Warning: the discussion is drifting to DIY electronic gadgetry. :)
On Sun, 31 Jan 2016 11:16:26 +0100,
Bo-Erik Sandholm <bosses...@gmail.com> wrote :
I have decided to simplify the DIY head tracking dongle build and
setup in some aspects, now I have ordered this sensor that do not
need initial calibration.
This is the new sensor module:
https://learn.adafruit.com/adafruit-neopixel-uberguide/overview
The page is about addressable LED modules. Is it an error?
I would use a GY-85 board and a micro-controller, as seen here:
http://www.rcgroups.com/forums/showthread.php?t=1677559
This is a good starting point.
It will initially be combined with a esp8266 module for WiFi
connectivity or maybe Bluetooth
http://www.esp8266.com/wiki/doku.php?id=getting-started-with-the-esp8266
http://www.aliexpress.com/item/Promotion-Brand-NEW-HC-05-Wireless-Bluetooth-RF-Transceiver-Module-serial-RS232-TTL/32367579918.html
Could there be some added latency when using wifi or bluetooth? A
direct usb connection should be faster, but avoiding a cable would
be desirable because many android devices cannot easily use their
usb port for communication. If using wifi, I would try multicast
udp.
Here's a page that explains how to use the bluetooth module:
http://www.instructables.com/id/Cheap-2-Way-Bluetooth-Connection-Between-Arduino-a/
Power will probably be from one of these, giving around 10 hours
of operations:
http://www.aliexpress.com/item/4PCS-Hot-Sale-Soshine-900mAh-14500-battery-3-2V-LiFePO4-AA-Rechargeable-Battery/32242320597.html
Nice!
I will send OSC (open sound control
https://en.m.wikipedia.org/wiki/Open_Sound_Control) directly from
the sensor.
OSC is a good protocol, but an application specific protocol could
be designed to be more compact, reducing the latency.
This should simplify the build of the head tracked sensor,
reducing the soldering need.
There would be 4 modules involved: a sensing assembly, a
micro-controller, a wifi transmission module, and a power supply.
Going usb-wired would remove the wifi transmitter and the supply.
A custom firmware can be programmed for the ESP-8266, which have
GPIOs, so maybe it could be used as a micro-controller:
http://hackaday.com/2015/03/18/how-to-directly-program-an-inexpensive-esp8266-wifi-module/
If a micro-controller is required, the trinket is an alternative to
the arduino nano: https://learn.adafruit.com/introducing-trinket/
It's much smaller, works at 3.2V. For a 5V USB wired version, it can
provide 3.2 volts for other boards.
This should simplify the setup of playback using
http://www.matthiaskronlachner.com/?p=2015
And maybe later ambiexplorer can be modified to accept OSC data?
It could even be used with a browser (chrome) based player.
In the end, the first problem to avoid is latency, and it can
invalidate many potential solutions.
This will allow you to use any headphones and DAC and amplifier
Best regards
Bo-Erik
I already bought some of the parts to create a head-tracking device,
months ago. Let's do it and share the designs. Even if we have
personalized HRTFs with order 1024 decoders, we need head-tracking.
The other solution is to use the sensors in phones or oculus-like
devices, but they are all too big or a bit expensive for the task
of listening to binaural audio only (not combined with visuals).
_--------------
I want to see a good quality over the ear stereo headphone with
all necessary electronics built into the headband. It will have
single usb connector which will provide power and digital audio
(24 bit) and carry head tracking information back to the
computer, which will have the software to play standard first
order B-format files decoded to binaural, using simple HRTF
filters. The computer can be your desktop, a tablet computer or a
mobile.
umashankar
For a DIY project, integrating head-tracking and audio would be a
lot, and the resulting device could be rather large. But I may be
wrong.
--
Marc
