GH1 firmware research volunteers required

Yes, I understand that we cannot byte compare the two versions. Each firmware needs to be disassembled back to assembly code and then the two text files need to be compared for differences. Because we know that the only change between 1.21 and 1.22 firmware is the battery test logic, then it should be relatively simple to figure out what the changes are to disable the battery check. If Panasonic are devloping in C or C++, then the task is more difficult as the compiler will optimize the code differently, but figuring out the changes should still be possible. I am just suggesting that this may be the easiest problem to attack out of all the suggested changes people want to the firmware.

In reality you propose task that is more difficult than original one.
As we know error code number, know main camera state structure and offset in this structure where error code is stored. We just don't know place that set this error code.
 
Hi there,

I salute your initiative with this. Have you considered starting a wiki for this, like the CHDK folks have done? I work doing DSP and firmware for an audio / video company, and would be happy to pitch in a little time annotating the disassembly and identifying subroutines. If there were a wiki that would include IDA files and note what parts of the code need attention, that would make it easier to contribute.

Also, what's the state of the gcc support for this processor? Would the panasonic tools be helpful? While we don't develop for these processors where I work, getting an eval copy of the toolset would likely be easy.

Cheers,
VP
 
I salute your initiative with this. Have you considered starting a wiki for this, like the CHDK folks have done? I work doing DSP and firmware for an audio / video company, and would be happy to pitch in a little time annotating the disassembly and identifying subroutines. If there were a wiki that would include IDA files and note what parts of the code need attention, that would make it easier to contribute.

Seconding this. Even just annotations for the disassembly would be very helpful to others.

Also, what's the state of the gcc support for this processor? Would the panasonic tools be helpful? While we don't develop for these processors where I work, getting an eval copy of the toolset would likely be easy.

AFAIK, even with gcc support for MN103, it would be difficult to make use of it to work with this firmware, since we don't have Panasonic's libraries or headers or even an API (unless you write from scratch, working backwards from the disassembly). Whatever we write would have to be independent code (possibly redundantly reimplementing what's already in the firmware), compiled for MN and then inserted into the firmware asm, or maybe called from the firmware somehow (separate binaries on the SD card?). Working directly on the asm is probably the easiest way to go, and tester13 is doing remarkable progress already.

Also, another OT thought. Do you think there the ability to encode and save video is intrinsically tied to some hardware component in the GH1, or is this mostly done in firmware? Do you think implementing something like this on the G1 is feasible (or even worth looking into)?
 
Also, another OT thought. Do you think there the ability to encode and save video is intrinsically tied to some hardware component in the GH1, or is this mostly done in firmware? Do you think implementing something like this on the G1 is feasible (or even worth looking into)?

We need gcc support, writing on assembler is not good and productive.
Also we have ready Panasonic Windows based crosscompiler and evaluation version of this tool could be priceless.

I am not interested in G1. You must download ID and my module and look yourself.
 
Hi there,

I salute your initiative with this. Have you considered starting a wiki for this, like the CHDK folks have done? I work doing DSP and firmware for an audio / video company, and would be happy to pitch in a little time annotating the disassembly and identifying subroutines. If there were a wiki that would include IDA files and note what parts of the code need attention, that would make it easier to contribute.

Also, what's the state of the gcc support for this processor? Would the panasonic tools be helpful? While we don't develop for these processors where I work, getting an eval copy of the toolset would likely be easy.

Panasonic tools will be very helpful. As they are made by manufacturer.
If you could get documentation for used chip (see beginning of this thread for details) or get evaluation board with it, it'll extremely good.
 
Last edited:
PTool 3.00 is released now, all available patches can be seen on screenshot:

1272266010.jpg


And all are still marked as untested (U).
But I hope that in very near future we will be able to tell how they work.
Only time limit have code instructions change, and it is extremely slight (conditional branch only).
So, must not be any crucial patches here.

Download link and instructions on first page.
 
I am not interested in G1. You must download ID and my module and look yourself.

Pardon me. I wasn't expecting you to look into the G1, but since you are already familiar with the implementation in the GH1, I thought you might be able to make an informed guess as to whether this is done more in hardware (i.e. a discrete codec chip) or in software.

I guess that question goes to others reading this thread as well.

And yes I've already been looking at the G1 firmware myself.
 
Pardon me. I wasn't expecting you to look into the G1, but since you are already familiar with the implementation in the GH1, I thought you might be able to make an informed guess as to whether this is done more in hardware (i.e. a discrete codec chip) or in software.

I guess that question goes to others reading this thread as well.

And yes I've already been looking at the G1 firmware myself.

It is LSI chip.
LSI can set some parameters for hardware block.
This is universal approach in all such LSI.
 
Last edited:
Well, it didn't brick my camera, so that's good! :) Sadly in the movie screen it still lists 29.59 minutes remaining, but I could understand if the display would just be seperate from the actual recording. I have now started recording, I'll let you know in half an hour if it went over that time...
 
tester, you're rocking this thing!

Good luck Chibs, could be first successful code patch ...
 
There is a region code that is set when the main motherboard is installed or replaced. That code drives the supported language group, TV standard (PAL or NTSC) and the 30 min recording limit. Updating firmware has no effect on the language set.

Sounds like the HVs, where the region code is also stored in EEPROM.

Our HV hacker identified & changed just one value (from the service menu which allows changing EEPROM values directly) and it apparently changed his HV20 from NTSC to PAL. He got as far as playing back PAL HDV tapes successfully and confirming that the recording framerates were PAL too when his daughter trashed the cam.
 
the version change works, but I changed on one test only the bitrate to 40mbps and after beginning to record the camera freezes, even the off switch, had to remove battery to switch off. fortunatelly it only happens after hit record, so I reinstalled the firmware.

Also tested 60p->24p, unfortunately it does not work:

the video is 23.97 fps , correct, but the file video structure appears messed, only 1 I frame and the rest all P frames, and it records to 12fps if shutter is 1/60 or less and 6 fps is shutter is over 1/60.

I think this is due to the probable fact that panasonic crippled the codec to give the same low quality video when using 1/30 shutter speed in 720p mode.

tomorrow will do some testings, BTW did you change both variables to 23.97?
 
Back
Top