Hi Nuno,
Thanks for this opportunity to allow me to help you understand audio
latency on Android better.
Contributions to total round-trip audio latency on Android are best thought
as stemming from two parts.
A) Android service space (AudioFlinger etc) ← developers and middleware
cannot affect
Hello Nunos,
I just joined https://groups.google.com/forum/#!forum/andraudio
I think that is a good place for discussion.
If you repost your questions there then I will try to answer them.
Phil Burk
On Wed, May 25, 2016 at 6:15 AM, Nuno Santos
wrote:
> I have been
Patrick,
Now that I had the chance of using Superpowered all I can say is that it
doesn’t make miracles. It will not provide lower latency than OpenSL already
gives. Android latency problem is much OS and device dependent.
Looking forward to understand what is new and if it will bring any
On Wed, 25 May 2016 09:29:30 -0300
Eder Souza wrote:
> Python is known as a friendly language, rapid developments, but with poor
> performance, I've never seen anyone here saying that tried something in
> real time using python, so I decided to test using some lib's (numpy,
Speaking of Android, we just released a fully configuarable and performant
USB Audio and MIDI library for Android. It’s the Android equivalent of iOS
Core Audio and Core MIDI for USB devices.
http://superpowered.com/android-usb-audio-android-midi
The Superpowered USB Audio and MIDI SDK provides
Also, only way to actually know if your compiler generates Neon is to ... read
the assembly!
Some common DSP tasks are implemented in the Ne10 library
http://projectne10.github.io/Ne10/doc/ which compiles to Neon on systems that
support it.
I have seen that nova_simd is good at generating neon
Hi,
This demo is running using buffer=4096 (this provide a good resolution for
numpy compute an FFT), I need stream with a buffer=3072, and its left
the remaining
samples(1024) to overlap and add again when the new frame arrives.
So it is a big buffer size, but numpy and pyaudio seems fast for
I have been testing DRC in 2 different Android devices: Nexus 9 (48000/128
buffer size) phone, Bq Aquaris M5 phone (48000/912) buffer size
Nexus 9 is a beast. I was able to run DRC with full polyphony (8 voices)
without any kind of glitch. The fake touches hack was essential to make this
Hi,
I do not see drives in the andraudio group since he migrated to [1]
Some years ago I tried see how some of my codes works in android, I wrote
effects to work in real-time (automatic pitch correction, time stretch,
pitch shifters in time domain and frequency domain) and I get a lot of
At JUCE / ROLI we've been working with Google for over a year to optimize audio
latency, throttle, etc for cross platform apps. Our app is featured also in the
Youtube video from Google IO, and it runs on some devices with performances
comparable to iOS.
Whether you are using JUCE or not,
Hello,I watched the google/android video the other day and I was very surprised
of the basic recommendations that were given to developers:- use native
sampling rate (don't hardcode sampling rate, request it from the device)
- use native buffering (don't hardcode it. But then maybe use double or
Hallo!
Thanks Nuno, it was a great demo ;)
LG
Georg
On 2016-05-25 12:46, Nuno Santos wrote:
> Hi George,
>
> I would be interested in such a community as well. Specially regarding
> audio performance. We have recently released DRC (one of the apps that
> has been featured on Google I/O Android
12 matches
Mail list logo