diff --git a/Bootup Logos/README.md b/Bootup Logos/README.md index 55f3431..40687a8 100644 --- a/Bootup Logos/README.md +++ b/Bootup Logos/README.md @@ -1,8 +1,12 @@ -# Preview of the logos -||   TS100   |   TS80    |   TS80P    |   Pinecil    |   IronOS     --|-|-|-|-|- -static
(right)|TS100|TS80|TS80P|Pinecil|IronOS -static
(left)|TS100_L|TS80_L|TS80P_L|Pinecil_L|IronOS_L -animated|||||IronOS -notes|||||   ^^^^^^^^
*(This loops just for
demonstration purposes,
the real thing
only plays once.)* +## Boot up logo's are logos or animations shown on boot of IronOS + +These are programmed into the device just like the normal firmware. +They can be (re)programmed as many times as desired after flashing the normal firmware. + + +### Data storage format + +The data is stored into the second last page of flash, this gives 1024 bytes of space for the entire payload of bootup logo data. + +The first byte is marked purely to \ No newline at end of file diff --git a/Bootup Logos/img2logo.py b/Bootup Logos/img2logo.py index f7ee3f9..ebed819 100755 --- a/Bootup Logos/img2logo.py +++ b/Bootup Logos/img2logo.py @@ -11,11 +11,7 @@ from output_dfu import DFUOutput try: from PIL import Image, ImageOps except ImportError as error: - raise ImportError( - "{}: {} requres Python Imaging Library (PIL). " - "Install with `pip` (pip3 install pillow) or OS-specific package " - "management tool.".format(error, sys.argv[0]) - ) + raise ImportError("{}: {} requres Python Imaging Library (PIL). " "Install with `pip` (pip3 install pillow) or OS-specific package " "management tool.".format(error, sys.argv[0])) VERSION_STRING = "1.0" @@ -24,6 +20,9 @@ LCD_HEIGHT = 16 LCD_NUM_BYTES = LCD_WIDTH * LCD_HEIGHT // 8 LCD_PAGE_SIZE = 1024 +DATA_PROGRAMMED_MARKER = 0xAA +FULL_FRAME_MARKER = 0xFF + class MiniwareSettings: IMAGE_ADDRESS = 0x0800F800 @@ -41,9 +40,7 @@ class PinecilSettings: DFU_PINECIL_PRODUCT = 0x0189 -def still_image_to_bytes( - image: Image, negative: bool, dither: bool, threshold: int, preview_filename -): +def still_image_to_bytes(image: Image, negative: bool, dither: bool, threshold: int, preview_filename): # convert to luminance # do even if already black/white because PIL can't invert 1-bit so # can't just pass thru in case --negative flag @@ -67,11 +64,7 @@ def still_image_to_bytes( if preview_filename: image.save(preview_filename) # pad to this size (also will be repeated in output Intel hex file) - data = [0] * LCD_PAGE_SIZE - - # magic/required header - data[0] = 0xAA # Indicates programmed page - data[1] = 0xBB + data = [] # convert to LCD format for ndx in range(LCD_WIDTH * LCD_HEIGHT // 8): @@ -80,7 +73,15 @@ def still_image_to_bytes( for y in range(8): if image.getpixel((ndx % LCD_WIDTH, y + bottom_half_offset)): byte |= 1 << y - data[2 + ndx] = byte + data.append(byte) + + """ DEBUG + for row in range(LCD_HEIGHT): + for column in range(LCD_WIDTH): + if image.getpixel((column, row)): sys.stderr.write('█') + else: sys.stderr.write(' ') + sys.stderr.write('\n') + """ return data @@ -93,9 +94,27 @@ def calculate_frame_delta_encode(previous_frame: bytearray, this_frame: bytearra return damage -def animated_image_to_bytes( - imageIn: Image, negative: bool, dither: bool, threshold: int -): +def get_screen_blob(previous_frame: bytearray, this_frame: bytearray): + """ + Given two screens, returns the smaller representation + Either a full screen update + OR + A delta encoded form + """ + outputData = [] + delta = calculate_frame_delta_encode(previous_frame, this_frame) + if len(delta) < (len(this_frame)): + outputData.append(len(delta)) + outputData.extend(delta) + # print("delta encoded frame") + else: + outputData.append(FULL_FRAME_MARKER) + outputData.extend(this_frame) + # print("full encoded frame") + return outputData + + +def animated_image_to_bytes(imageIn: Image, negative: bool, dither: bool, threshold: int): """ Convert the gif into our best effort startup animation We are delta-encoding on a byte by byte basis @@ -106,75 +125,32 @@ def animated_image_to_bytes( The naïve implementation would save the frame 5 times But if we delta encode; we can make far more frames of animation for _some_ types of animations. This means reveals are better than moves. - Data is stored in the byte blobs, so if you change one pixel in upper or lower row, changing another pixel in that column on that row is "free" + Data is stored in the byte blobs, so if you change one pixel, changing another pixel in that column on that row is "free" """ + frameData = [] frameTiming = None for framenum in range(0, imageIn.n_frames): print(f"Frame {framenum}") imageIn.seek(framenum) image = imageIn - if image.mode != "L": - image = image.convert("L") - # Resize to lcd size using bicubic sampling - if image.size != (LCD_WIDTH, LCD_HEIGHT): - image = image.resize((LCD_WIDTH, LCD_HEIGHT), Image.BICUBIC) - if negative: - image = ImageOps.invert(image) - threshold = 255 - threshold # have to invert threshold as well - - if dither: - image = image.convert("1") - else: - image = image.point(lambda pixel: 0 if pixel < threshold else 1, "1") - - frameb = [0] * LCD_WIDTH * (LCD_HEIGHT // 8) - for ndx in range(LCD_WIDTH * LCD_HEIGHT // 8): - bottom_half_offset = 0 if ndx < LCD_WIDTH else 8 - byte = 0 - for y in range(8): - if image.getpixel((ndx % LCD_WIDTH, y + bottom_half_offset)): - byte |= 1 << y - # store in endian-reversed byte order - frameb[ndx] = byte + frameb = still_image_to_bytes(image, negative, dither, threshold, None) frameData.append(frameb) # Store inter-frame duration frameDuration_ms = image.info["duration"] if frameDuration_ms > 255: frameDuration_ms = 255 - if frameTiming is None: + if frameTiming is None or frameTiming == 0: frameTiming = frameDuration_ms - print(f"Found {len(frameData)} frames") + print(f"Found {len(frameData)} frames, interval {frameTiming}ms") # We have no mangled the image into our frambuffers - # Now create the "deltas" for each frame - frameDeltas = [[]] - - for frame in range(1, len(frameData)): - - frameDeltas.append( - calculate_frame_delta_encode(frameData[frame - 1], frameData[frame]) - ) # Now we can build our output data blob # First we always start with a full first frame; future optimisation to check if we should or not - - bytes_black = sum([1 if x == 0 else 0 for x in frameData[0]]) - if bytes_black > 96: - # It will take less room to delta encode first frame - outputData = [0xAA, 0xCC, frameTiming] - delta = calculate_frame_delta_encode([0x00] * (LCD_NUM_BYTES), frameData[0]) - if len(delta) > (LCD_NUM_BYTES / 2): - raise Exception("BUG: Shouldn't delta encode more than 50%% of the screen") - outputData.append(len(delta)) - outputData.extend(delta) - print("delta encoded first frame") - else: - outputData = [0xAA, 0xDD, frameTiming] - outputData.extend(frameData[0]) - print("Used full encoded first frame") - - # Now we delta encode all following frames + outputData = [DATA_PROGRAMMED_MARKER] + outputData.append(frameTiming) + outputData.extend(get_screen_blob([0x00] * (LCD_NUM_BYTES), frameData[0])) """ Format for each frame block is: @@ -184,26 +160,13 @@ def animated_image_to_bytes( OR [0xFF][Full frame data] """ - for frame in range(1, len(frameData)): - data = [] - if len(frameDeltas[frame]) > LCD_NUM_BYTES: - data.append(0xFF) - data.extend(frameData[frame]) - print(f"Frame {frame} full encodes to {len(data)} bytes") - else: - data.append(len(frameDeltas[frame])) - data.extend(frameDeltas[frame]) - - print(f"Frame {frame} delta encodes to {len(data)} bytes") - if len(outputData) + len(data) > 1024: - print( - f"Animation truncated, frame {frame} and onwards out of {len(frameData)} discarded" - ) + for id in range(1, len(frameData)): + frameBlob = get_screen_blob(frameData[id - 1], frameData[id]) + if (len(outputData) + len(frameBlob)) > LCD_PAGE_SIZE: + print(f"Truncating animation after {id} frames as we are out of space") break - outputData.extend(data) - if len(outputData) < 1024: - pad = [0] * (1024 - len(outputData)) - outputData.extend(pad) + print(f"Frame {id} encoded to {len(frameBlob)} bytes") + outputData.extend(frameBlob) return outputData @@ -237,22 +200,19 @@ def img2hex( except BaseException as e: raise IOError('error reading image file "{}": {}'.format(input_filename, e)) - """ DEBUG - for row in range(LCD_HEIGHT): - for column in range(LCD_WIDTH): - if image.getpixel((column, row)): sys.stderr.write('█') - else: sys.stderr.write(' ') - sys.stderr.write('\n') - """ if make_erase_image: data = [0xFF] * 1024 elif getattr(image, "is_animated", False): data = animated_image_to_bytes(image, negative, dither, threshold) else: - data = still_image_to_bytes( - image, negative, dither, threshold, preview_filename - ) + # magic/required header + data = [DATA_PROGRAMMED_MARKER, 0xBB] + data.extend(still_image_to_bytes(image, negative, dither, threshold, preview_filename)) + # Pad up to the full page size + if len(data) < LCD_PAGE_SIZE: + pad = [0] * (LCD_PAGE_SIZE - len(data)) + data.extend(pad) deviceSettings = MiniwareSettings if isPinecil: deviceSettings = PinecilSettings @@ -266,9 +226,7 @@ def img2hex( deviceSettings.DFU_PINECIL_PRODUCT, deviceSettings.DFU_PINECIL_VENDOR, ) - HexOutput.writeFile( - output_filename_base + ".hex", data, deviceSettings.IMAGE_ADDRESS - ) + HexOutput.writeFile(output_filename_base + ".hex", data, deviceSettings.IMAGE_ADDRESS) def parse_commandline(): @@ -305,9 +263,7 @@ def parse_commandline(): "--threshold", type=zero_to_255, default=128, - help="0 to 255: gray (or color converted to gray) " - "above this becomes white, below becomes black; " - "ignored if using --dither", + help="0 to 255: gray (or color converted to gray) " "above this becomes white, below becomes black; " "ignored if using --dither", ) parser.add_argument( @@ -317,9 +273,7 @@ def parse_commandline(): help="use dithering (speckling) to convert gray or " "color to black and white", ) - parser.add_argument( - "-f", "--force", action="store_true", help="force overwriting of existing files" - ) + parser.add_argument("-f", "--force", action="store_true", help="force overwriting of existing files") parser.add_argument( "-E", "--erase", @@ -327,9 +281,7 @@ def parse_commandline(): help="generate a logo erase file instead of a logo", ) - parser.add_argument( - "-p", "--pinecil", action="store_true", help="generate files for Pinecil" - ) + parser.add_argument("-p", "--pinecil", action="store_true", help="generate files for Pinecil") parser.add_argument( "-v", "--version", @@ -346,17 +298,11 @@ if __name__ == "__main__": args = parse_commandline() if os.path.exists(args.output_filename) and not args.force: - sys.stderr.write( - 'Won\'t overwrite existing file "{}" (use --force ' - "option to override)\n".format(args.output_filename) - ) + sys.stderr.write('Won\'t overwrite existing file "{}" (use --force ' "option to override)\n".format(args.output_filename)) sys.exit(1) if args.preview and os.path.exists(args.preview) and not args.force: - sys.stderr.write( - 'Won\'t overwrite existing file "{}" (use --force ' - "option to override)\n".format(args.preview) - ) + sys.stderr.write('Won\'t overwrite existing file "{}" (use --force ' "option to override)\n".format(args.preview)) sys.exit(1) img2hex( diff --git a/Bootup Logos/output_dfu.py b/Bootup Logos/output_dfu.py index 2ee8057..5194a60 100644 --- a/Bootup Logos/output_dfu.py +++ b/Bootup Logos/output_dfu.py @@ -14,17 +14,14 @@ class DFUOutput: def writeFile( cls, file_name: str, - data: bytearray, + data_in: bytearray, data_address: int, tagetName: str, alt_number: int, product_id: int, vendor_id: int, ): - data: bytearray = b"" - - for byte in data: - data += byte.to_bytes(1, byteorder="big") + data: bytearray = bytearray(data_in) data = struct.pack("<2I", data_address, len(data)) + data data = ( diff --git a/README.md b/README.md index 52d7885..e30bc91 100644 --- a/README.md +++ b/README.md @@ -1,2 +1,8 @@ # IronOS-Meta -Storing meta information about devices that dont need to be in the main repo + +Storing meta information for IronOS. +This are things that are not part of the core "OS". +This includes photographs of hardware, datasheets, schematics and of course **bootup logos**. + +This repository uses github actions to automagically build the logos for each device. +Periodically a "release" will be tagged and pre-compiled logo's will be put there as well to make it easy.