Announcing disptec project
RidgeRun is working along Costa Rica TEC into a project for developing xDAIS DSP algorithms for TI processors.
Here is the press release for the resources donations.
Here is the project’s blog.
RidgeRun is working along Costa Rica TEC into a project for developing xDAIS DSP algorithms for TI processors.
Here is the press release for the resources donations.
Here is the project’s blog.
Thanks to the collaboration between RidgeRun, Costa Rica Institute of Technology, and Texas Instruments, we are announcing our first conference on the subject of embedded technology. We will be doing presentations for students on several subjects and introducing them to the research project on DSP algorithm development.
For more information visit the conference page.
Cristina has finished her graduation project with the Leopard board, and now we have (almost) completely AAC-LC and MP3 encode/decode support with the gstreamer plugin.
We have tested the encoders to properly generated standalone AAC or MP3 files, as well as muxplex them into a mp4/quicktime container format. The only remaining issue is related to the the encoders being unable to be used at the same time as the alsa capture driver (Cristina already ask Ittiam about the problem, but we haven’t got any feedback yet). In any case the issue is reproducible with Ittiam’s command line demo apps, so we don’t think is a gst issue.
As usual the beauty of a gstreamer solution is on the simplicity of the examples:
Encoding AAC audio into a quicktime container (two steps due the encoder bug):
gst-launch alsasrc ! audio/x-raw-int, width=16, depth=16, endianness=1234, channels=2, rate=44100, signed=true ! filesink location=/audio.pcm
gst-launch -e filesrc location=/audio.pcm ! audio/x-raw-int, width=16, depth=16, endianness=1234, channels=2, rate=44100, signed=true ! dmaienc_aac ! qtmux ! filesink location=/audio.m4a
Decoding the AAC audio from quicktime container:
gst-launch filesrc location=/audio.mp4 ! qtdemux ! dmaidec_aac ! alsasink
Encoding AAC audio elementary stream:
gst-launch filesrc location=/audio.pcm ! audio/x-raw-int, width=16, depth=16, endianness=1234, channels=2, rate=44100, signed=true ! dmaienc_aac outputformat=2 ! filesink location=/audio.aac
Decoding AAC elementary stream:
gst-launch filesrc location=/audio.aac ! aacparse ! dmaidec_aac ! alsasink
Encoding MP3 elementary stream:
gst-launch -e filesrc location=/audio.pcm ! audio/x-raw-int, width=16, depth=16, endianness=1234, channels=2, rate=44100, signed=true ! dmaienc_mp3 ! filesink location=/audio.mp3
Decoding MP3 audio:
gst-launch filesrc location=/audio.mp3 ! mp3parse ! dmaidec_mp3 ! alsasink
The support for these features is available on DDOMPE’s branch, and will likely be available on the 2.0 release of the gst-ti-dmai plugin.
Enjoy.
Some news surface last week about the hawkboard.
We are looking forward on RidgeRun to get our hands in doing some work with it. We already sign up two students from the Costa Rica Institute of Technology to do their internship project with us developing algorithms with the powerful DSP available on this boards during Q1/Q2 of next year. We will be applying next week to be part of the early adopter program.
The L138 already have support for GStreamer on a branch of arago-project, but this would be a good opportunity to push for stabilization of it (and maybe include it on 2.0 release of the gst-ti project?). We could have some nice demos (as Cristina and Maria did for Leopard board) with extended DSP algorithms as elements on the pipe.
Thanks to the kind invitation from Claudio Navarro (director of the “programming school”), Diego was invited to speak at the Colegio Universitario de Cartago (CUC) in Costa Rica about the work we do at RidgeRun programming embedded systems during last week.
They are on the process of refreshing their study plan, and now considering including some elements based on the feedback provided by the industry. We hope to see the CUC students doing some work with embedded in the future, as the ITCR is doing.
Well let’s face it, good embedded systems developers are still a rare breed. An embedded engineer takes a good mix of hardware and software skills, but a really good one also needs great communication and soft skills. We at RidgeRun are always looking for good embedded developers, and although we do our R&D in a wonderful country, with great investment on the high tech market and a growing technological environment, the embedded developers are still on short supply.
We have worked hard in the last six years with different players on the education sector to improve the skills for this market. The ITCR (the top tech university according to all tech sector surveys) has come to rescue with a new degree for computer engineering, and we are very happy that a member of our team (Miguel Aguilar) is giving the introductory course to embedded systems.
This is the first iteration of the course and it will be evolving over time, and by now is imparted to advance EE students. The course labs will be done with Texas Instruments MSP430 (16 bit microcontroller), and some BeagleBoards (32 bit SoC). On the software side they will be programming with Linux, and if time allows, DSP/BIOS for RTOS on the DSP of the Beagle.
I’m pretty sure Miguel will do a great job teaching, as he does as an outstanding engineer at RidgeRun. The first draft of the program course can be found here (in spanish).
Diego