We have a RK2705, the arm core runs up to 240 MHz and DSP up to 176 MHz.
Make sure you have the numbers right in the PLL (i think there is a list of PLL values in the HW_PMU module).
Increase the values accordingly...

Well, I didn't notice any difference in performence of video playback between original fw and sdk. Even in original firmware, videos in 'supported' range of resolution plays not good enough. RMVB file I have plays nice, until some big action come up in the movie, then it slows down really bad.

But still, I tested so many different video formats, and I am impressed. FLV, MPG, 3GP plays very nicely. Though I heard some clicks in audio of FLV file. I watched two such files, and there were noticable clicks on audio so I am not sure whether it is a particular file issue.

EDIT: Well, I tested more videos, and others play very nicely. I am positively surprised by resizing algorithm. It's anti-aliased! ;-)

Has anyone got any clue?

He probably can change it, but using assembler. It also depends on what changes have to be applied. If it's only modification of keyboard, screen driver etc it is not so difficult. That way you don't even modify ASM code, but just values. If you are talking about greater changes in the code, you need huge ASM skills.

Could you guys give us a helping hand ('us' being the folks from the mp4nation forums). The RK2806 is the same, the Wolfson audio chip is the same (if my information is correct), it's just another screen and touchscreen (or would the memory need changes too?). My skills go as far as that I know what assembly is, but I got no clue on what you're saying about changing the values with an assembler. To me it looks like I need something that can turn thise BASE.rkw file in something readable, and then copy paste certain parts from the original firmware?:confused:

Edit: I actually got some readable text out of it by running it through "ida disassembler" But I really got no clue what to do now. It refers to files like guidisp.c but I don't know how to find and disassemble those files?

well, I barely disassembled RK270X fw, so I definitely won't be able to help you in this.

Though, I am sure that screen driver is easy to modify, but if you want to modify touchscreen driver, you are in big troubles. I couldn't make a driver for my RK260X player using provided SDK in C (and I gave up). But in ASM it is even harder, and in my opinion impossible. I mean, everything is possible, but maybe you don't realize how much more you need to know to do such a small thing like modifying the keyboard driver.

I don't mean to discourage you but, there aren't people enough knowledgeable to tell you what to do without experimenting with the player. It is so much easier if someone has the player in his hands.

To be honest, we are just amateurs trying to squeeze everything we can from the players we've got. That's not even programming forum, but just mpx, though it's a pretty good luck that some of us have experience in programming.

I hope you'll make progress in it.

BTW. What for do you want to get this FW working? How does it expand the functionality of your player?

Ah well, didn't realize it would be that much work, thought it would be a matter of locating the right pieces and copy paste 'm, but I guess that's just not the way it works.

Ramos FW has some bugs like it messes up song order (when you want to play all the songs of an artist, it will also play songs of other artists). And overall you can see that Oppo has put much more effort in the FW than Ramos did (it also looks nicer etc).

I will just hope that Ramos releases a better FW themselves, this is way over my head I'm afraid. And I'm sure they will, Ramos is a pretty big name in China.

This is an interesting find.

To be certain it's a ZSP500 we should be able to find 32-bit instructions. ZSP400 (or rather, the ZSP G1 instruction set) only has 16-bit instructions.

I haven't been able to find a document describing the ZSP G2 instruction set. Also the ARM7EJC in the rk2705 remains a mystery, assuming it's not a typo and ARM7EJ-S was meant instead.

In RealFunctions.c there's two references to zsp500

That certainly suggests a ZSP500, but it's still just a comment. I'm not convinced yet because there are discrepancies: the "ZSP500" in the rk2705 doesn't run at 400 MHz, but only at 176 MHz; furthermore it is said to have not a 24-bit, but a 32-bit address space.

A simple test is to check a driver using a ZSP200/ZSP400/ZSP410 disassembler and see if it skips instructions or produces errors. If it doesn't we're back to square one.

Actually I'm more interested in the question if the rk2705 has a MMU.