Ursadon wrote: ↑11 Mar 2019, 13:22
All images stored in Flash in this format:
===================================
00 00 00 FF 01 00 00 00 - Header
...
... RLE-Encoded (with zero-pad)
...
XX XX - width
XX XX - Height
XX XX - width
00 00 - <probably> Height - but seems like always 00 00
30 XX XX XX - PTR to start of data frame
30 XX XX XX - PTR to start of image
00 00 56 38 - End of image (EOI)
===================================
Today I found out that the above "Header" of the image contains actually other information and in certain situations might be different.. so the way we were using this format in CH to find images in Main or Extended Flash is ... wrong
The above table translates to:
===================================
00 00 00 XX - number of colors for image - usually, yes, XX = 0xFF = 255 (Curious if I can put more than 0x00FF and Convers will be able to show it)
01 00 00 00 - unknown for now
XX XX XX XX - offset of where is LUT table for this image (BGR format). If all are 00, then use the current Theme LUT table.
...
... RLE-Encoded (with zero-pad)
...
XX XX YY YY - X = Width, Y = Height (Offset in flash for the first byte = Offset4Main in CH and represents the only information that IPC algorithm needs to know in order to show an image)
XX XX 00 00 - X = number of bytes per line (usually this is equal with width); 00 00 seem to be present only to keep data aligned in a 4-byte format
YY XX XX ZZ - pointer / offset to image data (the RLE-Encoded image part); YY = 00 then it's in Main Flash; 30 it's in Extended Flash; ZZ = (VV - 0x0C)
YY XX XX VV - pointer / offset to image frame; YY = 00 -> in Main Flash; 30 -> Ext. Flash; This is the offset where we find the first bytes that are explained in this table
00 00 56 38 - End Of Image (EOI) - I think this exists only to be able to try to see if there is an image here when you don't have an "Offset4Main"; for the designing software... or tools like CH
===================================
In CH we were looking for the Header then the EOI and what was in between together with those 2 was considered to be image frame.
Because we were looking only for "00 00 00
FF 01 00 00 00" we couldn't detect some images... that are present in custom firmware
Also, because of this, there were images in custom FW with logos that couldn't be decoded or made CH crash - the bytes that contain the offsets for image data and frame are changed for the original image with bytes that contain offsets for the new image. This means that if you want to change an image all you have to do is to go to original image and change them. There is no need to go to Main Flash and change there the offset "Offset4Main" - even if it also worked this way.
Flashing back Main Flash is faster that flashing Extended Flash. So if you want to replace an image, from my point of view, if you also modify Main Flash then you only need to write a sector (or how many are needed for that image to be stored in Ext. Flash) instead of entire Extended Flash.
The easy way which takes longer to flash back is to modify only Extended Flash.
I'll come back with an updated version of CH that will have this new algorithm in order to find all the images... and maybe also the option to have a VBF that contain all what is needed to be flashed back to Extended Flash.
