I'm trying to create a wav file from multiple other wav files.
I use AVAsset, AVAssestReader and AVAssetWriter.
The format setting used for the AVAssetWriterInput and AVAssetReaderAudioMixOutput is created like this:
AVAudioFormat(commonFormat: .pcmFormatInt16, sampleRate: 44100, channels: 2, interleaved: true)
And the AVAssetWriter is created like this: AVAssetWriter(url: outputURL, fileType: .wav)
Btw I noticed 2 weirds things:
1) When I create an AVAsset from a wav file I haven't any metadta.
The asset creation is:
let url = URL(fileURLWithPath: mWaveFilePath)
let asset = AVAsset(url: url)
I cannot do simpler, and when I look for metadata properties of this asset I get always empty array with wav file...
2) The most important is when I write a wave file I've the feelling that AVFoundation makes some errors in the wave header. Maybe it comes from me but I manage to create a wave file with audio, and followed some tutorial I've a bad time for finding where the error could come from.
Here is an example of good an and bad header:
The good header before importing the file.
We can see that the format tag is set to 1 which mean PCM. That's what we want.
Now the wrong header after the creation of my audio file:
-2... It's clearly wrong.
So did I miss something on using AVFoundation for creating a wav file, should I do something special?
Related
The thing which I am doing right now is that I am playing RTMP streaming on media server using ffmpeg command and also creating an audio file using google text to speech.
So I want to update mp3 file with silence if there is no content, so that it will keep will keep stream running.
I have tried 2 approaches:
By writing raw binary data to mp3 file but not working as it says content is not accurate.
Concatenate the audio content with the silence data and export file. In this scenario, I am able to update file but stream broken at the point while we are exporting file.
I have tried to write the audio file with binary data and also tried to concatenate audio content with silence and then export file but it break stream while we export the file.
I want to create an M4A file from an MP4, I want to attempt this from scratch without using other libraries but just the raw data.
So far I am able to locate the atom moov and parser it. And as a result I can pull the audio data from the mdat. So then I create my own M4A file with the right ftyp (M4A isomiso2) then add a new mdat with just the audio data I previously recovered, finally I add the moov with the same mvhd, and only the audio trak but with an updated stco to reflect the change in offsets of the chunks of audio data (as they are just one after each other now). I am sure I am doing all of this right.
However the M4A file just plays silence. I believe it is because I have to edit more in the moov but I am not sure what - I put it into FFmpeg corruption and I got:
"Sample rate index in program config element does not match the sample rate index configured by the container."
"Too large remapped id is not implemented."
So as a result I think it is something to do with the stsd atom but I am not sure how to change it.
I have a folder which contains lot of MP3 files, some of them are encoded using mp3PRO.
Since this format is now obsolete, I'd like to convert them back to MP3 (converters can be found easily).
Is there is a way to detect programatically if a file is encoded using mp3PRO format ? (eg : by looking at file header or specific signatures using an hex editor)
The official player is able to detect if file is encoded using mp3PRO (the logo is highlighted or not) so I suppose this is technically possible.
What I found so far is that bitrate of mp3PRO file appears to be pretty low (50% of non encoded file) : eg : a 128 kbps file will appears as 64kbps. However a 320 kbps file will appears as 160 kpbs (which are pretty common) so it cannot be used as a rule.
Here is what I found out and how I fixed it. I wrote in here in case somebody would need it :
MP3Pro files does not contains any special flag in the mp3 header that would help to recognize them.
They are technically very similar to usual mp3 files, except they are encoded half the bit and sample rate (eg : a 128kpbs 44100hz file will be encoded as a 64kps 22050hz file, resulting in mp3pro file being approx half the size of original file).
This has been made for compatibility, so default players can play them without any change.
They also contains some SBR data, which allow to synthetically rebuild the lost audio part (high frequencies) and to play them it was before the mp3 pro conversion.
Detecting the SBR data seems very hard if not impossible : it would require to decode the actual mp3 frames. Also there is no documentation to be found about mp3pro format.
What I did (which works but required some manual effort) : I added all files to be checked to playlist of an mp3 player (foobar 2000 in my case) then sorted the files on the sample rate column : most 22050 hz mp3 files were indeed mp3 pro files.
They were converted back to mp3 using winamp + the mp3pro plugin made for it, available here : http://www.wav-mp3.com/mp3pro-to-mp3.htm
I have successfully changed the muxing.c sample to use video frames that I generate on runtime.
I am trying now to replace the get_audio_frame function with a function that decodes an existing audio file, and writes its samples instead of the synthesized audio-samples in the example code.
I've tried using the "audio decoding" example to decode the audio file, but the not sure how / when to write the samples decoded.
I suggest to check the source of my Karaoke Lyrics Editor which is doing exactly what you need based on ffmpeg. See ffmpegvideoencoder.cpp, see createFile and encodeImage functions.
I have an application for iPAD.
This application records the voice of the microphone.
The audio formats of the item must be PCM, MP3 and WAV files. The MP3 file I get it starting from the original raw file and then convert using LAME.
Unfortunately I have not found any example that allows me to convert a PCM file to a WAV file.
I just noticed that if I put the file extension to WAV format, starting from the raw application saves without problems, so I think that there is no type conversion from PCM WAV files.
Correct?
PS: Sorry for my english ... I use Google Translate
WAV is some kind of a box. PCM is in the box. There are many container formats like MP4. MP4 can contain audio, video or both. It can also contain multiple video or audio streams. Or zip files. Zip files can contain text files. But zip files can also contain images, pdfs,... But you can't say "how can I convert a zip file to the text file inside the zip".
If you want to convert PCM data to a WAVE file you should not many problems because WAV files are quite simple files. Take a look at this:
(See also WAVE PCM soundfile format.)
You first need that header and after you can just append all your pcm data (see the data field).
Converting PCM to WAV isn't too hard. PCM and WAV both format contains raw PCM data, the only difference is their header(wav contains a header where pcm doesn't). So if you just add wav header then it will do the tricks. Just get the PCM data and add the wav header on top of the PCM data. To add wav header with PCM data, check this link.
I was working on a system where it accepts only wav files, but the one I was receiving from amazon Polly was pcm, so finally did this and got my issue resolved. Hope it helps someone. This is an example of nodejs.
// https://github.com/TooTallNate/node-wav
const FileWriter = require('wav').FileWriter
let audioStream = bufferToStream(res.AudioStream);
var outputFileStream = new FileWriter(`${outputFileFolder}/wav/${outputFileName}.wav`, {
sampleRate: 8000,
channels: 1
});
audioStream.pipe(outputFileStream);
function bufferToStream(binary) {
const readableInstanceStream = new Stream.Readable({
read() {
this.push(binary);
this.push(null);
}
});
return readableInstanceStream;
}