Hi all,
Does anybody have any experience of using the ADV202 JPEG2000 codec and
its peculiarities?
I am particularly keen to hear if anyone has experienced lock-ups while
encoding or permanently full buffers while decoding.
I'm willing to share experiences of using this chip. I can't seem to
find any other discussion forum or user group for it.
Regards,
Paul.

Michael S wrote:
>Does anybody have any experience of using the ADV202 JPEG2000 codec
>and its peculiarities?
>
I remember a few tests some 2 years ago we had an old revision which
was still very buggy and not matching the docs make sure you have a
recent chip version and all errata sheets

the PCI demo-board was working fine I even had the FAE address for
that chip but I can't find it any more - sorry

but if you find a user group - please let me know!
--
bye,
Michael

Hmmm. The silicon revision, firmware revision, docs, errata etc. don't
seem to match even now. There is a rather strange firmware version
numbering system which is confusing me. It looks like they might have
seen "stalls" in some shape or form but I can't make sense of it. I've
emailed an AD engineer. What I would love is another ADV202 customer to
share the experience with.
Their PCI demo-board hangs on me sometimes too. DMA errors.

usenet4@databuzz.co.uk wrote:
I'm (finally!) starting to ship product using this device. I am not
aware of any problems at the moment, but in view of they many different
ways the device can be interfaced and the huge variety of configuration
options, I certainly can't claim to have explored all the possibilities.
For example, one of the earlier firmware versions worked fine except for
certain combinations of tile size and decomposition levels (but not the
limitations described in the app notes) . Next release, everything was
fine. If I hadn't hit that particular combination of parameters then I
wouldn't have been aware of the problem.

It's disgraceful that the very simple "release notes" for the firmware
are several version behind, leaving us to guess what features do and
don't now work.

I am not aware of any forum to discuss this device and share your
frustration.

Can you describe more fully the circumstances of the failures you have
seen?

Yes, we've got PAL/NTSC video going into the ADV202 from an ADV7185
decoder. It is pretty much the same system as the ADV202 evaluation
board. We are configuring the ADV202 for PAL or NTSC reception and
DMAing compressed data from the CDE FIF small chunks at a time through
a PLX9656 PCI chip. The DMA engine is the one in the PLX and it's
reading from the ADV202's CDE FIF and writing to host memory as
directed by a Linux driver. Again, this is much the same set up as the
evaluation board. The only major difference is that the evaluation board
comes with a Windows driver that does some scatter/gather DMA thing. We
are just doing small single DMA transfers and queuing up the next one
when the DMA completion interrupt comes in.
We've got an issue with flow control between the PLX and the ADV202,
which is why we do these small transfers. DMAing more than is actually
reported available by the CDE FIF count register causes video corruption.
Sometimes, every 10 minutes or so, a DMA transfer refuses to complete.
The req/ack signalling between PLX and 202 freezes the DMA.
That's the first symptom.
The other is in the replay direction. The CDE FIF count/threshold
registers don't seem to work properly. They get into a state whereby the
CDE FIF always looks full. Again this is intermittent. It may be data
dependent.

My evaluation board can lock-up on me too. It reports DMA transfer
errors. This is why I suspect the 202 rather than our design. It could
be dependent on tile size/decomposition levels etc, but if so, nailing
it down with so many configuration options is pretty much impossible.

Regards,
Paul.

usenet4@databuzz.co.uk wrote:
of curiosity, is there any interest in getting a forum/mailing list
going for the ADV202?
, I'd be interested for one. Analog Devices have a Blackfin
forum, but nothing that covers the 202. Grrr.

Hi,

I would be interested in a mailinglist/forum for the ADV202. We have developed two PCB's, each featuring an ADV202. One board is at the receiving end (monitor board) and the other is the transmitter (camera board). Compressed JPEG2000 video data is being transmitted over an ethernet link between the two boards. The system essentially works but I am encountering frequent lock-ups of what seems to be the receiving ADV202 on the monitor board. It appears the ADV202 suddenly just stops asking for new data via DMA...
This one puzzles me a lot. After extensive testing and reading the posts here I am under the impression that there is indeed a bug with the ADV202 chip or the firmware.

I am also interested in a mailinglist/forum for the ADV202.

I may be a little late to the party but I started using the ADV202 just recently. I also found that the eval board would hang up. When I asked technical support they said they never saw the problem. I sent them back the eval board and they still did not see it. I found that it turned out that it would only lock up on store-bought DVD's. I did not get a good response from ADI except maybe it has to do with Macrovision.

I have my encoder board working well. I can save the data to a file and look at individual frames. ADI still does not have a mechanism to view the motion from a stored file. We are working on that.

I just got my decoder board and am bringing it up now. I have a problem where the first frame is flashed on the display and then the ADV202 does not accept any more data. Then the display is just on the bottom 1/3 of the screen with the original frame scrolling through until it eventually breaks up and dies. Have you seen this problem? Does the decoder have to be set for the same parameters as the encoder i.e. deinterlace, NTSC, etc? I think that that is my problem and I will look into that on Monday.

Answered my own question: All that was missing was me telling the decoder that the data stream was 'deinterlaced'.

Have you found a solution? I am having the same type of problem.

We are also facing similar kind of problems with ADV02.We
had configured ADV202 in DMA (single FIFO read mode) TEST mode as per the application note in ADV202 site (AN 799).It was working fine.When we configured ADV202 in operational mode with PAL interface, we could not get DREQ.We are using Single FIFO read DMA mode. The video decoder ADV7183A from
Analog Devices is interfaced to ADV202.We have been degguging the circuit and we have ensured that there is nothing wrong with our circuit. We are using the chip ADV202BBCZ SURF -115 #0540 0.4 759831.1 singapore.

Is this a kind of problem which U faced with ADV202.
What did U mean by stalling exactly.Is it that U are not getting dreq signal after configuring the device successfully. How did U rectify Ur problem.What are the things to be addressed and verified to make the device functional.

When we mailed to AD enginner, he had send a new firmware (encode_2_9_2.dat and encode_2_7r.dat).Even after loading it , we could not get DREQ though the application id was found to be right.

We are also experioencing lockups. Any pointers on how u solved the issue?

Hi Mike,

After extensive reading i found out that the ADV202 has to be incorporated with a Microcontroller/processor to load the firmware. But i'm having a doubt of how to load the firmware into adv202. Can you give me brief explanation of loading the firmware into ADV202. I've planned to use MCF 5282 for this purpose.

Any help in this regard will be appreciated.

Regards
Sathish

ADI's documentaion on loading the firmware is poor. Basically, the '.sea' file consists of 32KB. Think of them as 32-bit words (big endium). Each word has to be written to the ADV202's RAM address starting at 0x0005000. That is, if the first 8 bytes in the '.sea' file are 11 22 33 44 55 66 77 88, write 11223344 into address 0x0005000 then write 55667788 into address 0x00050004; then so on.

That is the easy part. While I have used the ADV202 previously using an NTSC decoder to write the uncompressed data to the VDATA bus, I am having problems writing a still image using the custom speficic mode. I can load the f/w, check that the f/w is running, read and write to all registers (and check that I have done so), but the ADV202(212) does not accept any video data I send it for encoding (tried HVF, EAV/SAV and HIPI.) It is as if the f/w is not running - yet the f/w checks comes out OK. Has anyone else seen this problem?

Hi,
i also tried to configure adv202 in custom specific mode for still image applications.But i am not able to find a way to make it work,when i contacted the AD engineers they suggested me some register settings,but those settings also are not working.
But i could make it work in normal host mode by simply writing the image data in pixel register and then reading the compressed data from code fifo,so you can also try that,but custom specific mode using VDATA lines doesnot seems to work.

Even i will be eager to join such a forum.

hi jim, it seems that you have used adv202 quite a lot .Whats ur take on adv202 getting hanged up,is there really some combination of parameters which can make it hang.
if you have any information plz forward it.

can u plz tell me what parameter values you had used specially in jpeg 2000 parameter registers.
thank you.