Just a simple quick question, how do I can know about AudioUnit host application changes to sample rate?
My custom AUAudioUnit v3 subclass seems not to have any property to observe and AUHostMusicalContextBlock do not provide a sample rate inspection.
I need to react for sample rate changes in host application.
Thank you
I believe a safe place to do it is in allocateRenderResourcesAndReturnError. e.g.,
mySampleRate = self.outputBus.format.sampleRate;
This is where Apple's iOSFilterDemoApp does it
Related
Link to the bug report on 'Feedback Hub'
An audio endpoint device, from here on referred to as 'endpoint', is a physical or virtual audio output or input device.
With the Windows 10 April Update 1803 the long overdue 'App volume and device preferences' have been introduced. These settings allow more control over audio stream management as it is now possible to set different endpoints for different applications, no matter whether that particular application comes with an endpoint selection or not.
However, there is an issue where the audio of a program, whose endpoint is non-default, is streamed through the default endpoint (or not at all) after it has been closed and launched again, although the endpoint is displayed correctly in the settings:
As far as I know the issue can be recreated on a Windows 10 machine (version 1803 or higher) with any virtual or physical endpoint and an affected program. I used 'VLC Media Player' in this example (disregarding the fact that it comes with an endpoint selection) as it is well known and widely accessible, which should make it easier to recreate the issue.
What I'm searching for...
... is a programmatically solution to switch between endpoints, which ideally can be launched in form of a script to set the correct endpoint with an application launch.
For my purpose it would be enough to have to adjust the device instance path manually, as the device would be always the same, but I'm not going to complain about a solution which retrieves the device instance path from the registry, too.
Defined endpoints and the device instance path of the device they are using can be retrieved from the subkeys of the key HKEY_USERS\# YOUR SID #\Software\Microsoft\Multimedia\Audio\DefaultEndpoint. I don't know how windows generates the name of the subkeys or where they can be found. If I had to take a wild guess, I'd say these are Application IDs (feel free to correct me if I'm wrong).
The device instance path itself can be found in the Device Manager (under 'Audio inputs and outputs' double click the desired device, navigate to the tab 'Details' and select 'Device instance path' from the 'Property' drop-down menu).
Additionally the entry about Audio Endpoint Devices and Stream Management in the Microsoft Docs might be helpful, but that is way above my head.
A possible but impractical workaround...
... would be, to manually set another endpoint for the application and switch back to desired endpoint at every launch of said application (as shown above).
But not just takes this at least 10 seconds at each and every launch, you might even forget to do this as the audio might just get streamed through the default endpoint *¹.
The alternative to the latter is, that no audio will be streamed at all *² or in some cases it actually works *³.
*¹ e.g.: VLC Media Player, Tom Clancy's Rainbow Six Siege (although the audio will be streamed correctly during the splash screens)
*² e.g.: Call of Duty 4: Modern Warfare, Call of Duty: Modern Warfare 2, Call of Duty: Modern Warfare 3
*³ e.g.: Window Media Player, Microsoft Edge, Firefox
Observations
VLC Media Player comes with an endpoint selection, but so does TeamSpeak 3 and, unlike VLC, it skips the Windows settings completely.
Call of Duty not streaming any audio most likely is connected to the engine as I didn't encounter any other application doing something similar.
Windows Media Player, Microsoft Edge and Firefox are the only programs (I tested so far) which work fine. They have no endpoint selection (I'd know of) and will use the correct endpoint after closing and launching it again. It should be noted, however, that Firefox and Microsoft Edge will show multiple instances in the "App volume and device preferences" when adjusting the endpoint.
Disclaimer
I already tried two 3rd party softwares: 'Audio Router', which didn't work at all and 'CheVolume', which doesn't solve the issue and constantly crashes while doing so.
This question is based on one I asked over at Super User (here), where I didn't get an answer I was able to work with due to my lack of knowledge regarding actual programming (I'm only somewhat familiar with Batch and PowerShell). I'm well aware that neither Stack Overflow nor Super User are script writing services, however, the issue is not being fixed with the Windows 10 October Update 1809 and I see this as a problem which is affecting not just me and with that would be helpful for multiple people after me. Feel free to write a comment or propose an edit if you see this differently.
I'm also not sure whether the tags 'audio-streaming' and 'endpoint' should be used in this context, please propose an edit if they shouldn't or you can think of any better.
Edit - 05/11/18
Using the 3rd party software 'EarTrumpet' I was able to overcome the issue with the 'Call of Duty' games (no audio at all after restarting), however, 'VLC Media Player' would not restart after I assigned a non-default endpoint with 'EarTrumpet' until I closed 'EarTrumpet' again and the issue with 'Tom Clancy's Rainbow Six Siege' remains the same.
Edit - 18/01/19
Added link to a bug report I created on the 'Feedback Hub' 2 month ago.
Edit - 20/01/19
After doing some testing again it should be noted that having 'EarTrumpet' run in the background will keep a non-default endpoint for 'VLC Media Player' across restarts, however, 'VLC Media Player' will only (reliably) restart when the non-default endpoint was set in the 'App volume and device preferences'.
I do not have any solution regarding a programming language to handle such events.
But I can recommend EarTrumpet app to handle this change more quickly https://www.theverge.com/2018/6/13/17457778/eartrumpet-windows-10-audio-app
(Windows store: https://www.microsoft.com/en-us/p/eartrumpet/9nblggh516xp?ranMID=24542&ranEAID=nOD%2FrLJHOac&ranSiteID=nOD_rLJHOac-hUn6PgKuMKwQLdrzRqnPTA&epi=nOD_rLJHOac-hUn6PgKuMKwQLdrzRqnPTA&irgwc=1&OCID=AID681541_aff_7593_1243925&tduid=%28ir__qwqlg6jd0jba3y9hpnbvikaite2xk6kuyv9udtr100%29%287593%29%281243925%29%28nOD_rLJHOac-hUn6PgKuMKwQLdrzRqnPTA%29%28%29&irclickid=_qwqlg6jd0jba3y9hpnbvikaite2xk6kuyv9udtr100&activetab=pivot:overviewtab )
I will update the answer if I find a easy way to script/program a change of output on each app.
I have a pet project with webrtc audio-video calls. Currently calls from browser are working exactly as expected.
And I'm trying C++/Qt client based on Native C++ WebRTC (it is just a modified peer_connection_client example with modified signalling).
It works, but I've found an issue: when I make a call from one PC to another, the quality of the video is extremely reducing until bandwidth became around 250-300kbps (or 500kbps total, not sure).
As I told, there is a web version working on same signalling and I tested it out on the same PCs. The result was really surprised for me: no quality reducing and around 2000kbps network load.
Moreover, if I connect PC with cable to my router, the quality is fine and 2000kbps load, as expected.
I suppose that the problem is somewhere around wifi bandwidth estimator or so, but can't realize how can I control it's behavior.
Have someone any ideas how can I improve the quality and make webrtc use 2000kbps not 500kbps?
Thanks in advance,
Br,
Sergey
While going forward, I've found out that I have to call rtc::Thread::ProcessMessages() eventually to prevent being stuck on signalling_threads events.
But after that a new problem occured. It is "UDP send of XXX bytes failed with error 10035" this is described at https://groups.google.com/forum/#!topic/discuss-webrtc/wmYo7AU3evI.
I want to try some C and C++ programming with audio processing, such as synthesizers, chorus, delay etc, but I only know working with a console as output. I wish to have, instead of a console application, a window that would be capable of sending an audio signal to the speakers, running the code at the background and working with it in a similar way as it goes with printf: every time I would call the "output function", it would send to the speakers (or sound card) a sample value, indicating current oscilator position. This output operation could be executed every time it is requested or in the end of a built in loop. Doing all this with a high sample rate would be just great.
I think I could do all this using AudioWorker on Web Audio API, plus a flexible GUI on HTML5 canvas, but I'm new at this API and I'm not sure whether its resulting sound quality is good enough.
Thanks in advance.
Edit: I use Windows 8.1, but any answer for other platform is welcome.
Edit2: Any programming languages other than C, C++ or JavaScript suggestions are also welcome.
I do lots of sound synthesis with the Web Audio API and I think it sounds great. Javascript is really all you need. Well, if you want to use audio files, you need a web server to serve those audio files, but the audio synthesis all happens in javascript.
It doesn't matter so much what OS you use, but different browsers have different levels of support for the web audio API. Chrome tends to have the best support, and Internet Explorer definitely has the worst support.
Some people suggested https://github.com/calabash/calabash-ios for iOS tests. I am using cocos2d-iphone, can I write tests for it using calabash?
Still calabash turned me off because it says it will only work with the simulator which is not enough and also if devices I have to use a service which just sounds like trouble. Any clean solution?
After completing this guide you will be able to run tests locally
against the iOS Simulator. You can also interactively explore and
interact with your application using the Calabash console.
Finally, you will be able to test your app on real, non-jailbroken iOS
devices via the LessPainful service.
Edit: https://github.com/calabash/calabash-ios/wiki/07-Testing-on-physical-iDevices suggests that device testing is possible... I guess I'll just have to try it myself...
Edit 2: I did get it all working. Even hooking up to a remote device that is not even plugged into the computer (just needs to be on same WIFI). When hooking up to the device it helps using the UDID of the device, this was not stated on calabash-ios docs, or I missed that.
For anyone coming here later, this is the final command that will work:
DEVICE_ENDPOINT=http://192.168.36.180:37265 BUNDLE_ID=build/tabeyou-cal DEVICE_TARGET=thelongudidofyourdevicegoeshere OS=ios6 cucumber
Just replace the ip to the ip of your device, the udid, and BUNDLE_ID which should be the target name (I think).
My current question is, how would I identify cocos2d stuff like CCMenu, CCSprite etc? These all seem to support accessibility identifiers and I'm sure the Ruby-iOS part could find anything under the hood - in turn that should make it possible to write tests interacting with cocos2d elements.
I try to send/read a control message with a specific setup packet to an USB device.
I've found at MSDN this documentation: http://msdn.microsoft.com/en-us/library/ff537344%28v=vs.85%29.aspx and at the usbuser.h the struct "_USB_SEND_RAW_COMMAND_PARAMETERS" which can be filled with the parameters for a setup packet. The problem ist that MSDN says about this function: "Do not use this request".
The next try was the request code "USBUSER_PASS_THRU" but I don't know what the parameters mean and I don't think that it is possible to send a specific setup packet with this request code.
I can't use WinUSB because I would like to solve this without any installation or other requirements to the target PC.
Has anybode solved this problem or has an hint to solve this?
Thank a lot.
Regards
If you want to avoid driver installations, you could make your device emulate an Human Interface Device (HID) or Mass Storage Device; both of those types of devices work automatically with the built-in Microsoft drivers.
If you figure out another way to do it, I'd be interested to hear the answer.