David McGriffy wrote:
Oh how I wish I had time to play with this stuff. At this moment on my
breadboard I have a Teensy 3.2, a 6DOF gyro module, etc. Bluetooth modules
should be in today. 10DOF modules on order. I think the magnetometer is
important for HT to prevent drift. The altitude, perhaps not so much. Of
course, this is for a drone project (I'm writing a book about building
drones). And I have audio projects backed up ...
For audio and head tracking, I have been playing in Unity to Google
Cardboard and now GearVR. I have some demo code for anyone interested in
Unity development.
The standard I hear in the VR world is 20ms "motion to photons". Minimum
frame rate of 60Hz, preferably 90-120Hz. These faster rates do not allow
convolution with the full block size of the listen database, though careful
truncation should be OK.
David McGriffy
VVAudio
https://depositonce.tu-berlin.de/bitstream/11303/1318/1/Dokument_28.pdf
(System latency)
Chapter 4.1.3
Figure 4.4 displays the results together with the mean value of all 17
persons and a
95 %-confidence interval depending on the total latency time. An
average value of 0.5
was assumed as the threshold that needs to be passed to allow a
perception. Hence
if the 95 %-confidence interval of the mean value exceeded this limit,
the respective
total latency time was counted to be perceptible.
Each (total) latency time falling short of 85 ms was ignored by the
subjects. The
transition between “not perceptible” and “perceptible” occurred in a
range between
85 ms and 101 ms. For latency times exceeding 101 ms localization
artifacts were
perceived.
5.2.1
Sandvad and Wenzel [110, 136, 137] have previously investigated the
influence of
the system’s total latency time on localization. They concluded that
for producing
real time auralization the auralization system has to meet the
following requirements:
The total latency time has to be below 91 ms, the update rate has to
be at least 60 Hz
and a minimal spatial resolution of about 2± is required.
The experiment at IRT which focused on latency (described in section
4.1.3) produced
similar the results, i. e., a maximal latency time of 81 ms using an
update rate
of 120 Hz and a spatial resolution of 5± (due to the step-motor). The
head tracker
system itself had a latency time of theadtracker = 8.3 ms and an
angular resolution of
0.1±.
Since the electronic auralization system (BRS Processor) itself has a
shorter latency
time (tBRS < 6 ms), the system’s total latency time is in the order of
magnitude
of about ttotal = theadtracker + tBRS ' 15 ms. Thus does not exceed
the upper limit
of 85 ms and fulfils the requirements previously mentioned.
The link to this dissertation originally came from Aaron Heller.... (I
was combing some "old" mails... :-) )
The standard I hear in the VR world is 20ms "motion to photons".
I somewhat would like to express my doubts!
You must be aware that update rate and latency are s.th. quite different.
20ms (max.) latency looks (suspiciously) close to 1s/(min. update rate).
I would have thought the same. However, max latency should be <
significantly greater > than (1/min. update rate) s.
(?)
The standard I hear in the VR world is 20ms "motion to photons".
Sources?! No "hearsay" accepted, at least not at this point!
Best,
Stefan
P.S.: It was all about "Never do MATH in public"... ;-)
_______________________________________________
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit
account or options, view archives and so on.
