Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Comparing MP3 and Aiff? (Read 7748 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Comparing MP3 and Aiff?

Hi,

Is there a software package for Mac OSX to compare a MP3 and Aiff visually so you can see the frequency differences i.e what you are losing by using 128 MP3 encoding?

The reason I want to do this is because I am doing a project where I want to visually show the differences between and lossy and raw format.

Thank you for your help in advance.

Darren.

Comparing MP3 and Aiff?

Reply #1 –
Quote from: darren1408 on
Hi,

Is there a software package for Mac OSX to compare a MP3 and Aiff visually so you can see the frequency differences i.e what you are losing by using 128 MP3 encoding?


This is not something I've done.  But I think these programs are called Spectrum Analyzers.

The main places to search for Mac software would be Version Tracker and MacUpdate

Here's what I get at VT.

version tracker: search: spectrum analyzer

Comparing MP3 and Aiff?

Reply #2 –
I'm not sure what you expect to see visually. If you expect to see a low pass filter in the lossy file then that is certainly not universal. The encoded file may contain all of the high frequencies of the original or it may not.

There are other clues that can show up in the spectrum, but these are much more subtle. Hopefully folks here will provide some examples or links to demonstrate what you may be able to see. Ultimately, however, no visual representation is going to do an adequate job of showing what has really been lost in the lossy encoding.

Comparing MP3 and Aiff?

Reply #3 –
A pdq states, the visual differences that are expressed may not result in audible differences.  You can show different lowpass filters and whatnot but that doesn't mean that the lowpass filter is audible.  A better visual representation would be your blind ABX results comparing an AIFF file to a 128kbps AAC file.  That way people understand that a visual spectrum is nice but it is not always a good indicator of audio quality.  Remember that we listen to music with our ears, not our eyes.

Comparing MP3 and Aiff?

Reply #4 –
Hi Everyone,

Thank you for your help! Is there any way I can get any mathematical data then rather than visually?

Thank you again.

Darren.

Comparing MP3 and Aiff?

Reply #5 –
Quote from: darren1408 on
Is there any way I can get any mathematical data then rather than visually?


Get?  Represent?  You can represent mathematical data with equations.  It doesn't really matter whether they're "visual" - i.e., on paper - or spoken or represented in Braille (if that be possible) I guess.

But by "visually" perhaps you mean "graphically".  Did a university teacher ask you do the task or something?  If so, why not seek further guidance from whoever set you the task?

I for one was certainly not going to give you some pompous lecture on why you shouldn't do what you said you wanted to do, particularly as I had no idea why you wanted to do it.  So I just answered your question - told you where to look to find Mac software.  (But at Hydrogen Audio I fear you'll get many people who will tell you what they think you ought to do when you haven't asked and even when they know nothing of your purposes.)

As I say, if someone's asked you to do something and you're a bit unsure about teh details, I'd suggest you ask them for further pointers.

Comparing MP3 and Aiff?

Reply #6 –
Quote
Is there any way I can get any mathematical data then rather than visually?
I would assume MatLab[/u] (or a MatLab Clone) can do that. 

I don't use MatLab, but I understand it can open WAV files (and perhaps other audio formats).  I assume it has a rather steep learning curve, but it does have FFT functions built-in, so you won't have to develop your forumla/algorithm from scratch.

Comparing MP3 and Aiff?

Reply #7 –
Quote from: Nick E on
I for one was certainly not going to give you some pompous lecture on why you shouldn't do what you said you wanted to do, particularly as I had no idea why you wanted to do it.


I didn't realize that certain users (including myself) were being pompous.  The question has already been answer but some of us decided to give further explanation.  Many users often come here to hydrogenaudio showing visual differences between lossy and lossless files trying to make valid audio quality claims.  The claims are then shutdown after the users fail blind ABX tests.  We were just trying to say that visual differences don't always result in audible differences.  So I am sorry if explaining that is being pompous but I would rather have someone explain something to me than feed my false perception.

Comparing MP3 and Aiff?

Reply #8 –
Quote from: kornchild2002 on
.  Many users often come here to hydrogenaudio showing visual differences between lossy and lossless files trying to make valid audio quality claims.  The claims are then shutdown after the users fail blind ABX tests.


Maybe I'm prone to giving the benefit of the doubt, at least on the first round.  It sounded like he wanted this visualization to see/show how lossy works, not to "shoot it down"

Speaking of shooting...  would it be unfair to do some spectral/temporal/whatever computer analysis of a particular song mp3 vs original as an aid to an ABX test?

That is you could have any printout or display of the analysis to watch during the test, but still have to distinguish with your ears.  Is that any less fair then word of mouth hints like focus on the cymbal at 17 seconds?

Comparing MP3 and Aiff?

Reply #9 –
My point was that one cannot judge audio quality from looking at graphs alone.  I do not know what the OP's intentions are.  I just wanted to make sure that they aren't trying to use graphs alone to justify audible differences without actually conducting blind ABX tests.  In other words, I didn't want them to show these graphs and say that there is a difference and one should go with lossless even though that difference may not be audible.  It is just that I have seen too many people come onto these boards, post spectrum analysis graphs, and then try to make valid claims about audio quality without conducting any blind ABX tests (or variant ABX tests that are applicable to their situation).

I think that it would be perfectly acceptable if someone were to conduct blind ABX tests and then post the spectrum analysis graphs.  I just don't think that people should rely on those types of graphs alone as visual differences do not always translate over to audible differences.

Comparing MP3 and Aiff?

Reply #10 –
It is human nature to want to represent things using our most developed sense. The best way to show the audible differences though between codecs is obviously by using your hearing though! You could show the graphs simply to illustrate that what is being mathematically removed DOESN'T equate to audible differences, but you certainly shouldn't use a graph to try and prove the opposite!

Comparing MP3 and Aiff?

Reply #11 –
In answer to the OP - Amadeus - http://www.hairersoft.com/Amadeus.html

It can generate a variety of different graphical visual representations of sound files.  Under the Analyze menu you get the choice of:

Spectrum
Animated Spectrum
3D Spectrum
Sonogram
Real time Spectrum
oscilloscope
Multiband RTA
Waterfall

Comparing MP3 and Aiff?

Reply #12 –
Why not use matlab or octave, time align, take the difference, and show the error spectrum and the original spectrum on one plot, with the error spectrum on top?
-----
J. D. (jj) Johnston

Comparing MP3 and Aiff?

Reply #13 –
Quote from: Woodinville on
Why not use matlab or octave, time align, take the difference, and show the error spectrum and the original spectrum on one plot, with the error spectrum on top?

Please don't encourage people to subtract the lossless version from the lossy version and then either look at or listen to the difference. People are bound to misunderstand what they see/hear.

Comparing MP3 and Aiff?

Reply #14 –
Quote from: pdq on
Quote from: Woodinville on

Why not use matlab or octave, time align, take the difference, and show the error spectrum and the original spectrum on one plot, with the error spectrum on top?

Please don't encourage people to subtract the lossless version from the lossy version and then either look at or listen to the difference. People are bound to misunderstand what they see/hear.


Why not? It makes it clear just how much of the signal YOU CAN NOT HEAR.

That, in and of itself, I think, is an interesting demonstration.

Furthermore, plotting the difference spectrum vs. the original is very enlightening as in regards to both short-term and long-term spectra.

I encourage people to experiment.
-----
J. D. (jj) Johnston
 

Comparing MP3 and Aiff?

Reply #15 –
Quote from: darren1408 on
Hi,

Is there a software package for Mac OSX to compare a MP3 and Aiff visually so you can see the frequency differences i.e what you are losing by using 128 MP3 encoding?

The reason I want to do this is because I am doing a project where I want to visually show the differences between and lossy and raw format.


The first complexity is the fact that AIFF like WAV is a container file format. They can contain either lossy or lossless coded information. Tables of contents and other things can be stuffed away inside, as well.

The best way to visually show the difference between lossless and lossy compression is to encode something that is designed to have a representative and known appearance, after being coded and decoded.

Examples of signals that get butchered by lossy coding (or not so much butchered) would be multitones. A multitone might be composed of sine waves on 1/12 octave intervals, or some such.

You can find examples of this approach at

http://www.pcavtech.com/techtalk/compare_p...ptual/index.htm

I'll warn you in advance that all links to www.pcabx.com are broken, but the basic web page may be helpful.