Hi all:
I'm looking for the driver we use to operate the camera under Android.
I was hoping to possibly check it over and redesign it for our Kaiser and possibly Polaris/Vogue machines as well.
I have experience in coding and development, just not much in the field of Linux these days. I worked with SCO Xenix back in the 90s' into 2000 (I know... old school), but not with much more recent Linux builds except Ubuntu and Slax.
If anyone has an ORIGINAL copy of the unmodified driver we use, please post to this thread. The only hold-up to having Android fully functional on our Kaisers is the camera/video issue and I'd like to try and fix it, you know, "do my part to help".
Thanx
PoX
PoXFreak said:
Hi all:
I'm looking for the driver we use to operate the camera under Android.
Click to expand...
Click to collapse
It's got 2 parts, the android driver and the kernel part. The android part is in /hardware/msm7k/VogueCamera if you pull froyo or gingerbread from git, or you can just clone the camera driver git on sourceforge here: http://androidhtc.git.sourceforge.net/git/gitweb.cgi?p=androidhtc/camera.git;a=summary
I haven't looked into the kernel part that hard, you will need to look in /arch/arm/mach-msm for the driver. I think a lot of it happens in the /audio/qdsp4 directory which is what handles most of the media stuff.
Have fun
Thanx scoot!
I think what might be happening is a memory allocation error when the camera is switched to video, either that or the driver is asking the camera to do something it can't, like capture video at 1280x1024.
If I get a chance tonight, I'll play with it on the laptop and try it by pushing the changes via C+P in root explorer.
Any clues as to what resolution the camera is supposed to record video?
PoXFreak said:
Thanx scoot!
I think what might be happening is a memory allocation error when the camera is switched to video, either that or the driver is asking the camera to do something it can't, like capture video at 1280x1024.
If I get a chance tonight, I'll play with it on the laptop and try it by pushing the changes via C+P in root explorer.
Any clues as to what resolution the camera is supposed to record video?
Click to expand...
Click to collapse
Uhm... our kaisers are probably using a hack. We have manual focosing with the wheel, perhaps the video is trying to auto-focus constantly and failing. I noticed that the FC is not right after switching to video, just a few seconds after... I'll try something and report
Do you know if camera is working for polaris? I mean, not video just camera
Thanks
daedric said:
Uhm... our kaisers are probably using a hack. We have manual focosing with the wheel, perhaps the video is trying to auto-focus constantly and failing. I noticed that the FC is not right after switching to video, just a few seconds after... I'll try something and report
Click to expand...
Click to collapse
I haven't had a chance to mess with it much but you may be right. A fixed-focus lens cannot have an auto-focus command sent to it because there's nothing there to recieve the command.
I have a processor upgrade to do on my old "Android-use-only" laptop (YAY, I got a Prescott!..lol), but I'll dig in and see where there might be a command for auto-focusing. Who knows, you might have stumbled upon the answer we needed 2 years ago...
Aren't you confused ?? Kaiser doesn't have a fixed-focus, it's a auto-focus lense.
Without any working Android phone to use as a model I have no point of referrence for what the video size should be, but I can only guess it's not supposed to be 1600x1200, or YUV420.
Daedric, I apologize... I was in a hurry and saw "manual focus" and just blubbered out whatever came to mind.
I did look into the commits made for the camera and all I saw was the size intergers being high. I would assume the video recorded with our camera would only fit the WinMo parameters for 640x480 (max)...
Must....dig.....deeper....
I'm on vacation at this time and although I brought a laptop, it's not my "development" machine. I'm also beginning to believe my theory about the video requiring some "cache" file to record ideo to before sending it to the "Video" folder created on the SD.
I'll dig in when I get back...
PoXFreak said:
I'm on vacation at this time and although I brought a laptop, it's not my "development" machine. I'm also beginning to believe my theory about the video requiring some "cache" file to record ideo to before sending it to the "Video" folder created on the SD.
I'll dig in when I get back...
Click to expand...
Click to collapse
We must first figure out how to auto-focus. The manual focus is intercepting wheel up-wheel down to move the lens. Now, as to this, i belive we don't have the focus range we had on winmo.
Second, after this is done, focusing on our cameras (i belive this is true to EVERY cellphone camera) works on a contrast basis. You probably noticed that it's very difficul to focus on a clean, white, bright wall. This must be implemented in software, either at driver level, or some other way...
The best course of action will probably be to dig how other low-spec/initial android phones are working, and try to replicate. Perhaps HTC Magic's camera...
Before vacation I was looking into the kernel drivers from the Dream and Sapphire, but since these have slightly different hardware, I'm afraid to start playing around in kernel drivers, especially where hardware control is concerned.
If this was MS-DOS/Windows compiled, I would be in heaven and it would be so much easier.
Any news?
Do you have any good news?
Related
Don't get any of your hopes up, CorePlayer will solve nothing for us:
First of all, what they "fix" will only work inside Coreplayer
And second:
http://www.corecodec.com/forums/index.php?topic=596.15
page 2, second post, blue one:
"ok... so let me clarify... so its a little of BOTH to be true in that... we _ARE_ using HW overlay with the qtv drivers for the color conversion and bilinear zooming and we are _NOT_ using the builtin mpeg4 hw decoder."
So they only fix the problem which made our kaiser slower (about 3 times slower) than all similar devices when using DirectDraw... you know the issue, we were going around it by using Raw Framebuffer in Tcpmp
What this means: they found a workaround for a bug in the Kaiser, they do not have video drivers, they do not have 3D drivers.
I wish it wasn't so... but it is
And this needs a new thread because?
This has been covered in the driver development thread, and the CorePlayer thread. Yes, we know it's not the mythical new driver set that will bring about world peace and cure famine, but for anyone who wants to watch decent video on their Kaiser it's a solution for now.
Sorry if i'm bloating the forum, but i keep seeing threads on the first page about this, and was hoping to put an end to it all
Thank you for your clarification! This clears up the situation
I was really hoping they would be accessing the hw acceleration of decoding as well :/
Well, I just wanted to ask something very odd.
In my HTC Tytn I (Hermes) using black satin rom,
coreplayer 1.2 plays videos with very slow performqnce whatever
the settings, but when I try benchmarking it plays smoothly and gives
125% using rawbuffer.
I ve reached to a point here i will watch the movies in benchmark mode...
Redundant Thread.
And with a mighty
[Thwomp]
The Colossal Hammer of Closure
ends the life of yet another redundant thread...
Video recording app now in the market, tested and works ok overall but no sound. To stop recording hit back button. Very buggy!!
watts da name of it
yeah it's working well.
very, very buggy mind you, alot of force closing but, hey, it works i suppose.
quality's not too bad but it's really slow, so as long as you keep still it works fine
i'll get a little video uploaded to show you.
illrific said:
watts da name of it
Click to expand...
Click to collapse
Video Recorder/Video Camera
can anyone find where it is on the sdcard and here is more info[URL="]http://www.android7.org/forum/viewtopic.php?f=10&t=5[/URL]
santaninja said:
can anyone find where it is on the sdcard and here is more infohttp://www.android7.org/forum/viewtopic.php?f=10&t=5
Click to expand...
Click to collapse
Been trying to find it myself, all it has in the program file is a bunch of blank pics! Man its a start but buggy as hell!!!
on the website it says there's gonna be an update in the next 2-5 days so lets hope the bugs get ironed out. also i'd quite like to see more stable recording and sound
but maybe i'm asking too much
Interesting. Pity about the bugs and no sound, but it's a start, which is very good.
Good start. Bring on the updates.
Keeps asking to update! very annoyin
i saw this app today, for some reason decided not to download it
but its a "good" start
hope to see this being updated constantly.
How could u not even try it out .
I grabbed it as soon as physically possible.
It needs a lot of work.
And yes, MJPEG
Really needs some work haha.
EDIT: I just realized that my post has a lot of smilies
Update Released!
New update with a much cleaner interface. The videos are stored(i use this loosely) onto the sd card in /.org.android7/videorecorder. Uses about 16m for 10 seconds of video. The audio works as long as you also have the VoiceClient installed, it's not very loud but it's there.
After opening the app press record to open the recorder window(your not recording yet) hit menu-record to record and the same to stop recording.
I'm curious how the recording is working? It would be nice to at least be able to re-assemble the recordings on my machine.
In the video folder there are two file types blah.3gp [16kb] (appears to only be the audio) and a group of blah.jpg [~225kb] images (tho these are not viewable even on my pc?) any ideas?
The videos it makes are not viewable on any of the other viewers on android either. (the only thing it's really missing)
Well, it seems to get better with updates.
2 things i realized...
1) When you replay a video after you record it, it plays like its on fast-foward or 2x.
2) If you delete a video from your phone, the video gets deleted(Why is 30 seconds of recording 32 Megs by the way?) BUT the audio files from the video(which the file type is .jpg also, weird?)stays in the anroid7.org folder so they never really delete, you have to manually go in there and delete them when you USB to your comp.
Either way, i consider the effort and applaud him for the first video recording app of its kind on the android OS, i see potential in the future.
built-in video player?
I'm confused.. I know g1 doesnt have a built-in icon for video player but i just noticed that i can open a video file using an astro file manager (unless astro has a video player itself).. it has an auto-rotate function too and looks like the Video Player in the Market.. hhmm.. And currently, I dont have any video player program installed in my G1... Can someone clarify this?
Ya, I've tried rebuilding the 'video' into something useful. The Audio file name is xxxxx.jpg.3gp (amr i believe) and the images 225k each ~2/sec don't appear to actually be .jpg, nor can i get them open with .gif .bmp .jpeg .png as extensions
this is on windows, maybe they'll turn up something more promising on *nix.
I quit testing this.
It's not encoding nothing only take fast pics and save de audio on 3gp format.
It's good to see that we can take pcitures so fast because this show us the real posibilities with this cam on HTC G1.
Google has to build something using the internal h264 posibilities of decode/encode(i guess) to encode the pics and audio on real time(too heavy processing ?)
Till that time....just wait and see
It CAN.
WinMo guys did.
Prove me wrong.
HTC Touch Diamond T-mobile G1
Main specs:
Qualcomm MSM7201 CPU Qualcomm MSM7201 CPU
192 MB of RAM 192 MB of RAM
3.2 MP Camera, no flash 3.2 MP Camera, no flash
OS: WinMo OS: Google Android
Videorecording:
352x288x30fps None.
So:
1) It is definately technically possible.
2) Hey, WinMo does Android, the open source, WTF???
Where's Our Video Player?
Does anyone remember TCPMP / Core Player on WinMo?
Why can't we have that satisfaction on our G1's???
Is it the hardware-decoder thing? Is it OTW?
Anyone know what's up?
There is a video player, It's called Video Player....
Is there any way at all to do a fix for either the youtube client or the X1's display drivers such that it display the Youtube vid correctly?
I am using Wild Child's latest rom and for some strange reason when I play the tube file the first time in landscape mode a white line is in the button of the screen and in the left hand side of the screen, but then I play the same file the second time those stripes are gone?
This happen with every second youtube file
Is it a hardware or software issue??
As far as I can tell this has been there in all of the youtube applications with various drivers, the new manila rom doesnt use r3 drivers...and its still there...
WC
to be honest, i don't consider this a bug at all. so there's a thin white line - so what? what bug's me more is the fact that the videos still have to be fully buffered before playing or you'll have video-lag. (happens via 3g and wlan with every rom i've tried so far, not via activesync)
if i remember correctly this had to something with drivers too, but i didn't follow the corresponding threads lately. any news on that?
caliban2 said:
to be honest, i don't consider this a bug at all. so there's a thin white line - so what? what bug's me more is the fact that the videos still have to be fully buffered before playing or you'll have video-lag. (happens via 3g and wlan with every rom i've tried so far, not via activesync)
if i remember correctly this had to something with drivers too, but i didn't follow the corresponding threads lately. any news on that?
Click to expand...
Click to collapse
It did not happen with R1A, only with subsequent R2/R3 ROMS and I'm assuming custom ROMs based off of it.
I also really wish someone with the expertise would step up and help us out with this one.
The reason why it is buffered is to somewhat avoid the lag in video, this was dne on purpose...I mean even on your computer when accessing some sites to view video's through say windows media player it will buffer the video... As I stated in my earlier post...I have tried various drivers and that software bug is still there.......
WC
you ever noticed, how video-playback is lagging, but audio is normal while buffering? that's what i meant. hit "pause" and let it load the full video first, everything will be normal.
so no, that's not like on the computer...
caliban2 said:
you ever noticed, how video-playback is lagging, but audio is normal while buffering? that's what i meant. hit "pause" and let it load the full video first, everything will be normal.
so no, that's not like on the computer...
Click to expand...
Click to collapse
I may be wrong but if you try on an old computer with a 528MHz processor and the same ram size, I think it will be laggy too...
Mopral said:
I may be wrong but if you try on an old computer with a 528MHz processor and the same ram size, I think it will be laggy too...
Click to expand...
Click to collapse
depends on what OS this pc has to run... and yes, an old pentium-m underclocked to 600 mhz plays them fine.
anyhow, look above: it didn't happen with the first firmware. it doesn't happen if it's connected via active sync. it obviously doesn't happen on other phones (the iphone has an even slower cpu), etc etc...
but that wasn't the point.
so, no news in that direction.
It´s a driver issue, install this and YouTube will be fine in HQ without full buffering.
YYYYY said:
It´s a driver issue, install this and YouTube will be fine in HQ without full buffering.
Click to expand...
Click to collapse
I have just try it: black screen when playing video with the youtube app...
Than try this, it´s a little bit difficult because every cooker has different driver combinations, try this if you want.
YYYYY said:
Than try this, it´s a little bit difficult because every cooker has different driver combinations, try this if you want.
Click to expand...
Click to collapse
Black screen too!
Maybe because I use Agent47's rhodium drivers and Neos2007 OpenVG...
Ah ok, I have no idea why it is not working, as I said, it has sth. to do with the driver combinations the cookers are using, maybe you should try it without the openVG drivers. Just try
Just for the hell of it i added youtube to the r1 rom....this issue of the white line around the bottom of the screen and running up the left side is a software issue....not a driver issue...
WC
Wild Child said:
Just for the hell of it i added youtube to the r1 rom....this issue of the white line around the bottom of the screen and running up the left side is a software issue....not a driver issue...
WC
Click to expand...
Click to collapse
I am not a software engineer or a programmer, but to me its like the client is made to talk to a set of display drivers which aren't 100% compatible with the X1s displaydrivers. Thereby causing it not display fullscreen video in youtube client correctly always?
if the youtube client had ben open source would it have possible to take the base code apart and fix the problem?
just for fun i've installed the official cab from youtube - no white stripes, but lagging as well while buffering...
Wild Child said:
Just for the hell of it i added youtube to the r1 rom....this issue of the white line around the bottom of the screen and running up the left side is a software issue....not a driver issue...
WC
Click to expand...
Click to collapse
Reverting to the R1A ROM fixes all issues with youtube constantly buffering and not playing back HQ video smoothly, it does not fix the issue with the white line. Sorry for the confusion.
I think the buffering issue and not playing back HQ videos smoothly is a much bigger issue than the white line.
YYYYY said:
It´s a driver issue, install this and YouTube will be fine in HQ without full buffering.
Click to expand...
Click to collapse
Can you reupload the fixes you had posted? I would like to try and see if they work on my R2A X1a ROM
Sure, please report if it´s working or not, which Youtube version are u using?
-click-
never noticed this before:
lag while buffering obviously depends on the video being played. there's non of it while playing a cartoon (try "southpark"). play anything else - lag until fully buffered. so either the initial drivers had some sort of hw-acceleration (don't believe that) or the newer ones are less efficient and therefore need more cpu-ressources?
the "pc vs.mac" videos are playing almost smooth, only minor hickups there -> white background without details.
@yyyyy:
are these fixes of your's uninstallable? i'm on itje's v10.2 and it's running perfectly, i don't want to risk a hardreset right now...
When accessing the browser version of youtube on cm7, I can play 320p and below videos fine, but if I get a 480p video it stutters. When I try to press the button to change the resolution it pops up and disappears almost instantly. Is there a way to set it to 320p as default or a way to get this button to behave? I know dsp is probably a ways off so I'd like to get this working. Thanks.
In case anyone is interested, theres a setting in the youtube account options, but it only lets you set it to the lowest quality.
Go to www.youtube.com/xl . it's a much better version of the site for tablets. You can go from SD to HD with one click. Hope this will work for you while we all wait for hardware acceleration, so we can use the proper youtube app.
Thanks, Why was it possible to have hardware accel in Froyo but not GB? Is it that different?
chadamir said:
Thanks, Why was it possible to have hardware accel in Froyo but not GB? Is it that different?
Click to expand...
Click to collapse
If I remember right, it's a driver issue. Like alot of things for the nook, there were never proper drivers put out for froyo and gingerbread. most of them had to be backported I think. Froyo does have hardware acceleration, but I think it's a hacked up driver. That's no knock to the people who made it, I have a great amount of respect for the devs work.
Hey guys,
Long time lurker, but just wanted to share this as a quick fix for people struggling with the bluetooth streaming audio quality issues that Google still hasn't fixed in 4.3. I've been reading through one of the open issues at Google:
http://code.google.com/p/android/issues/detail?id=39632
and found in post #157 that someone noticed this commit elsewhere that supposedly fixed the low quality issue:
https://github.com/djp952/android-p...mmit/2ee778598e3b2e3f826e3f6b4eeafbb11005b21b
All credit to the people who found all this information in the first place, but I've been so tired of this stupid quality thing that I downloaded the source for 4.3 and tried to toss this one line in there to see what would happen. Long story short, I built the JSS15J branch for mako and it seems to have -vastly- improved the quality (back up to where CM was to my ears, at least).
Afaik, all you have to do is backup and replace these two files living in system/lib/hw and set permissions to 644. Then reboot and all should be well - with the added benefit of AVRCP 1.3 working too now.
Please don't be to harsh if I've forgotten something crucial... it's my first time sharing anything at all
UPDATE:
I've uploaded a different set of files built with DEFAULT_SBC_BITRATE=345 (mako-JSS15J-345-btaudio.zip). The original was set with DEFAULT_SBC_BITRATE=328 (mako-4.3-JSS15J-btaudio.zip). Those of you who were having issues with stuttering...maybe this will fix it? I have no idea...not a real developer But give it a shot!
UPDATE #2:
From the sounds of things, these files are only compatible with AOSP compiled roms - no one's gotten it to work with CM10.2 yet, I believe. Just fair warning if it doesn't work for ya.
Cbaronne, this particular issue has been such a pain as I'm a heavy music listener in the car so I'm excited to see this.
However, any idea if this work on JWR66V?
As I understand it, the difference between the two are mainly wifi related so I'm willing to guess that those two files would be the same.
Your thoughts?
AdamUpNorth said:
Cbaronne, this particular issue has been such a pain as I'm a heavy music listener in the car so I'm excited to see this.
However, any idea if this work on JWR66V?
As I understand it, the difference between the two are mainly wifi related so I'm willing to guess that those two files would be the same.
Your thoughts?
Click to expand...
Click to collapse
I originally compiled JSS15S because I had aspirations of running a clean new build to test the WiFi stuff, but then just took those two files out to try 'em. I'm on stock 66V rooted and they work just fine for me ^^
You Da Man.
Trying this right now.
UPDATE: Works like a charm. No more nasty hissing/crackling of the higher frequencies. Sounds great. Thank you sir, you just made my week.
I'm using the same files extracted from GNex factory image and is working great. However I think I may be missing BLE as GNex doesn't support it. I will give this a try and report back.
Thanks for this. I upgraded to 66V from AOKP the other day & have been seriously considering going back due to the horrible highs (listen to some dubstep and you won't be able to focus on anything else.)
Here's a flashable zip I made. Don't mind the updater script I just used one from another zip I had I my phone that mounted
http://db.tt/LZPJNzjV
Sent from my Nexus 4 using xda app-developers app
getting quite a few glitches using this. audio seems to randomly stutter like its being stretched; may have something to do with cpu speed scaling not sure yet.
quality does sound better tho with this.
That happened to me too initially when I was playing with GNex files. Try repairing the device.
---------- Post added at 12:03 PM ---------- Previous post was at 11:32 AM ----------
Just tested this. Working great.
Does AVRCP 1.3 really work? For me, it looks like this in the image appended ... trackname + album working fine, but band does not
I get track number, time played, artist, track name and album name on a cheap sony car stereo unit.
As for sound quality I don't really have a reference point so haven't noticed it being all that bad. If I wanted to flash the attached zip how easy is it? The limit of my experience flashing was a stock 4.3 using this tutorial http://forum.xda-developers.com/showthread.php?t=2010312
Thanks for supplying this patch. It works excellently
I've noticed though that last.fm scrobbler can't 'see' what Play Music is playing anymore though, reckon the patched file breaks that connectivity?
noobdeagle said:
getting quite a few glitches using this. audio seems to randomly stutter like its being stretched; may have something to do with cpu speed scaling not sure yet.
quality does sound better tho with this.
Click to expand...
Click to collapse
I've never personally experienced the stutter problem, but afaik, that's an ongoing issue that other people have reported (and perhaps why Google still hasn't made this small change). I know it isn't perfect yet, so while waiting for CM10.2 I'm keeping my eyes open for other tweaks to that file
muff99 said:
Does AVRCP 1.3 really work? For me, it looks like this in the image appended ... trackname + album working fine, but band does not
Click to expand...
Click to collapse
I've only ever used this fix with Play Music so far, and it all seems to be working (at least artist/title/album). I don't have a lot of experience with the new Bluetooth stack outside of that x.x
Thanks for the fix! The audio quality is much improved, and the AVRCP stuff still works completely. I was going to make a CWM zip of this, but someone earlier in the thread beat me to it.
Thanks!
Edit: I'm using Google Play Music, which seems to work very well with AVRCP. I can't speak for other players.
cbaronne said:
I've never personally experienced the stutter problem, but afaik, that's an ongoing issue that other people have reported (and perhaps why Google still hasn't made this small change). I know it isn't perfect yet, so while waiting for CM10.2 I'm keeping my eyes open for other tweaks to that file
Click to expand...
Click to collapse
Darn. I'm getting the occasional shutter as well. Although, I'd much rather deal with the odd stutter than terrible audio 100% of the time.
Looks like google is having the same issues. See post #182 https://code.google.com/p/android/issues/detail?id=39632#c182
cbaronne, could I bother you to create another set of files using the max bitrate of 345 for 48Khz? After reading the comments on the Google code site and from here (http://soundexpert.org/news/-/blogs/bluetooth-audio-quality-a2dp), I'm curious to see what happens.
AdamUpNorth said:
Darn. I'm getting the occasional shutter as well. Although, I'd much rather deal with the odd stutter than terrible audio 100% of the time.
Looks like google is having the same issues. See post #182 https://code.google.com/p/android/issues/detail?id=39632#c182
cbaronne, could I bother you to create another set of files using the max bitrate of 345 for 48Khz? After reading the comments on the Google code site and from here (http://soundexpert.org/news/-/blogs/bluetooth-audio-quality-a2dp), I'm curious to see what happens.
Click to expand...
Click to collapse
Sure thing - now I'm curious too. Changed that value to 345 and kicked off the build... will have something to post when I get home from work!
Edited my post with a more friendly link for the flashable zip.
Sent from my Nexus 4 using xda app-developers app
I just ended building JSS15J just to get this fixed and found this post lol... an incredibly annoying bug and an easy workaround. Thanks for posting it.
muff99 said:
Does AVRCP 1.3 really work? For me, it looks like this in the image appended ... trackname + album working fine, but band does not
Click to expand...
Click to collapse
If I use Pandora I have the same issue:
Artist shows up in the song name field on my car radio
Album shows up correctly
Song Title is missing
If I use Google Play Music everything shows up correctly.