Back to the relevant info :

True. You have gone straight and I am admittedly the villian now.

I just though you might were ready for some more entertainment....you know; us trolls in between

But fair enough. Good luck walking the line then.

Peace

Android's weakness in real-time audio is the reason why I bought overpriced iPhone 3GS instead of some 2-3x more powerful Android phone. it's a great phone tho, so i'm not complaining + iMachine is fabulous!

i'm in exactly the same boat. as soon as theres a version of amplitube or ampkit for android then i'll immediately ditch my iphone.

i'm waiting for android4 to fix this latency issue - otherwise i'll regrettably be upgrading to an iphone 4S.
ive actually used my 3GS as an emergency amp(sim) in sudden live situations.

Did you just buy it recently? The 3Gs and the original Droid (which I had) were out around the same time, and the performance was pretty much the same.

Got android 4.0.3 today, my transformer tf101 tablet is now really fast and music apps are somewhat more responsive!
Notice less latency while hitting notes, not perfect but definitely better then before.
Caustic still shows the same latency, but i'm happy with this update.

Recently got a Galaxy Nexus with ICS. Just tried Caustic demo on it. It reports 89 ms latency. From how it feels I'd guess it's even more. (Though how much of that is on the GUI side is hard to say.) I'm also guessing this figure reflects the buffer size Caustic is pushing to android.media.AudioTrack, which could very well have further buffering after it that's transparent to the app.)

I'll be following this one closely, there's no good reason why latency performance should be so poor. Even if they gave apps the ability to request shorter buffers, that would do it. (And a proper callback function for the audio process, instead of pushing and blocking until the buffer is ready for more.) Then you could tweak latency to what the particular hardware is capable of, just like in the desktop/laptop world. If the hardware can't keep up, oh no, dropouts. We're all already used to that. But to the semi-pro audio uninitiated, I could see them thinking that's failure, thus they might hard code in unnecessarily long buffers. This is exactly why for so long Windows needed a 3rd party driver format, ASIO, in order to do low latency audio before WDM came on the scene and finally caught up/surpassed that performance. Unfortunately, I don't think there's such thing as installable drivers for these phones. (At least not without rooting the thing, replacing the entire OS, and voiding the warranty. Obviously not an option for anyone hoping to widely market an app.)

By the way, this issue has been pushed to #11 on Google's bug list! http://code.google.com/p/android/issues/list?can=2&q=&sort=- stars&colspec=ID%20Type%20Status%20Owner%20Summary%20Stars

Keep hitting that star folks! (On this page. http://code.google.com/p/android/issues/detail?id=3434 ) Maybe we can make this problem #1.

And you know what it's the #2 defect on the list. I think seeing as how it's defect is next to the Nexus Battery life it's should be getting some attention. The android team has already said they know about the issue. Hopefully they are working on it, they said they were hopeful it would be in ICS but we know it's not.

Someone (Windows X) has been working on 5ms - 10ms latency on Samsung Nexus S :

http://forum.xda-developers.com/showthread.php?t=1621914

Maybe others will follow...
Last edited by UltraJv on Fri May 11, 2012 10:05 am; edited 1 time in total

I don't see hacking the OS as a viable solution. We need a sanctioned way to do this so we can sell products to everyone. Not just those willing to tinker.

Its a start, it could end up in sanctioned Android code or it could end up as ASIO4ALL did.

Can anyone tell me how to test Audio low latency on Android 4, ICS ??
it will be of great help to me

To test, load up a music app, press a key and listen for delay. (?)

If you want a measurement, you can install Caustic and look in the menu. The app's buffer size is converted to milliseconds and displayed as latency. Note that this does not account for Android's internal mixer buffers or touch latency, there's no easy way to measure that.

You could record audio using an external recorder. The test would be to tap the screen "loudly" and measure the time between the screen tap and the audio coming out of the device.

Thanks for the reply.

By the way can this Audio Latency be tested as part of CTS verifier?

Samsung devices are the worst in regards to latency. Try a Motorola Device, their response time is really, really good. I used to have a Droid I, and even playing drum pad apps it "felt" really low latency. Now I have a Galaxy Nexus (Samsung) and the latency is terrible. So it depends on what manufacturer you go with. It's not just and "android" OS problem.