Question of mp3 Being a Lossy Format

[Raj Ganatra]

Question of mp3 Being a "Lossy" Format

Hi,

How is mp3 "lossy"? For example, would I lose quality in an mp3 file if I move the file around: say that the mp3 is stored on my c drive, but then I burn it in the same format to a disk, then I copy it from the disk back to my c drive, then I again burn it to another disk, and then from there I re-copy it back to my c drive. Does this process cause the file to lower in quality each time I do this? Or does lossy just mean that if I convert to mp3 to wave, and then back to mp3, then the quality goes down?

Second question: on all those music editing software, can it only edit the song in .wav format? For example, let's say I open an mp3 file in one of those software, does the software first convert it to .wav, edits it, and then re-encode it back to mp3, thus causing a "lossy" effect? Or does the software directly edit the file in mp3 format so that there is no "lossy" effect?

Thanks!

 

[Raj Ganatra]

Another question please: CD quality is said to sample sound at 41100 values per second. What is this "value"? Are they referring to 41,000 individual frequencies of sound per second? So if a sound is producing 100,000 frequencies for a certain second, does the recorder take a statistically accurate sampling of this (adequately randomized) with the sample size being 41,000?

Thanks again!

 

[Pengwuino]

No the file will stay the same unless you are burning those files as CD-Audio because it will have to be encoded into a format stereos and radios and anything that accepts cd's can read. Every other step does not change the file quality however, it is a perfect copy.

 

[-Job-]

Another question please: CD quality is said to sample sound at 41100 values per second.

It's 44100 i think, and measured in Hz. Typically 1 Herz = 1 cycle per second. When talking about sound files 1 Hertz = 1 frame per second.
So a sound file, such as WAV, with "Sample Rate" of 44100Hz means it has 44100 sample frames per second.
A sample frame may be divided into a number of sample points,1 if Mono, 2 if Stereo ...etc. When playing a sound file, a program reads the sample frames and plays them back.
The difference between MP3 and WAV is in the size of the sample frames and the sample rate. The reason is that there is information in a WAV file that can be thrown away without the user noticing significantly. This is what we mean by losing quality, the WAV playback is much more faithful to what was recorded than the MP3, which throws small chunks away because they won't be very noticeable.
You can think about this in the following manner. When recording, a sound recorder needs to read in a sample every so often. If you read a sample 100,000 times per second, you will have a very high quality sound file, but most of the time you don't need that much. Compare this to a video. A video is also composed of frames. When creating an animation, if you use 2 frames per second, then the animation isn't smooth. As you go up in the frame rate the differences between each frame become smaller and the animation becomes smoother and smoother. It reaches a point at about 30 frames per second where you can't tell the difference between the animation and the real thing. This means that if you go up to 60 you won't see any improvement and you will only get a much bigger file size.
So the WAV is like the 60 frame per second animation, and the MP3 the 30 frames per second animation. Of course there's other things an MP3 encoder does to throw away useless information. For example, in your animation, instead of storing every frame, you might just store the first frame and the differences between each frame. You can do similar things to audio.
For my senior design project I'm doing an Audio Mixer so I've had to learn a lot about this stuff, and it's pretty interesting stuff, for example, did you know that in order to mix two sounds together all you need to do add the values in each frame together? And that to increase the volume by a factor X all you need to do is multiple the value of each frame by the value X?