release.txt Driver File Contents (dm6437_1day_1_21.zip)

*******************************************************************************
* DM6437 Demo Application Release Notes
*
* Texas Instruments Incorporated
* Copyright (c) 2007, All Rights Reserved
*******************************************************************************

-------------------------------------------------------------------------------
DM6437 Demo 1.30.00

-------------------------------------------------------------------------------
Disclaimers:

This is demo code only and although every effort has been made, TI makes no
claims pertaining to performance nor robustness of the demo. The demo should
serve as a reference to see how all the various pieces of the DVSDK are
integrated together into a single coherent application. The demo is designed
to be somewhat generic and flexible and therefore may not be implemented in
the most optimal way. Also, the demo implementation is complex, so there are 
more simplified code samples in the DVSDK examples directory.

The demo code has not been through rigorous testing and hence is not
considered to be production code. It's quite possible undiscovered bugs exist 
in the demo. TI will make every effort to fix bugs for future releases.

The demo is not guaranteed to decode all H.264/G.711 clips that the user
obtains from outside sources, as testing in this area has been minimal.

TI is not positioning this demo code as production software nor does TI plan
to make a product out of the demo code. It is here purely for example
purposes.


-------------------------------------------------------------------------------
Known Issues & Things to Take Note of:

- When the host app is connected over Ethernet, it can introduce extra CPU
  load because of network latencies. Sometimes the CPU load may spike up to 16%
  momentarily and then come back down. We suspect this is due to the two-way 
  handshaking that happens between the host and target once per second when 
  the host makes a series of RPC calls to query statistics from the target. If 
  you disconnect the host application, the CPU load does indeed go down to a 
  steady expected value.  This does not impact audio/video performance, since 
  the RPC is at a lower priority. This is still being investigated and 
  improvements might be made in a future release.

- When using the IP discovery feature of the host application, the host
  machine and the EVM are required to be on the same subnet. One common 
  situation to watch out for is when a laptop is on a wireless network and the 
  EVM is plugged into a wired LAN. In general this won't work because in most 
  cases the wireless network will be on a different subnet than the wired LAN. 
  You can still manually enter the IP address into the host app and it will 
  work, but you can't use the IP discovery feature. In this case, you will 
  have to run the code from CCS to see the IP address that gets printed to the 
  CCS console. 

  There is one other IP discovery issue to watch out for. A laptop with a
  wireless network connection that is also docked and connected to a wired LAN 
  may have both connections active at once, and the laptop will actually have 
  two IP addresses, each of which on a different subnet. This will prevent the 
  IP discovery protocol from working.  The solution is to disable the wireless 
  network while docked.

- The CCS project now has multiple 'Configurations': Debug, Release, iDebug,
  and iRelease. The difference between debug and release is that the debug
  configuration is built with -o3 compiler optimization. The 'i' configurations
  use instrumented versions of the PSP drivers.

- There is no other demo application documentation. However, you can find a
  high level PowerPoint presentation in the demo doc folder.

- The hardware resizer is used to both de-interlace the input and also to
  scale the video input from D1 to CIF. For optimal video quality, D1
  video is set to 704x480 for NTSC and 704x576 for PAL instead of 720x480
  and 720x576. This is because the CIF resolution of 352x240 NTSC 
  (352x288 PAL) is exactly half of 704 lines horizontally and yields the 
  best quality.

- When running the host app in encode to file or decoding from file mode, the
  demo is not performing streaming, but rather using file IO fread/fwrite
  operations.  The demo has implemented a simple file IO over TCP/IP sockets 
  system that allows the target to open/close/read/write host files.

- In some instances, depending on the video source equipment used (DVD,
  camera, NTSC, PAL, etc.), you might observe one or more black lines at the 
  edges of of the video input.

- Occasionally, audio may output a brief moment of noise. At other times you
  may hear "ticking" artifacts. Many times if you hit "stop" and then "play" 
  again, the noise artifacts will disappear

- Many times when you first start the demo with audio enabled, you will hear
  some initial noise artifacts.

- You may observe that when the audio produces noise, that the synchronization
  between audio and video is disrupted.

- Certain instabilities have been observed, mainly with using the host
  application to start and stop the demo.

- For best performance, remember to "Run Free" from CCS as opposed to just
  "Run".

- "Disconnect" the host application to eliminate any rate loss due to network
  latencies. This of course won't be possible when encoding to a file or 
  decoding from a file, since the host application is required for those modes.

- Occasionally, initiating a play or stop from the host app will fail to work.
  Usually repeating the action again after the error dialog box appears seems 
  to work.  Also, occasionally clicking on the Play button from the host 
  application will not work at all.  Usually trying it a second time
  will work.  This is being investigated.

- It has been observed that attempting to encode to file over a slow network
  connection (such as a wireless connection), will cause the demo application 
  to crash.

- When you decode from file, it can take a moment before you see the output.
  This is because the application has to initially read from the file to
  buffer up the video before the decode begins. If you then decode again from 
  the same file, it can happen quicker because Windows has the file cached and
  the read operation happens quicker.

- When you hit Stop while encoding to a file, it can take a moment for the
  application to actually stop.  This is because the application has to flush 
  all in-flight buffers and also flush the file I/O.

- It is generally a good idea to reset the board frequently if making
  modifications to the application and re-running it. Sometimes the EVM board 
  will get in a bad state and must be reset.

  We strongly recommended you at least initiate a 'Debug->Reset CPU' from CCS
  before loading a .out file.

  In addition, you may find it necessary to actually reset the board using the
  POR reset button on the EVM. To do this, use Debug -> Disconnect Emulator 
  from CCS, hit the POR reset button, then use Debug -> Connect Emulator from 
  CCS to re-establish emulation.

  On rare occasions, it will be necessary to do a complete power cycle of the
  board.  To do this, you must quit out of CCS.

- If you modify the DSP/BIOS TCF file and rebuild your CCS project, all source
  files will rebuild. However, the very next incremental build will also cause
  all source files to be rebuilt.

- The NDK Network Stack included in the DVSDK is the evaluation version and
  will expire after 24 hours of continuous operations. Currently, the demo 
  only printf's a message saying the NDK is about to shut down in 5 minutes. 
  No other action is taken by the demo. This means the app has the potential 
  to hang upon reaching the timeout.

- By default, the demo is configured to use DHCP to obtain its IP address.
  You can assign a static IP address by modifying app_main.c and changing
  "dhcp" to 'static' in the configuration section.

- The demo defaults to using a variable bit rate for the H.264 encoder. This
  can be changed by setting the max bit rate in the configuration section
  of app_main.c.

- The video decode task uses a CPU memcpy() to move some data.  A future
  enhancement may change this to use DMA for better performance.

- A future enhancement may add a Pause/Resume feature in addition to just
  Play/Stop. There is some code is in the demo to support this, but is not 
  functional at this time.

- The code is set up to light user LEDs on the EVM but there are currently
  some issues with performing I2C transactions to set the LEDs during runtime,
  so the lines of code that set LED's on/off has been commented out.

- If you want to run the demo using instrumented device drivers, you will need
  to rebuild it using one of the project configurations. Select
  Project->Configuration from CCS to change this.

- The demo is a very dynamic system -- just about everything is created
  dynamically at runtime. Currently, there is no real support to tear down
  these resources.

- While decoding from file, if you are decoding both video and speech files
  and the "Repeat on Decode" box is not checked, the demo will stop playing
  when the first file finishes.

- Mic-in is not supported in this release

- There are many spots in the demo code where a runtime assertion can
  occur, such as a memory alloc fail. In these cases, the demo will abort
  and stop running. If you have CCS open, you will see a message on the
  console. However, if you are just using the host app, you will not see
  any error messages and the host application will freeze up.

- There is the possibility for a memory leak to occur due to buffers not
  getting freed on certain audio driver error conditions.

  A memory leak has been discovered when audio is enabled. 3300 bytes will be
  lost each time play & stop are hit. You can observe this by looking at the
  system stats in the host app, the heap free decreases.

  A new 48 byte memory leak was introduced in 1.10.00. The heap loses 48 bytes
  each time you play then stop.

-------------------------------------------------------------------------------

*******************************************************************************
* End of File
*******************************************************************************
Download Driver Pack

How To Update Drivers Manually

After your driver has been downloaded, follow these simple steps to install it.

  • Expand the archive file (if the download file is in zip or rar format).

  • If the expanded file has an .exe extension, double click it and follow the installation instructions.

  • Otherwise, open Device Manager by right-clicking the Start menu and selecting Device Manager.

  • Find the device and model you want to update in the device list.

  • Double-click on it to open the Properties dialog box.

  • From the Properties dialog box, select the Driver tab.

  • Click the Update Driver button, then follow the instructions.

Very important: You must reboot your system to ensure that any driver updates have taken effect.

For more help, visit our Driver Support section for step-by-step videos on how to install drivers for every file type.

server: web3, load: 2.65