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: HELP! Please URGENT (mp3 vs cd-audio comparison) (Read 68412 times) previous topic - next topic
0 Members and 2 Guests are viewing this topic.

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #100 –
Quote
Quote
This is the problem with non DBT.  You "like" to know what you compare, which also means you probably know what you'd "like" to win as well, and thus bias is incurred.


Dear, Dibrom, for God's sake, why should I want to win?!

You misinterpreted what I said.  I did not say you wanted to win, but rather that you will likely want a particular format to win.  What I was trying to illustrate is in non-DBT tests there is an inherint bias because the listener will likely prefer some format ahead of time, at least on a subconscious level.

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #101 –
Quote
But what I appreciate, that you admit, YOU WERE CHEATING! Because it is not ORIGINAL ANYMORE. Just read your words carefully again.   You promised me 1 ORIGINAL!!! But what you sent to me, was no original wave... 

Heh.. I was trying to prevent cheating.. I don't know was it necessary or not.
But as you know, even masters are sometimes lowpassed in recodring studios at 20khz. I did a pretty standard Blackman 16384 point (iirc) lowpass in CoolEdit2Pro. So it was my original, my mastering  And no encoding noise added..

If I wouldn't have done it, it had been trivial to see in one second which is the original, using spectral view..
Juha Laaksonheimo

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #102 –
Quote
JohnV:

*argh*!

Why did you have to give out the results, I would have loved to see some ABX results.


bugdie:

We've heard excuses like yours (need a complete song for comparison, etc.) a million times before.  The fact that the sample was slightly manipulated doesn't make any difference.

Quote
We've heard excuses like yours (need a complete song for comparison, etc.) a million times before.  The fact that the sample was slightly manipulated doesn't make any difference.


As you may see, I admit, I lost. I really did. But I WAS CHEATED and this is fact, too... 

Volcano: For you Germans maybe doesn't make a difference slight manipulation, as Deutsche Telekom has also manipulated our government proimising (when they bough our telecommunication company) to invest 1 000 000 000 EURO in year 2003 and now they say, they "underestimated some factors and maybe they'll invest just 80% in 2003..." 

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #103 –
Trolling:

Does this means that the mp3 is the most transparent sample as you eliminated it immediately with the original, probably meaning that both samples sound similar to you?

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #104 –
Quote
Quote

Such settings were used which give lots of high freqs, and I did to all samples (including the original) a 20khz lowpass in CEP2pro, in order to make the spectral recognition not TOTALLY obvious.. Yes, I tried to make spectral recognition harder, so the files are not optimal quality..


Well, I lost. Sad, but true  I am not very pleased by the fact, indeed.

But what I appreciate, that you admit, YOU WERE CHEATING! Because it is not ORIGINAL ANYMORE. Just read your words carefully again.  You promised me 1 ORIGINAL!!! But what you sent to me, was no original wave... 

Heh... I think you're missing the whole point here.  He said he lowpassed all of the files at 20khz, including the original.  Thus, any deviation is normalized and is not going to really be a factor in determining a difference between the files (they are all equal in this regard).  Besides, it's very highly doubtful that you would be hearing the 20khz part anyway.  You still haven't provided any ABX results, so it's also called into doubt whether you even heard any difference at all.  I don't really see what you're complaining about.

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #105 –
Quote
As you may see, I admit, I lost. I really did. But I WAS CHEATED and this is fact, too...  :lol:

lol.. if that makes you feel better..  I think you know the real effect of this so called "cheating" though to your testing..  But if it makes you happy.. 
Juha Laaksonheimo

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #106 –
Quote
Heh... I think you're missing the whole point here.  He said he lowpassed all of the files at 20khz, including the original.  Thus, any deviation is normalized and is not going to really be a factor in determining a difference between the files (they are all equal in this regard).

That is correct.
Juha Laaksonheimo

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #107 –
Quote
You misinterpreted what I said.  I did not say you wanted to win, but rather that you will likely want a particular format to win.  What I was trying to illustrate is in non-DBT tests there is an inherint bias because the listener will likely prefer some format ahead of time, at least on a subconscious level.


Huh... now I see 

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #108 –
Quote
Volcano: For you Germans maybe doesn't make a difference slight manipulation, as Deutsche Telekom has also manipulated our government [...]

Why don't you just pack it in, get some samples, and provide us some ABX results. You seem so confident that you can hear a difference, so there should be no problem to prove it. Right?

Usage of a complete song for direct AB or ABX comparison is nonsense (that's another one of those typical statements you see from "audiophiles" BTW). No brain is capable of remembering exactly enough how 5 minutes of music sound to be able to tell where there were artifacts, especially if they were only subtle. 20 second clips are the only way to pick up subtle differences reliably.

If you still insist on using a complete song, go ahead and do it - you won't do yourself any favour though. (Another problem is that a complete song cannot be uploaded for us guys to verify, for copyright and bandwidth reasons.)

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #109 –
It would have been more rigorous to lowpass the original before encoding, although I interpret the result as "no clear difference was heard".
This has happened to me quite often doing some ABX blind tests... sure of myself about what I hear, and when I click the "test over" button of the ABX program... all wrong !

Fortunately, I had once the opposite experience : I didn't think I succeeded, so alike were the samples, and in fact I got 15/16 in ABX    It was Fronhaufer 192 CBR stereo vs original.

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #110 –
Quote
Why don't you just pack it in, get some samples, and provide us some ABX results. You seem so confident that you can hear a difference, so there should be no problem to prove it. Right?

Problem is, can you trust the ABX results of a person, who desperately wants to prove that "mp3 sounds like crap"? No offence to budgie, but imo the test like I did is the only way to prove to people like budgie, that not all lossy audio sound can be instantly recognaized.

The sample was very hard to encode. Very strong stereo separation and lots of high frequency cymbals which is usually very hard for lossy audio, and also few sharp attacks, a bit guitar and piano.
Juha Laaksonheimo

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #111 –
This answer goes mainly to JohnV and Dibrom:

Once again, I admit, I lost. I let myself get fooled... It doesn't make me very happy...  But I won't really commit suicide due to this sad fact...  Anyway, I would prefer more correct test, i.e. without any kind of adjusting/arranging the test samples. And I would rather prefer more complex and longer song. That's all... nothing more and nothing less. And that's no elusion or excuse from my side. I am old enough to acknowledge I swallowed a bait. I should have been more cautious... 

As I noticed, there was used obviously good tweaked loss compression utilities/programs, so I would test some of them this weekend. Just to feel if I could live comfortably with some of them, when needed... Internet is a real deep source, so it should be no problem.

Have a nice day!

P.S. If you could explain me just shortly how to make a correct blind test, maybe I would try it, too...

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #112 –
Quote
It would have been more rigorous to lowpass the original before encoding, although I interpret the result as "no clear difference was heard".

I tried this at first, but it didn't provide smooth hf response in specral view, making again the recognition with spectral view totally trivial.

Now that I think about it, I shold have made the lowpass somehow "uneven", so that it's not so smooth in spectral view and enocde with that. Then it's maybe not so easy to see which is encoded and which is the original. Maybe should have cutted the top portion of freqs from some lossy encode, add it to the original so that the 20khz area is uneven in spectral view, and make it the original, then use it to encode with lossy..
Juha Laaksonheimo

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #113 –
Quote
Problem is, can you trust the ABX results of a person, who desperately wants to prove that "mp3 sounds like crap"? No offence to budgie, but imo the test like I did is the only way to prove to people like budgie, that not all lossy audio sound can be instantly recognaized.

You're right of course, but you always take that risk when you ask "strangers" (couldn't find any better word...) for ABX results. People participating in ff123's listening tests can do it, everybody can.

If ABX results are provided and they seem odd, there's still the possibility to conduct a test like you did which males cheating near impossible.

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #114 –
Quote
P.S. If you could explain me just shortly how to make a correct blind test, maybe I would try it, too...

See ff123's site (I linked to it in my previous reply to you), there are instructions on how to use the ABC/HR tool.

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #115 –
Quote
Usage of a complete song for direct AB or ABX comparison is nonsense (that's another one of those typical statements you see from "audiophiles" BTW).

I wouldn't be so categoric. Though controversial, Oohashi's experiment about utrasounds did use long samples, and, related to audition or not, each time, it took some tens of seconds for the EEG to react.
I'm not saying it works really this way, but there is no evidence of the absence of long term listening effects.
Also, I don't know if you remember the "16 bits dithered vs truncated" listening challenge. The first people to manage the ABX did perform one listening session per day ! In this perspective, long samples can be considered.

But I agree that for compresion artifacts, short, or even very short samples are easier, at least for me.

I've got the link to Oohashi's at home, I'll add it tonight. I lost the link to the 16 bits challenge.

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #116 –
Quote
Problem is, can you trust the ABX results of a person, who desperately wants to prove that "mp3 sounds like crap"? No offence to budgie, but imo the test like I did is the only way to prove to people like budgie, that not all lossy audio sound can be instantly recognaized.


I don't desperately want to prove MP3 is CRAP... Maybe my wrong experiences with MP3, there was Xing involved and Blade or so... And as I know people are satisfied with 128 kbps or lower bitrates on very, but really very cheap equipment, has led me to a little bit being big-mouthed. But I really don't want anything to prove or so. I really don't need it, believe me. 

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #117 –
Quote
I tried this at first, but it didn't provide smooth hf response in specral view, making again the recognition with spectral view totally trivial.

I think Budgie can be trusted. No need to lowpass anymore.

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #118 –
Quote
As I noticed, there was used obviously good tweaked loss compression utilities/programs, so I would test some of them this weekend. Just to feel if I could live comfortably with some of them, when needed... Internet is a real deep source, so it should be no problem.

Excellent. Your words of wiscom sound like sweet music to my ears 
Well... of course - as you certainly guessed - JohnV didn't choose the worst encoders and settings...    By the way I'm sure you'll actually find an interesting selection in his post !
One encoder that you might want to try, is MPC (aka Musepack). A bit like Vorbis and AAC it has been designed for quality from the ground-up, and is very well-tuned, fast and solid. Unfortunately it's based on some MPEG1 Layer2 (=MP2) mechanisms, so it's not clear whether it's free from patents or not.
Anyway it's open-source now. (btw, JohnV, are the licensing terms now settled for the encoder part?)

For now, Musepack has the edge when targeting for transparent quality. And it won't smear even the sharpest transients :-)

Quote
I don't desperately want to prove MP3 is CRAP... Maybe my wrong experiences with MP3, there was Xing involved and Blade or so... And as I know people are satisfied with 128 kbps or lower bitrates on very, but really very cheap equipment, has led me to a little bit being big-mouthed. But I really don't want anything to prove or so. I really don't need it, believe me.

Good point !  IMHO here's the worst weakness of MP3:  good potential, good encoders but careless people who are so tempted to use either (or both) of:
- a fancy-looking, proprietary, bloated all-in-one interface with pretty crappy mp3 encoder 
- a fixed  bitrate and - very often - too low bitrates for the music to be treated well... 
- the default settings of the encoder (which, except for MPC/Vorbis, is quite low quality)..

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #119 –
Quote
Now that I think about it, I shold have made the lowpass somehow "uneven", so that it's not so smooth in spectral view and enocde with that. Then it's maybe not so easy to see which is encoded and which is the original. Maybe should have cutted the top portion of freqs from some lossy encode, add it to the original so that the 20khz area is uneven in spectral view, and make it the original, then use it to encode with lossy..

After number 4, number 3 looked closest to what I expect the original to look like in spectral view.  But 3 had some very small dropouts in it which number 4 didn't have.  Yeah, I think you probably could have fooled me if you had cut and pasted the tops from one of the others onto the pre-encoded file to make it the new original.  But I think people could have legitimately cried "Foul!" then.  Post lowpassing at 20 kHz is probably safe enough, though.

BTW, I could not distinguish any of these from each other by just listening.

ff123

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #120 –
Quote
Very difficult.
First, the information over 20 kHz was not artificial but naturally recorded at high sampling rates.
And the most tricky part : the playback did use bi-amplification to avoid intermodulation distortion under 20 kHz.

Yeah, I know. However, I want to try first if I can hear any difference on long-term listening, even when the conditions are not the ideal ones. It's only a home-based approach to the experiment, just to see what happens. Depending on the results, I'll check the HF intermodulation of my equipment.

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #121 –
OK, I see the discusion is going on. As a starter of this thread I would like to inform everyone that tonight I recieved 2 CDs:

1. JM Lab - Le Grand Spectacle De Son

2. The Greatest Hits of Sting: Fields of Gold 1984 - 1994

The second one is probably well known to most of people here.  The first one is mix of classic and acustic music. I was told that this CD comes with these high quality speakers (JM something???). This is the one that worries me.

For testing purporse I plan to encode both of them with lame 3.92 --alt-preset insane. Should I use the -k switch to preserve the highest frequencies? Excuse my ignorance... Please someone explain if using of the  -k switch is useful for this?

Details of the test:
Two persons agreed to take the test, not one as said before.
The two persons will test the CDs separately.  Person A will test the first CD and person B will test the second CD (see above CDs).

For person A and B I will prepare two CDs. The first one will be mp3 only (this is the one the bet is about) and on the second one I can use any other format (lossless or lossy, i.e. mp3, ogg, mpc, monkey,....), BUT the second one is not bet.

For the first CD I will select some tracks and copy them directly from wav (ripped with EAC). For the same tracks I will record MP3 alt preset insane tracks. All mixed to be compared. Not all tracks from the original CD will fit, but I suppose that doesn't matter.

On the second CD, which is not part of the bet, I will use for person A (test with JM lab CD) maximum possible quality (he's supposed to be "the big ears") with lame, ogg, mpc aac, monkey and maybe others.

On the second CD ( for the person B, who is not "the big ears"), I will use lame only, but mixed with lower quality parameters. I will also "plant" a track or two with really "lame" settings, just to make sure he finds something (I hope).

Thats all for now,

Thank you all for participating,

qshtr

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #122 –
Quote
Should I use the -k switch to preserve the highest frequencies? Excuse my ignorance... Please someone explain if using of the   -k switch is useful for this?

No, don't use -k. --alt-preset insane gives over 20kHz response anyway.
Juha Laaksonheimo

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #123 –
Quote
The first one is mix of classic and acustic music. I was told that this CD comes with these high quality speakers (JM something???). This is the one that worries me.

JM Lab speakers... the factory is 500 meters away from my office 

Those kind of CDs often feature audience clappings recorded especially from inside the audience  you may be in bad trouble !

HELP! Please URGENT (mp3 vs cd-audio comparison)

Reply #124 –
Quote
Quote
So that TV tone is 19khz?  I can hear that clearly, even a few rooms away (tells me someones watching the tube  )   Guess loud music hasn't totally destroyed my ears... yet.

Eeh, wasn't it something like 16kHz? Also tv-sets can give many kind of high pitched sounds.. I hear different hf sounds pretty much from every tv.. 

Man..  budgie sure can stir up stuff...

Someone's probably already done it, (I'm far from reading to the end of thread yet) give him the link to the ABX software and let him do true double blind tests on his own choices of music..

BTW on the TV thing.. in the US,  the horizontal scanning frequency is 15.75Khz (NTSC video)..  in Europe and other parts of the world,  where PAL is used,  I think it's a bit higher,  (too lazy now to look it up) but I don't think it's 19Khz...

Interestingly awaiting the Audiophile Showdown Bet results..    I want to do something like this with a friend who has a $$$$$ system, and considers himself a Real Audiophile...  he is in the "vinyl sounds so much better than those inferior CD's" crowd..  It would be interesting to see what he thinks of well made mp3's, ogg's, and MPC's..

Jon