Why sound isn't working on the TP2.

[slacksoul]

We did it!

Thanks to the efforts of Jonpry, (with special thanks to [ACL] and Wozzer) we finally have sound on the TP2!
please visit http://forum.xda-developers.com/showthread.php?t=627997 for more information.

Also, please consider donating to Jonpry via paypal at his email address: jonpry [at] gmail *dot* com

For Historical Purposes
I'm posting this thread because I noticed there's a lot of confusion regarding the status of sound in the Rhod android port, information on the exact reason why sound doesn't work is hard to come by and complicated; so hopefully this thread will clear some of that up.

The Rhodium, like most modern smartphones, has more than one processor, it has four to be exact. Of these processors, two of them are general purpose and two are media processors as follows:

- 1 ARM9 processor that handles the communications stack.

- 1 ARM11 processor upon which the OS (winmo or android) runs on.

- 2 DSP processors that handle the sound and camera functions, they are the QDSP5000 and QDSP4000, they are under the control of the ARM9 processor.

Whenever the phone's OS (WinMo or Android) wants to play sound, it needs to invoke the QDSP5000 chip from the ARM9 processor through a special communications channel between the ARM11 and ARM9 processors. Because this is such a common arrangement in smartphones we already have a standardized way of doing this in the Linux kernel, unfortunately when we attempt to invoke the QDSP5000 chip through this method, it simply does not respond.

This has absolutely stumped developers, there is no reason it shouldn't work, there is nothing wrong with it and it works on every other HTC phone with this chipset, except the Rhodium and Topaz.

How come sound during calls works then?
We have absolutely no idea, it's theorized that WinMo somehow sets it up so that sound during calls is handled exclusively by the ARM9 chip, so when we boot up Android via HaRET, this carries over and in-call sounds are handled all by themselves. Unfortunately this does not carry over to application sounds like the ringer, music and other app sounds.

What can we do to help?
The issue is quite complicated, and at this time it's agreed that the only way to make headway on the issue is to decompile the Rhodium's WinMo sound driver: wavedev.dll. Without the crucial information in this driver, we will never be able to get sound working in the TP2, decompiling drivers is a very tedious, time-consuming and complicated task, and none of the current devs are able to do it at the moment.

If you, or someone you know, has experience in reverse-engineering and decompiling as well as strong C++ and C skills, please take a look at the code. The only way we are going to crack this is by having more eyeballs look at the code.

~ALERT!~(07/16/2010)

Jonpry has developed a wrapper that loads the WinCE "wavedev.dll" driver, eliminating the need to reverse engineer it. Unfortunately loading the wince driver in Android is still not giving us sound, so we need people to help test and develop Jonpry's new 'Vial' wrapper system. The git repository for the source is at http://gitorious.org/vial, please contact Jonpry so he can bring you up to speed on the project.

Thanks for all your hard work bringing sound (et al.) to Android Jonpry!

[Unknownforce]

Thanks for posting this, I was looking for a good write up on the reasoning for this. I didn't know where the DSP was buried and looking through lines of the current android code wasn't getting me anywhere...

I'm not particularly adept to reverse engineering, but I know a little bit, enough to try and take a look at it.

If you have any headway so far, I'd love to take a look at it and try to continue or make some more headway. shoot me a msg or something.

[x10guy]

+1. Probably the best explanation (concise) I've read about "sound" and Android. Thanks.

Let's hope this thread brings more "eyes" to the project to help solve it.

[jamice4u]

You are the man. I just sent you some thanks. Thank you for all your hard work try to get the sound working on android. 1 post and a 35 reputation impressive. It surprises me that 60 people read this post and only 8 people had the common courtesy to say thanks. This guy is doing his best to get the sound working!!!! We should show he more appreciation than this!!!! Only one more thanks for me to make 200 reputation.

[shaggylive]

Originally Posted by slacksoul

If you, or someone you know, has experience in reverse-engineering and decompiling as well as strong C++ and C skills, please take a look at the code. The only way we are going to crack this is by having more eyeballs look at the code.

well, I've never really considered myself a programmer, but I've got the flight time to certify..
I've gotten very good at reverse engineering/trouble shooting alot of R&D things over the years (like voip, before it was called that) which brings me to my unique skill set of working in both voice and data. I think alot of data type people have a hard time understanding phone logic.
I have been wanting to get into linux for many years, but never had much reason. till now... sounds like a mystery needs solved..
actually I've never more than looked over a sholder of someone working in c, so I don't even know what the good programming tools for it are, but am willing to give it a go if someone wants to hold my hand till I get walking.

[undarated]

As this is my first post to the site, I have to say THANKS!!!!! I'm constantly checking the site for any news or updates for android on the Rhodium. I wish you guys the best of luck and I can't wait til it's fully function.

[Unknownforce]

Originally Posted by shaggylive

well, I've never really considered myself a programmer, but I've got the flight time to certify..
I've gotten very good at reverse engineering/trouble shooting alot of R&D things over the years (like voip, before it was called that) which brings me to my unique skill set of working in both voice and data. I think alot of data type people have a hard time understanding phone logic.
I have been wanting to get into linux for many years, but never had much reason. till now... sounds like a mystery needs solved..
actually I've never more than looked over a sholder of someone working in c, so I don't even know what the good programming tools for it are, but am willing to give it a go if someone wants to hold my hand till I get walking.

just FYI, ARM ASM is much different than x86 reversing. I looked into it myself and it's much different.

[coolsilver]

My guess is either these are controllers that need firmware from the os to control and run the port for the speakers etc.

Much like most broadcom wifi drivers in linux need a firmware and driver (kernel module) to work. Like most of those you can see you have a wifi card but doesn't work. Firmware can also cause issues like WPA not to work. Most of these chips in the beginning used ndiswrapper which uses the windows firmware to work on linux. Though now there is a built from scratch driver which does have a firmware made for linux.


I also bet more needs built from scratch or hacked from windows to run on linux/android.

Anyway to make a wrapper and process calls through that dll? Is there any specific windows htc audio services that might be the link?

[ghost_deini]

Anybody find this file? i can see wapdrv.dll and waveapi.dll with total commander...
I've used IDA for x86 stuff and programmed in arm asm so i'd like to at least look in to it

[JDM_SOHC]

Thanks for the info, to bad it's not exactly GOOD news lol..