front-ends for Linux Audio Modules) first
release. Among other things, FLAM intends to allow programmers and
non-programmers alike to create their own (external) GUIs for audio
plugins. At this moment only Rosegarden as a host and LADSPA as plugin
type are supported, but this is hopefully just a first step.
Project page:
http://vagar.org/code/projects/flam
Tutorial with a few screenshots:
http://vagar.org/asciidoc/flam/primer/primer.html
Source code repository:
git clone http://vagar.org/git/flam
I'd welcome any feedback, here or in the FLAM forums (registration
required):
http://vagar.org/code/projects/flam/boards
Thanks!
Luis
do any of the jack client examples show playing a file from disk?
if so, which?
if not, any links to simple code that does this?
c++ or c?
thanks in advance for any pointers.
drew
--
http://freemusicpush.blogspot.com/
Hi all,
This meditation isn't about gray hair, sagging flesh
or receding libido. :-)
In debugging Nama's new audio editing functions,
I'm noticing that I have more and more code dealing
with time. I am concerned seeing a multiplicity of
method names like adjusted_region_start_time
or unadjusted_mark_time.
I'm wondering if perhaps I can centralize
or at least systematize this functionality.
Nama deals with several kinds of time:
Ecasound time. Positions in seconds or samples from
the perspective of the Ecasound audio engine
WAV time. Displacements in audio files.
Track / Region time. Positions in a track or region
Mark time. Nama currently has only one type of mark:
marks anchored to an absolute project timeline.
I think it also needs marks for positions in a tracks WAV files,
which the user may trim or offset.
Edit / offset-run time. This is what started the entire
issue. To record a fix for a note at time T in a WAV file W,
I use Ecasound's select object to offset all the WAV files
in a project to start at time T. The fix, W', then gets
placed at T using playat.
MIDI time. There are plenty of references on this,
and it's a subject of its own. Nama needs at least to
know enough to work across the various systems for bridging
between ALSA, JACK and MIDI.
So do I need some Big Abstraction(tm), or shall I just
continue to work incrementally?
How do you think about time?
I don't expect a simple answer or an answer at all,
but it can help to formulate the question.
Regards,
Joel
--
Joel Roth
hi all,
for some time, i have been working on nova-simd, a cross-platform SIMD library
library written in c++. it abstracts the platform-specific APIs and provides a
`vec' class, which maps to a SIMD floating-point vector on the target machines
and which can be used to formulate more complex algorithms. for some commonly
used vector functionality, helper functions are provided, which are built using
the vec class.
features:
- same interface for different SIMD architectures
- supported backends: sse-family, avx, altivec, arm/neon
- some backends include vectorized implementations of libm functions
- extensively used in supercollider
- header-only c++ (no runtime dependencies, composable), no dependencies
canveats:
- little documentations/examples
- no release, no tarball, just a git repository [1] and a web interface [2]
- header-only c++ (no c support)
maybe it is useful for other people as well...
cheers, tim
[1] git://tim.klingt.org/nova-simd.git
[2] http://tim.klingt.org/git?p=nova-simd.git;a=summary
--
tim(a)klingt.org
http://tim.klingt.org
Art is either a complaint or do something else
John Cage quoting Jasper Johns