anal_mp3.c

#include <stddef.h>
#include <stdint.h>
#include <stdbool.h>

#include <stdio.h>
#include <stdlib.h>
#include <assert.h>

typedef struct {
	uint8_t* buf;
	int curr;
	int len;
} stream_t;

typedef struct {
	// all unchanged from the header, ids of these things
	uint8_t id;
	uint8_t level;

	uint8_t bitrate;
	uint8_t samplerate;
} frame_header_t;

// returns read byte
// or negative on error (eof)
int read_next_byte(stream_t* stream, uint8_t* byte) {
	if(stream->curr == stream->len) return -1;

	*byte = stream->buf[stream->curr++];
	return (int)*byte;
}

// returns header position
// or negative on error (eof)
int read_next_frame(stream_t* stream, frame_header_t* new_frame) {
	uint8_t byte = 0x00;

	while(true) {
		uint8_t byte1;
		int pos = stream->curr;
		if(read_next_byte(stream, &byte1) < 0) break;
		if(byte1 != 0xFF) continue;

		uint8_t byte2;
		if(read_next_byte(stream, &byte2) < 0) break;
		if(byte2 & 0XF0 != 0xF0) continue;

		uint8_t byte3, byte4;
		if(read_next_byte(stream, &byte3) < 0) break;
		if(read_next_byte(stream, &byte4) < 0) break;

		new_frame->id = (byte2 >> 3) & 0x01;
		new_frame->level = (byte2 >> 1) & 0x03;

		new_frame->bitrate = (byte3 >> 4) & 0x0F;
		new_frame->samplerate = (byte3 >> 2) & 0x03;

		return pos;
	}

	return -1;
}

// returns 0 on success
// negative on unsupported frame
int is_frame_supported(const frame_header_t* frame) {
	// only supporting MPEG I Level 3
	if(frame->id != 0x1) return -1;
	if(frame->level != 0x1) return -2;

	// invalid bitrate & samplerates
	if(frame->bitrate == 0xF) return -3;
	if(frame->samplerate == 0x3) return -4;

	return 0;
}

// returns location of the TAG header
// or negative on error (not found)
int id3v1_header_location(stream_t* stream) {
	while(true) {
		int pos = stream->curr;

		uint8_t byte1, byte2, byte3;
		if(read_next_byte(stream, &byte1) < 0) break;
		if(byte1 != 'T') {
			continue;
		}

		if(read_next_byte(stream, &byte2) < 0) break;
		if(byte2 != 'A') {
			stream->curr--;
			continue;
		}

		if(read_next_byte(stream, &byte3) < 0) break;
		if(byte3 != 'G') {
			stream->curr--;
			continue;
		}

		return pos;
	}

	return -1;
}

int main(int argc, char** argv) {
	assert(argc == 2);
	char* filename = argv[1];

	FILE* fp = fopen(filename, "rb");
	assert(fp);

	fseek(fp, 0, SEEK_END);
	size_t filelen = ftell(fp);
	fseek(fp, 0, SEEK_SET);

	printf("%zu bytes in file\n", filelen);

	uint8_t* buf = malloc(filelen);
	size_t bytes_read = fread(buf, 1, filelen, fp);
	assert(bytes_read == filelen);

	stream_t stream = { buf, 0, filelen };
	int tag_pos = id3v1_header_location(&stream);
	stream.curr = 0;
	if(tag_pos > 0) {
		printf("id3 header found at %d\n", tag_pos);
		stream.len = tag_pos;
	} else {
		printf("id3 header not found\n");
	}


	size_t num_frames = 0;
	size_t valid_frames = 0;

	while(true) {
		frame_header_t header;
		int pos = read_next_frame(&stream, &header);
		if(pos < 0) break;
		num_frames++;

		if(is_frame_supported(&header) < 0) continue;
		// printf("frame %zu at pos %d\n", num_frames, pos);
		valid_frames++;
	}
	printf("%zu frames in total, %zu valid\n", num_frames, valid_frames);

	free(buf);
	fclose(fp);

	return 0;
}