I'm writing a C++ library for decoding and encoding audio between different formats/codecs. I have a routine for quickly detecting the format before loading the required codec library.
For WAV files one can simple look for the ASCII values "RIFF" and "WAVE" at the start of the file. The same applies to FLAC, we can simply read in the first 4 bytes, which will be "fLaC".
But how can I quickly detect if a file is MP3? I can't rely on the file extension. I also can't try to decode the first MP3 frame, since there might be additional data at the start of the file (eg: ID3, cover image, etc).
Detecting if a file is an MP3 is more complicated than searching for a fixed pattern in the file.
Some concepts
(See http://www.codeproject.com/Articles/8295/MPEG-Audio-Frame-Header for details)
MP3 file consists of a series of frames and each frame has a header at the beginning.
Header starts at a byte boundary with an 11-bit sync word, which is all 1s. Hence the sync word is either 0xFFE or 0XFFF.
Length of each frame is calculated based on the header parameters.
Algorithm to determine if a file is MP3 or not
Search for the sync word in the file (0xFFF or 0xFFE).
Parse the header parameters.
Determine the frame length using the header parameters.
Seek to the next frame using the frame length.
If you find another sync word after seeking, then the file is mostly an MP3 file.
To be sure, repeat the process to find N consecutive MP3 frames. N can be increased for a better hit-rate.
Related
Say I have a 100GB compressed file, after 76% of uncompression, my device got rebooted by other events, then I simply want to recontinue the uncompression from that 76% mark where I last left off. That's it.
To help with this, I could control how files are compressed and archived.
But while uncompressing on device, no command line, only zlib APIs are available, or any new APIs that may require.
This is a repost, a reworded question, for clarity, I apologize for that. Previously Z_FULL_FLUSH was suggested, but I didn't understand how I will use that 76% mark's offset to initialize the zlib.
Much appreciate any feedbacks.
Thanks
Read thru the zlib's FAQ and annotated usage page for better understanding of how deflate, inflate are working together in compressed stream.
For this, you don't even need to specially prepare the gzip file. You can save the state of inflation periodically. If interrupted, roll back to the previous saved state and start from there.
You can use Z_BLOCK to get inflate() to return at deflate block boundaries. This will be noted in data_type as documented in zlib.h. You would pick an amount of uncompressed data after which to save a new state. E.g. 16 MB. Upon reaching that amount, at the next deflate block boundary would save the location in the compressed data, which is both a byte offset and bit offset within that byte, the location in the uncompressed data you saved up to, and the last 32K of uncompressed data, which you can get using inflateGetDictionary().
To restart from the last state, do a raw inflate, use inflatePrime() to feed the bits from the byte at the compressed data offset, and use inflateSetDictionary() to provide the 32K of history. Seek to the saved offset in your output file to start writing from there. Then continue inflating.
How can I read the compressed image data from the IDAT chunk of a PNG using libpng? I have not found a method to do this in the libpng documentation, but I may have overlooked it. Is it somehow possible to use the "unknown chunk" facility for the IDAT chunk?
The purpose of this is that I want to write a very fast PNG-to-PDF converter. Because PDF supports the PNG data format (with each scanline prefixed by a filter-type byte), it should be possible to just copy over the contents of the (concatenated) IDAT chunks and slap the right PDF headers around it (also copying the palette if necessary). This saves a decompression/re-compression step.
If libpng does not provide such low-level access, does any other library provide this functionality? Otherwise I'll just write a PNG chunk reader myself...
I've been looking for an answer everywhere and I was only able to find some bits and pieces. What I want to do is to load multiple mp3 files (kind of temporarily merge them) and then cut them into pieces using silence detection.
My understanding is that I can use Mp3FileReader for this but the questions are:
1. How do I read say 20 seconds of audio from an mp3 file? Do I need to read 20 times reader.WaveFormat.AverageBytesPerSecond? Or maybe keep on reading frames until the sum of Mp3Frame.SampleCount / Mp3Frame.SampleRate exceeds 20 seconds?
2. How do I actually detect the silence? I would look at an appropriate number of the consecutive samples to check if they are all below some threshold. But how do I access the samples regardless of them being 8 or 16bit, mono or stereo etc.? Can I directly decode an MP3 frame?
3. After I have detected silence at say sample 10465, how do I map it back to the mp3 frame index to perform the cutting without re-encoding?
Here's the approach I'd recommend (which does involve re-encoding)
Use AudioFileReader to get your MP3 as floating point samples directly in the Read method
Find an open source noise gate algorithm, port it to C#, and use that to detect silence (i.e. when noise gate is closed, you have silence. You'll want to tweak threshold and attack/release times)
Create a derived ISampleProvider that uses the noise gate, and in its Read method, does not return samples that are in silence
Either: Pass the output into WaveFileWriter to create a WAV File and and encode the WAV file to MP3
Or: use NAudio.Lame to encode directly without a WAV step. You'll probably need to go from SampleProvider back down to 16 bit WAV provider first
BEFORE READING BELOW: Mark's answer is far easier to implement, and you'll almost certainly be happy with the results. This answer is for those who are willing to spend an inordinate amount of time on it.
So with that said, cutting an MP3 file based on silence without re-encoding or full decoding is actually possible... Basically, you can look at each frame's side info and each granule's gain & huffman data to "estimate" the silence.
Find the silence
Copy all the frames from before the silence to a new file
now it gets tricky...
Pull the audio data from the frames after the silence, keeping track of which frame header goes with what audio data.
Start writing the second new file, but as you write out the frames, update the main_data_begin field so the bit reservoir is in sync with where the audio data really is.
MP3 is a compressed audio format. You can't just cut bits out and expect the remainder to still be a valid MP3 file. In fact, since it's a DCT-based transform, the bits are in the frequency domain instead of the time domain. There simply are no bits for sample 10465. There's a frame which contains sample 10465, and there's a set of bits describing all frequencies in that frame.
Plain cutting the audio at sample 10465 and continuing with some random other sample probably causes a discontinuity, which means the number of frequencies present in the resulting frame skyrockets. So that definitely means a full recode. The better way is to smooth the transition, but that's not a trivial operation. And the result is of course slightly different than the input, so it still means a recode.
I don't understand why you'd want to read 20 seconds of audio anyway. Where's that number coming from? You usually want to read everything.
Sound is a wave; it's entirely expected that it crosses zero. So being close to zero isn't special. For a 20 Hz wave (threshold of hearing), zero crossings happen 40 times per second, but each time you'll have multiple samples near zero. So you basically need multiple samples that are all close to zero, but on both sides. 5 6 7 isn't much for 16 bits sounds, but it might very well be part of a wave that will have a maximum at 10000. You really should check for at least 0.05 seconds to catch those 20 Hz sounds.
Since you detected silence in a 50 millisecond interval, you have a "position" that's approximately several hundred samples wide. With any bit of luck, there's a frame boundary in there. Cut there. Else it's time for reencoding.
For performance of reading and writing a large dataset, we have multiple threads compressing and writing out separate files to a SAN. I'm making a new file spec that will instead have all these files appended together into a single file. I will refer to each of these smaller blocks of a data as a subset.
Since each subset will be an unknown size after compression there is no way to know what byte offset to write to. Without compression each writer can write to a predictable address.
Is there a way to append files together on the file-system level without requiring a file copy?
I'll write an example here of how I would expect the result to be on disk. Although I'm not sure how helpful it is to write it this way.
single-dataset.raw
[header 512B][data1-45MB][data2-123MB][data3-4MB][data5-44MB]
I expect the SAN to be NTFS for now in case there are any special features of certain file-systems.
If I make the subsets small enough to fit into ram, I will know the size after compression, but keeping them smaller has other performance drawbacks.
Use sparse files. Just position each subset at some offset "guaranteed" to be beyond the last subset. Your header can then contain the offset of each subset and the filesystem handles the big "empty" chunks for you.
The cooler solution is to write out each subset as a separate file and then use low-level filesystem functions to join the files by chaining the first block of the next file to the last block of the previous file (along with deleting the directory entries for all but the first file).
I was able to use Ned Batchelder's python code, which I converted to C++, to extract jpgs from pdf files. I'm wondering if the same technique can be used to extract tiff files and if so, does anyone know the appropriate offsets and markers to find them?
Thanks,
David
PDF files may contain different image data (not surprisingly).
Most common cases are:
Fax data (CCITT Group 3 and 4)
raw raster data with decoding parameters and optional palette all compressed with Deflate or LZW compression
JPEG data
Recently, I (as developer of a PDF library) start noticing more and more PDFs with JBIG2 image data. Also, JPEG2000 sometimes can be put into a PDF.
I should say, that you probably can extract JPEG/JBIG2/JPEG2000 data into corresponding *.jpeg / *.jp2 / *.jpx files without external libraries but be prepared for all kinds of weird PDFs emitted by broken generators. Also, PDFs quite often use object streams so you'll need to implement sophisticated parser for PDF.
Fax data (i.e. what you probably call TIFF) should be at least packed into a valid TIFF. You can borrow some code for that from open source libtiff for example.
And then comes raw raster data. I don't think that it makes sense to try to extract such data without help of a library. You could do that, of course, but it will take months of work.
So, if you are trying to extract only specific kind of image data from a set of PDFs all created with the same generator, then your task is probably feasible. In all other cases I would recommend to save time, money and hair and use a library for the task.
PDF files store Jpegs as actual JPEGS (DCT and JPX encoding) so in most cases you can rip the data out. With Tiffs, you are looking for CCITT data (but you will need to add a header to the data to make it a Tiff). I wrote 2 blog articles on images in PDF files at http://www.jpedal.org/PDFblog/2010/09/understanding-the-pdf-file-format-images/ and http://www.jpedal.org/PDFblog/2011/07/extract-raw-jpeg-images-from-a-pdf-file/ which might help.