[Sir Nobax]

First, i don't want to hurt anyone's feeling, i never said that listening to encodings is bad and that i dislike everyone who does that (actually totally the opposite, i am quite jealous for them doing that, but i am the guy who needs it on paper to believe it for myself).

I just did it because some people are like me, wanting to see real proof instead of only hearing it, wich will be different for all people. I wanted to make my contribute to those fellow head-fi'ers.

Quote:

Originally Posted by zip22 how is looking at pictures more effective than listening when you are trying to judge the quality of an audio reproduction?

Just an example how i judge the quality of an codec; If you'd take a look at the pic of the 96kbps HE-AAC sample and a look at the MP3Pro one, you'll see that the MP3Pro one has an enourmous amount of disortion on the last second, while the HE-AAC, has'nt and still has all of the high tones of the spectrum, conclusion MP3Pro Jitters at 10khz, so when you want to encode classical music (for example) on a really low bitrate, use HE-AAC instead.

Quote:

Originally Posted by zip22 now, if you would ABX and then explain your findings with visual representations, that would be useful. but simply attempting to judge based on the images is illogical.

If you listen first and use the images to strenghten your findings then this threat has use for you, i do too, i just need the images first and then i listen to it to strenghten my findings.


Quote:

Originally Posted by hYdrociTy Thank you so much for your time Actually I have one request. AAC-LC @224 with CBR and VBR, since that seems to be the limit before the ipod buffer overfills and there is hard disk interaction.

The new pictures look less saturated, this is because i now use Audition 1.5, im still looking for the old program though, i think the right one would be CoolEdit Pro, but i suffer from slow internet speeds (Dial-in with the GPRS of my phone)

I have done both encodings, 224 CBR is quite good but i cant see any difference between the 192kbps and the 224kbps encoding, i think this is because it's al ready "maxxed", if the low-pass filter would be standard set to 19khz or 20khz you would see the difference. 224 VBR was unarchievable because the tone is too simple (the file was averaged at 192kbps [147kilobytes \ 8 * 6 seconds = 192kbps average] , it was a perfect as-good-as lossless audio file.

Here is the 224kbps CBR:

And here is the 224kbps VBR:

Quote:

Originally Posted by db597 If only we could do a similar test for headphones - e.g. feed in a reference signal and then record the output with a mic. Even on the best cans, I bet the graphs would be pretty bad with lots of holes.

My DT550

and my HQ-1700

 

[zip22]

Quote:

Originally Posted by Sir Nobax /img/forum/go_quote.gif If you listen first and use the images to strenghten your findings then this threat has use for you, i do too, i just need the images first and then i listen to it to strenghten my findings.

when you look at the images first, you have a preconceived notion of what you are going to hear. this significantly taints your audio test. if i listen to a particular codec and hear something wrong, there is no way for it to influence the graphs.

 

[Sir Nobax]

Quote:

Originally Posted by Sparky191 /img/forum/go_quote.gif Another vote for FLAC and how about ATRAC 256/292/352 and lossless. I'm not entirely sure what I'm looking at in these pictures. Anyone care to explain one of these images?

Comparing FLAC and other Lossless codec's would be quite useless, as it is a carbon copy of the original WAV.
However i did two Lossless encodings, WMA 9.1 Lossless and WMA 9.2 Lossless, the only thing that was different was that the encoded files where 0,0022 seconds longer ?!

And indeed ATRAC already came up in my mind, i just need to find myself a good Atrac-encoder.

A small explenation: The pictures show several sweeps of an tone wich starts at 0Hz and ends at 22Khz, it has a lot of reverb (the bouncing thin lines) to make it more complicated for the codecs, how closer the encoded sample looks like the original WAVE file the better is the codecs techniques, off course none can hear the tiny holes that even appear in the 256kbps encodings but it still shows you how close it is to the source.

 

[Sir Nobax]

Quote:

Originally Posted by zip22 /img/forum/go_quote.gif when you look at the images first, you have a preconceived notion of what you are going to hear. this significantly taints your audio test. if i listen to a particular codec and hear something wrong, there is no way for it to influence the graphs.

You are tottally right, but it makes me happier listening to 256kbps WMA 10 Pro's knowing that its as close as it gets at that bitrate. And its all about the happyness you have when you are listening to your favorite can's or IEM's, is'nt it?

 

[Febs]

Quote:

Originally Posted by Sir Nobax /img/forum/go_quote.gif I just did it because some people are like me, wanting to see real proof instead of only hearing it, wich will be different for all people.

But that is precisely the point. These graphs are not "real proof" of anything. They do not correlate to how the resulting files will sound. Take your specific example:

Quote:

If you'd take a look at the pic of the 96kbps HE-AAC sample and a look at the MP3Pro one, you'll see that the MP3Pro one has an enourmous amount of disortion on the last second, while the HE-AAC, has'nt and still has all of the high tones of the spectrum, conclusion MP3Pro Jitters at 10khz, so when you want to encode classical music (for example) on a really low bitrate, use HE-AAC instead.

First of all, while I have read several definitions of the word "jitter" with respect to audio, the phenomenon that is reflected in your spectrograph does not appear to be "jitter" under any definition of that term. Second, even if we assume for the purposes of argument that the spectrograph does in fact show some sort of artifact in the last second, we still cannot conclude from the spectrograph that the artifact is audible. You need to do a listening test for that.

Quote:

If you listen first and use the images to strenghten your findings then this threat has use for you, i do too, i just need the images first and then i listen to it to strenghten my findings.

Unless your listening test is conducted double-blind, this a seriously flawed methodology. If you develop an expectation about what you are going to hear, and then listen for the purpose of confirming that expectation, you are almost ensuring that your results will be influenced by an expectancy bias.

Edit: I stepped away from the computer while posting this, and in the interim, there were several posts made that make mine redundant.

 

[Thelonious Monk]

it's like a headphone frequency response graph; fun to look at, but doesn't tell you enough about the real sound...

 

[RotAtoR]

Too bad these graphs have very little meaning in terms of codec quality. The purpose of a lossy codec is, in essence, to add as much (imperceptible) noise as possible to the signal in a way that allows better compression. So the best lossy codec may in fact be the one whose graph looks the "worst" but is still produces audio perceptibly indistinguishable from the original signal. In the end you can only rely on listening tests to test the codec quality.

 

[Sir Nobax]

Quote:

Originally Posted by RotAtoR /img/forum/go_quote.gif Too bad these graphs have very little meaning in terms of codec quality.

Wow!!!

How original.
Your number six who says exactly the same

.

... serious guys, if you don't like it .. don't look and don't post, we arent flaming Headroom because they put headphone frequency response graphs on their site are we? , and like Thelonious Monk said, neither do those things say anything about the quality, however it gives an IDEA how it sounds.

 

[redshifter]

sir nobax,
wonderful effort! a+ for you!

A+

 

[davei]

Quote:

Originally Posted by Sir Nobax /img/forum/go_quote.gif Since the day i started 'listening' to music instead of 'hearing' it i'm worried about the quality of lossy codecs like MP3, AAC and WMA. Every time i hear a song i wonder if it is really how it has to sound, because of the loss of data in the compression.

That says it all right there, if it eliminates worries and makes the OP feel better about using a particular codec, great. Personally, with portable audio I'm usually far more interested in what's playing - I never listen to wide frequency sweeps, and a great/favorite song will always put a smile on my face even if it's recorded in mono on a ratty old Type I cassette

 

[audiomagnate]

Quote:

Originally Posted by Sir Nobax /img/forum/go_quote.gif Wow!!! How original. Your number six who says exactly the same . ... serious guys, if you don't like it .. don't look and don't post, we arent flaming Headroom because they put headphone frequency response graphs on their site are we? , and like Thelonious Monk said, neither do those things say anything about the quality, however it gives an IDEA how it sounds.

Get used to it. Anything remotely based on real science around here gets flamed. You have to understand that accurate sound reproduction is more of a religion than a science to most of these people, and you're an atheist who's telling them there is no god. You should have started a thread about how changing the power cord on your CD player made the most amazing improvement in sound you've ever heard through your KSC-75s. Then you'ld be feeling the love right now.

I think the complexity of the test wave form, combined with how easy it is to see aberrations, is a great idea. I think it's excellent, well thought out stuff. Keep up the good work!

 

[Digitalbath3737]

Quote:

Originally Posted by audiomagnate /img/forum/go_quote.gif You should have started a thread about how changing the power cord on your CD player made the most amazing improvement in sound you've ever heard through your KSC-75s. Then you'ld be feeling the love right now.

That made me spit Gatorade everywhere!

 

[wnmnkh]

How people can say that graphs do not represent the codec quality? I do not understand.

What version of the software are you using by the way?

 

[Febs]

Quote:

Originally Posted by audiomagnate /img/forum/go_quote.gif Get used to it. Anything remotely based on real science around here gets flamed. You have to understand that accurate sound reproduction is more of a religion than a science to most of these people, and you're an atheist who's telling them there is no god.

Oh, please. I am one of the biggest advocates that there is on this forum when it comes to applying scientific processes to test the differences between codecs. I reject the proposition set forth in first post in this thread that these graphs are a better way to test a codec than a listening test. That does not mean that I am rejecting "science" or the application of science to the comparison of audio codecs.

Sir Nobax says that he wants "real proof" about the differences between these codecs. The only way to establish "proof" about the quality of any one of these codecs versus any other of these codecs is first to establish that there is an audible difference between the codecs. Only then do observations about those differences have any merit.

Quote:

Originally Posted by wnmnkh /img/forum/go_quote.gif How people can say that graphs do not represent the codec quality?

Well, let me ask you this: what do these graphs tell you about how the codec sounds? Can you determine by looking at these graphs which codecs you would be able to tell from the original?

 

[zip22]

Quote:

Originally Posted by audiomagnate /img/forum/go_quote.gif Get used to it. Anything remotely based on real science around here gets flamed. You have to understand that accurate sound reproduction is more of a religion than a science to most of these people, and you're an atheist who's telling them there is no god.

i completely disagree. Filburt put it best. we are not disputing the science of it, we are disputing that this "science" is not nearly deep enough to accurately portray whether a codec does a good job reproducing a wave for human ears.