MP3 convertion with RecordRTC input about lamejs

[shanewho]

mp3encoder.encodeBuffer(samples);
    var mp3Data = mp3encoder.flush();
    console.log(mp3Data);

You get encoded data from the encodeBuffer call AND the flush call. You are only logging the result of the flush, which is just a tiny bit of the data. You need to capture the array returned from encodeBuffer as well.

[rehoehle]

What? That is exact the sample from the page? And that is working in my current version its only not working with the RecordRTC data and what do you suggest to get the correct data and the complete data?

lib = new lamejs();
mp3encoder = new lib.Mp3Encoder(1, 44100, 128); //mono 44.1khz encode to 128kbps
samples = new Int16Array(44100); //one second of silence
var mp3 = mp3encoder.encodeBuffer(samples); //encode mp3
//write mp3 here
mp3 = mp3encoder.flush();
//write mp3 here

This is the example from the Readme. What is the difference to my version?

[shanewho]

I could be wrong, but looking at this code I'm not sure how it would work at all:

Here you call encodeMP3(buffer) and use the result in the blob:

var mp3Blob = new Blob([encodeMP3(buffer)], {type: 'audio/mp3'});

But the encodeMP3 function doesn't actually return any value (i.e., there is no return statement in the function). So the value of the blob is created with an array whose only value is undefined.

The encodeMP3 function needs to use the result of encodeBuffer var data = mp3encoder.encodeBuffer(samples) as well as the result of .flush(). These will be 2 different arrays but you need the data from both of them to get your full mp3.

Here is some code I use, you can hopefully see the difference:
#7 (comment)

[rehoehle]

Yes i know ;) its for debugging... i know that there is no return... that is not the problem... the problem is that the output from the lame lib is empty... and to short. In my old version i use

mp3encoder = new lame.Mp3Encoder(1, 44100, 192);
mp3data = mp3encoder.encodeBuffer(e.data.buf);

that code. And its working well. That make not so much sense because encode buffer and the flush result is stored in the same variable and overwrite the preview one. So how should that work? Then flush have to return the complete data and overwrite the previews data its not appending i think or what do you think?

I think the for loop and iteration over the chunks is to show a progress vor example. I had tested that but it was not working.

Ok i see what you mean. In the other Issue they write both in the blocks array. Ok i will try that. But the questions is why is it working in my current code?

[rehoehle]

Ok i have tested some more and your right now i get the data but i sound like Darth Vader :D any suggestions on that? The record has the double length and is slower.

[shanewho]

My only guess is that the buffer you are recording is stereo and you are encoding it as mono/single channel. If that's the case, you'd have to figure out how to split the buffer into 2 arrays, and pass them both to lame encodeBuffer (and make the Mp3Encoder constructor specify 2 channel). This may not be correct but that is my best guess.

[rehoehle]

Ok i found in the documentation, that the recording from RecordRTC is stereo only. So i have a blob where i have the stereo data. I try to find how to get the left and right channel separately.

[zhuker]

Stereo is interleaved that is you have one 16 bit sample of left followed
by a sample of right.
You essentially split the stereo into two putting every i-th item into left
array and every i+1-th item into right.
They call it deinterleave.
On Aug 27, 2015 11:02 PM, "René H." [email protected] wrote:

Ok i found in the documentation, that the recording from RecordRTC is
stereo only. So i have a blob where i have the stereo data. I try to find
how to get the left and right channel separately.

—
Reply to this email directly or view it on GitHub
#8 (comment).

[rehoehle]

Thx for your reply. Then i try to implement that. Do you know some code where the interleave is implemented?

[zhuker]

You mean deinterleave?
It's easy:
Stereo = array...
Left = new array
Right = new array
For(i=0;i<stereo.length;i+=2) {
Left.push(stereo[i])
Right.push(stereo[i+1])
}

Interleave is the opposite of this operation

Sent from my iPhone

On Aug 31, 2015, at 17:01, René H. [email protected] wrote:

Thx for your reply. Then i try to implement that. Do you know some code where the interleave is implemented?

—
Reply to this email directly or view it on GitHub.

[rehoehle]

Yes i ment that :) thx then i can close that ticket :)

[shamseerahammedm]

@rehoehle do u have any implemented sameple code of this recordrtc to mp3?

[rehoehle]

@shamseerahammedm this is an older post and i don't use this anymore so i can't give you some code. But in the end it was working.