Steve,<br><br>When you say: <br><br>1) There is a problem related to specific firmware version of newer
(USB Video Class compliant) Logitech cameras -- the symptom is the
camera becoming wedged when streaming at certain resolutions.<br><br><br>What are those resolutions so we can avoid them?<br><br><br>Paul <br><br><br><br><br><div class="gmail_quote">On Tue, May 13, 2008 at 2:38 PM, Steve Sisak <<a href="mailto:steve.sisak@ioxperts.com">steve.sisak@ioxperts.com</a>> wrote:<br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Available here:<br>
<br>
<<a href="http://www.ioxperts.com/downloads/Video/Beta/1.1.2b7/" target="_blank">http://www.ioxperts.com/downloads/Video/Beta/1.1.2b7/</a>><br>
<<a href="ftp://ftp.ioxperts.com/IOXperts/Video/Beta/1.1.2b7/" target="_blank">ftp://ftp.ioxperts.com/IOXperts/Video/Beta/1.1.2b7/</a>><br>
<br>
This release expires July 1, 2008<br>
<br>
This release is primarily a checkpoint and regression test for an internal cleanup of the FireWire code in preparation to make another try at getting it working on Intel.<br>
<br>
At the moment, as far as I can tell, FireWire now works under Rosetta on Intel machines under both Tiger and Leopard, but isn't working in Leopard on my G5 Quad, for reasons I can't determine -- including wether the problem is isolated to my machine.<br>
<br>
(I had the entire FireWire bus on that machine vanish under Lepoard, although it was working if I rebooted to Tiger -- it came back after a clean install of Leopard this morning, but I'm not sure what to trust at the moment)<br>
<br>
Also the errors I'm now getting w/r/t FireWire native on Intel lead me to believe that I may be close to having it working (I'm getting what look like byte-swapping bugs rather than a hang stopping the video stream) -- not committing to anything yet, but it's looking encouraging.<br>
<br>
I'd appreciate any help in determining:<br>
<br>
1) That nothing works worse than the previous release (that I didn't introduce any new bugs in the cleanup)<br>
<br>
2) Where FireWire support works and where it doesn't at the moment. (It should be working everywhere but native Intel, including under Rosetta, but might be having an issue on Leopard+PPC -- I'd like a matrix to try and track this down)<br>
<br>
Also I bumped the expiration date out another month, just to be safe.<br>
<br>
Thanks again for your help,<br>
<br>
-Steve<br>
<br>
------------<br>
<br>
Know issues:<br>
<br>
1) There is a problem related to specific firmware version of newer (USB Video Class compliant) Logitech cameras -- the symptom is the camera becoming wedged when streaming at certain resolutions.<br>
<br>
2) There is a bug that sometimes creates strangely named files at the root of your hard drive -- I can not reproduce the problem here, but they appear to be harmless -- still looking for the cause of this.<br>
<br>
----- changes in 1.1.2b6 -----<br>
<br>
bump expiration date (no substantial changes)<br>
<br>
----- changes in 1.1.2b5 -----<br>
<br>
Fix bug where screen image ends up as two blue squares over some noise, like this:<br>
<br>
<<a href="http://www.laservall.com/%7Efrancesco.rolando/screenshot_04.jpg" target="_blank">http://www.laservall.com/~francesco.rolando/screenshot_04.jpg</a>><br>
<br>
when video is resized from 320x240 to 640x480.<br>
<br>
It may also fix the other reported crash in the Apple RAW codec, where the top of the stack of the crashed thread looks like this:<br>
<br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
Thread 0 Crashed:<br>
0 ...ickTimeComponents.component 0x99cffe1c Copy32ARGB32ARGBScale + 156<br>
1 ...ickTimeComponents.component 0x99cfcc88 RAW_CDDrawBand + 2104<br>
2 ...ple.CoreServices.CarbonCore 0x90bdf030 CallComponentFunctionCommon + 1044<br>
3 ...ple.CoreServices.CarbonCore 0x90bdebb8 CallComponent + 260<br>
4 ...ickTimeComponents.component 0x9a7826e8 ImageCodecDrawBand + 40<br>
5 ...ickTimeComponents.component 0x99c37cec BaseCodec_DecompressWorkFunction + 1324<br>
<snip><br>
</blockquote>
<br>
Improvements to internal resource loading.<br>
<br>
----- changes in 1.1.2b4 -----<br>
<br>
Bump expiration date, improved logging.<br>
<br>
----- changes in 1.1.2b3 -----<br>
<br>
Don't attempt to cache preferences at static init time -- causing coreaudiod to crash<br>
<br>
----- changes in 1.1.2b2 -----<br>
<br>
Change Isoc stream logic to be tolerant of Leopard's high USB call latency:<br>
<br>
Don't attempt to calculate if we're close to the current time: try a block of frames and see if we get a TooOld error<br>
<br>
----- changes in 1.1.2b1 -----<br>
<br>
Bump expiration date to 4/1<br>
<br>
----- changes in 1.1.2a47 -----<br>
<br>
Fixed the double-install bug -- you should now be able to install over any version of the driver on any OS version.<br>
<br>
----- changes in 1.1.2a46 -----<br>
<br>
The new serial number scheme is in place for all drivers (USB Webcam, FireWire Webcam, FireWire Industrial) so that new purchases will have a new (1.2) serial number.<br>
<br>
Old serial numbers are now properly recognized and supported until we have an automatic upgrade system in place. (Fixed a bug in a46)<br>
<br>
----- changes in 1.1.2a45 -----<br>
<br>
I think we have the new serial number scheme is in place and new purchase with get a new (1.2) serial number -- this release should work with a 1.1 serial number and 1.2 serial numbers should work with any previous driver.<br>
<br>
I also may have fixed a crash where QuickTime was crashing if fed a 0-length MJPEG buffer.<br>
<br>
----- changes in 1.1.2a44 -----<br>
<br>
Bump expiration date, fix web page.<br>
<br>
<<a href="http://www.ioxperts.com/products/webcamx_1.2.html" target="_blank">http://www.ioxperts.com/products/webcamx_1.2.html</a>><br>
<br>
----- changes in 1.1.2a43 -----<br>
<br>
This release is mostly to bump the expiration date before 1.1.2a42 expires<br>
<br>
I'm still working on the purchase and registration code, but made a number of improvements to the logging code so that it won't log anything (and fill up your hard drive) if you leave an expired demo running after the 30 minute limit.<br>
<br>
Also, this version is built with Xcode 2.5 (rather than 2.4.1) so that I can debug on both Tiger and Leopard.<br>
<br>
As always, please report any problems.<br>
<br>
Also, we've added updated compatibility notes on our website for Leopard and various previous operating system <<a href="http://www.ioxperts.com/leopard.html" target="_blank">http://www.ioxperts.com/leopard.html</a>> -- please let me know if you see anything incorrect.<br>
<br>
(We're still cleaning up the pages, but figured is was better to have something close to correct posted rather than waiting longer)<br>
<br>
Happy New Year,<br>
<br>
-Steve<br>
<br>
----- changes in 1.1.2a42 -----<br>
<br>
This is also primarily a regression test release as I've been cleaning up the purchase and registration code in preparation for doing a paid update.<br>
<br>
(If you're currently testing a demo version and are about to purchase a copy, please contact me off list as I could use some help verifying the process once it's ready)<br>
<br>
Also, I bumped the demo expiration date to 12/31 as the current betas expire in in 3 days on 12/15. Please speak up if there are any problems so they can be fixed before that happens.<br>
<br>
Known issues:<br>
<br>
1) Installing over 1.1.2a38 or later will give "no cameras" -- reboot and reinstall to work-around (2nd install will work).<br>
<br>
----- changes in 1.1.2a41 -----<br>
<br>
<br>
This is primarily a regression test and should be functionally identical to 1.1.2a39 except that I've done a major cleanup of the core logging code to allow better debugging of application initialization and IPC.<br>
<br>
I'd appreciate it if you could verify that no new bugs are introduced relative to a39 while I add code that takes advantage to the new functionaity.<br>
<br>
----- changes in 1.1.2a39 -----<br>
<br>
Fixed a crash in the logging version of the driver for the divio chipset (passing a class instead of POD type to printf())<br>
<br>
----- changes in 1.1.2a38 -----<br>
<br>
Change ioxdeviced from StartupItem to launch-on-demand launchd daemon.<br>
<br>
This fixes the problem with ioxdeviced quitting, resulting in no cameras being visible -- it also has a bit more logging in case there is still a problem.<br>
<br>
My suspicion is that the daemon was being loaded to soon and either getting in contention with one of the Apple drivers for the hardware or running before something was ready.<br>
<br>
Under Mac OS 10.4 and later, it is now registered with launchd and loaded only when the video digitizer is accessed by QuickTime, which will be much later.<br>
<br>
It will also be relaunched automatically if it falls over.<br>
<br>
One thing I did notice is that there may be an installer problem if you install 1.1.2a38 over itself -- in this case no cameras will appear in Camera Identifier. Please let me know if you see this problem, but you can work around it by rebooting and installing again.<br>
<br>
----- changes in 1.1.2a35,36,37 -----<br>
<br>
bump expiration date.<br>
<br>
----- changes in 1.1.2a34 -----<br>
<br>
This is functionally equivalent to a33, except that I've updated to the latest eSellerate SDK and added support for a few new Logitech cameras.<br>
<br>
----- changes in 1.1.2a34 -----<br>
<br>
Update to latest eSellerate SDK<br>
<br>
Support for new Logitech cameras<br>
<br>
----- changes in 1.1.2a33 -----<br>
<br>
Changes to Isochronous bandwidth allocation code and Isoc stream startup code to prevent stream from restarting when stopping<br>
<br>
----- changes in 1.1.2a29 -----<br>
<br>
Fix to correct Gamma on newer Logitech cameras<br>
<br>
Possible fix (and additional logging to detect) the Skype crash<br>
<br>
Prefer YUY2 over MJPEG in UVC video mode selection<br>
<br>
----- changes in 1.1.2a28 -----<br>
<br>
1.1.2a28 extends the expiration date and includes a few minor performance improvements.<br>
<br>
I've also added some additional logging to try and isolate the problem with the Philips SPC900NC not being detected after restart.<br>
<br>
----- changes in 1.1.2a27 -----<br>
<br>
Enabled Intel support for VLSI Vision CPiA-based Cameras<br>
<br>
Use physically contiguous Isoch buffers on CPiA and Philips-based cameras.<br>
<br>
Allow Isoch pipe reset when stopping stream -- should speed start/stop<br>
<br>
----- changes in 1.1.2a26 -----<br>
<br>
Added Intel support for cameras based on the Divio NW801 and NW802 chipsets.<br>
<br>
Also fixed a few byte swapping errors in common code that may have caused a few isolated problems.<br>
<br>
----- changes in 1.1.2a21 -----<br>
<br>
If you're having problems with crashes in Skype, this build _may_ help.<br>
<br>
The primary change is that it holds off on some initialization of the video pipeline until the as late as possible -- this may provide a slight performance increase as well.<br>
<br>
Hopefully there will be a matching Skype update to address the threading issues we've been seeing.<br>
<br>
The cause of the registration dialog popping up multiple times in Skype is that Skype is opening and closing the digitizer multiple times at startup -- this is not a problem, but somewhat annoying -- I'm looking at a fix to make the dialog appear only once per application launch.<br>
<br>
----- changes in 1.1.2a20 -----<br>
<br>
1.1.2a20 adds support for Face Tracking on Intel-based machines.<br>
<br>
In the process, we've done major work on the software pan/tilt/zoom code which should result in improved performance, but should be tested thoroughly for compatibility as there are significant differences in the math.<br>
<br>
----- changes in 1.1.2a18 -----<br>
<br>
This is primarily to bump the expiration date as 1.1.2a17 expired today, but we may have fixed a problem in the camera registration code that could cause a camera to not be recognized under certain circumstances.<br>
<br>
For anyone who has one of the newer Logitech cameras and is having problems with the audio, drop me a note off-list -- we have an audio driver that works aroud the bug in the camera about ready for testing.<br>
<br>
----- changes in 1.1.2a17 -----<br>
<br>
This should fix the problems a few of you reported with the QuickCam Pro 5000, Fusion, and (Black) QuickCam for Notebooks Pro.<br>
<br>
The issue was that the firmware in those cameras returns incorrect data for some of the controls related to the snapshot function, and the driver caught and reported the error, unfortunately causing it to fail on load with those cameras.<br>
<br>
I've added code in the driver to work around the bug -- please versify that the problem is fixed and that the fix didn't cause issues with any other cameras.<br>
<br>
There is still a known issue that the snapshot button on the new Logitech cameras in not working -- I was hoping to get it fixed, but didn't want to hold up getting something out with the previous version expired.<br>
<br>
---------<br>
<br>
As always, please post any problems to the list.<br>
<br>
-- <br>
_________________________________________________________________________<br><font color="#888888">
Steve Sisak, CTO <a href="mailto:steve.sisak@ioxperts.com" target="_blank">steve.sisak@ioxperts.com</a><br>
IOXperts, Inc. voice: +1 617 876-2572<br>
87 Bristol St #3A fax: +1 617 876-2337<br>
Cambridge, MA 02139 mobile: +1 617 388-6476<br>
_______________________________________________<br>
Video-Beta-Discuss mailing list<br>
<a href="mailto:Video-Beta-Discuss@mailman.ioxperts.com" target="_blank">Video-Beta-Discuss@mailman.ioxperts.com</a><br>
<a href="http://mailman.ioxperts.com/mailman/listinfo/video-beta-discuss" target="_blank">http://mailman.ioxperts.com/mailman/listinfo/video-beta-discuss</a><br>
</font></blockquote></div><br>