Hifidelio-User.de

Das (inoffizielle!) Forum für Hifidelio-User
Aktuelle Zeit: 14.12.2017 15:12

Alle Zeiten sind UTC + 1 Stunde [ Sommerzeit ]




Ein neues Thema erstellen Auf das Thema antworten  [ 23 Beiträge ]  Gehe zu Seite 1, 2  Nächste
Autor Nachricht
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 14:17 
Offline
Moderator

Registriert: 27.01.2005 21:11
Beiträge: 2436
Vincent hat geschrieben:
Funny. you talk about "hochwertige" and you want to play MP3 in 192 kbs?

Why not? Ever did a test by yourself? I did. More than 128kps are more or less useless for my ears.

And I'm not alone, very few people seem to have "golden ears". Look at the results of this public test for example:
http://www.audiohq.de/index.php?showtopic=1436, Detailed results are here: http://www.listening-tests.info/mf-128-1/results.htm

128kbps mp3 encoded with lame were tested with a score of 4.6.
A score of 5 means "transparent", "no difference to original"
A score of 4 means "difference noticeable, but not unpleasant"


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 14:27 
Offline
Moderator
Benutzeravatar

Registriert: 29.01.2005 00:43
Beiträge: 1750
Wohnort: CH, Bern
gigi hat geschrieben:
Why not? Ever did a test by yourself? I did. More than 128kps are more or less useless for my ears.

I began encoding all my CDs by the end of the 90ies on a Dual Pentium Pro 200 with Fraunhofer commandline-encoder, two task on WinNT, in HQ-mode each Song took about twice the length of the song.
I've made several hearing tests and found no difference for my (non-golden) ears between CD and 112 kbit/s mp3s!
Ok, I'm shure it was not the best hifi equipment and HD space was much more expensive at this time, today I use 128kbit/s...
(for shure my ears are not better today)

Regards, Bergi

_________________
Was war ST-64 :?:


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 15:16 
Offline
Benutzeravatar

Registriert: 22.02.2007 01:03
Beiträge: 441
Of course I did.
I took several different types of music and ripped them on the HF
WAV/AIFF/FLAC makes no difference to me.

Then I ripped to the highest quality MP3 the HF sports (320 kbs I believe)
The difference in a straight A/B comparison is clearly audible.

Now listening test are highly subjective so I asked my wife to listen. She doesn't have the high end neurosis I'm prone to but for her the differences where also easy to hear.

I don't think one is in need of a delicate 'ear' to spot these differences.
It often starts with the weakest spot of any audio system: the speakers. Most speakers are not able to reproduce small details. They more or less blurr the signal.
A lot of speaker systems have a heavy bass if not booming and brilliant high and the midrange is weak.

A lot of people are simply not sensitive to sound quality. Take car stereo as an example, excessive base, over brilliant high. It sounds as unfaithful to the original as possible but a lot of people likes it. In fact you can buy home stereo systems with 'atmospheres' like 'stadium', 'concert hall' and of course "car stereo".

Now I have a question for a technical well educated guy like you.
If one removes part of the information from a redbook audio CD, how could you expect the result to sound the same?

_________________
the Rough Guide to the Hifidelio has become part of The Well Tempered Computer


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 15:29 
Offline
Moderator

Registriert: 27.01.2005 21:11
Beiträge: 2436
Vincent hat geschrieben:
Of course I did.

When listening to a sample, did you know which one was the original and which one the mp3? I assume yes.
Then such a test makes no sense at all.
The only way to do such a test is NOT to know what you are hearing.
The best practical method is an ABX-Test: http://en.wikipedia.org/wiki/ABX_test
You cannot do this autonomously, but somebody else has to provide you with the files.
Once you know what you are listening to, you WILL hear a difference, because "Hearing is believing". If you have nothing to believe, you won't hear a difference.

Zitat:
Now I have a question for a technical well educated guy like you.
If one removes part of the information from a redbook audio CD, how could you expect the result to sound the same?

That's the whole magic with all kinds of bit-reduced audio.
Read here: http://en.wikipedia.org/wiki/Psychoacoustic

If you are interested, just provide me with a sample wav file. I will try to encode it with different bit rates and decode it to wav again.
Then you hear to these WAVs and tell me which one you prefer and which is closest to the original. I'm very sure that you are not able to detect 320kbps mp3, even 192kpbs is very unlikely.


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 16:34 
Offline
Benutzeravatar

Registriert: 22.02.2007 01:03
Beiträge: 441
Of course I knew, but my wife didn't. I simply let her listen to A and then to B and asked if she hear any difference. Well she did. Then I asked her preference which was A because it sounded more natural. A being the lossless format, B the MP3
N=1 is not a sample size acceptable for the scientific community but for sure it can't be my own bias.

Hearing is believing means:
- don't care about the specs
- don't listen to the sales talk
- don't give a hood about the brand
but listen

What you describe is exactly the opposite, believing is hearing or prejudice for short

I can tell bad audio equipment from good equipment.
One day I bought my current speakers. From that time on I hear a bird when playing the second part of Beethoven's string quartet no.8
On my previous system I simply didn't hear it. That’s what good speakers do, resolving all the information.
And that’s more or less the difference between high quality MP3 and lossless formats, the micro details are gone

_________________
the Rough Guide to the Hifidelio has become part of The Well Tempered Computer


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 17:23 
Offline
Moderator

Registriert: 27.01.2005 21:11
Beiträge: 2436
Vincent hat geschrieben:
What you describe is exactly the opposite, believing is hearing or prejudice for short

Not sure if you got my point or if I got yours here.
Prejudice: Yes, you cannot eliminate ANY kind of prejudice if not doing a double-blind test. Asking your wife is NOT a double blind test.
Read here: http://en.wikipedia.org/wiki/Blind_experiment

Zitat:
I can tell bad audio equipment from good equipment.

When it comes to mp3 above 160kbps, I simply do not believe you. My offer one post above is still valid.


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 17:33 
Offline
Moderator
Benutzeravatar

Registriert: 03.04.2005 07:40
Beiträge: 1062
Wohnort: Kasel bei Trier
Vincent hat geschrieben:
That’s what good speakers do, resolving all the information
I agree with that:
The "source-material" & the speakers are the most importend part of a HiFi-Chain.
But:
...even on a good chain I have no chance hearing a difference between MP3 at 192 kbit/s or above. We made a double-blind-test with some Freaks at 320 kbit/s vs. CD (good equipment) a few years ago ... no chance at all (Pop, Classic)... I´m sorry for that ;-)
And finally:
What do you think is the reason, for that there are no really independent and controlled double-blind-tests of "High-End"-Equipment ...
I guess I know the reason ...

_________________
... zur Zeit tatsächlich ohne echte Wünsche ... d.h. wann sollte der ST-64 kommen ?!


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 18:03 
Offline
Benutzeravatar

Registriert: 22.02.2007 01:03
Beiträge: 441
Double blind is not without flaws, there is the well known placebo effect.

But I wasn't out there to perform a scientific experiment.
Before loading 400 CDs in the HF I want to make sure what format suite me best.
As the differences where audible to me, I choose a lossless format.
If you don't believe, well you don't.
Personally I would have loved hearing no differences at all as MP3 support is much better than FLAC support.

Dirk,

It’s a bit unclear what you are suggesting.
If you mean that there is a lot of humbug in the High End world, you're right
There are even claims ignoring the most fundamental laws of physics, highly amusing.
On the other hand, do you know controlled double-blind-tests of any consumer product? What about a double-blind-tests of a car?

_________________
the Rough Guide to the Hifidelio has become part of The Well Tempered Computer


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 19:08 
Offline
Moderator

Registriert: 27.01.2005 21:11
Beiträge: 2436
Vincent hat geschrieben:
If you don't believe, well you don't.

You could easily prove me wrong, but you don't want to, right?


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 19:13 
Offline
Moderator
Benutzeravatar

Registriert: 03.04.2005 07:40
Beiträge: 1062
Wohnort: Kasel bei Trier
Vincent hat geschrieben:
... What about a double-blind-tests of a car?
maybe it´s not necessary because it´s easy to check really interresting parameters like stopping distance etc. And: no one has any problems with emotions like "I prefer a mini-cooper ... I´like it more than a Porsche ..."
But there are things you cannot check by measurement like manifestation of pain or good-feeling, or detecting a water-spring by using a dowsing-rod etc.
In this case you need a double-blind-test to stay serious.
I´ve no problem with the statement:
"I prefer this monster-cable cause I´ve the impression of a better sound"
But it´s not serious to say:
"There is a better sound with this monster-cable (in this or that way) - everbody with music-used young & good ears must hear - but we avoid a double-blind test"
The very few tests in this way have been a desaster for the distributors of very expensive High-End-Products.
BTW: I really like fine arts in manufacturing good products.
Dirk

_________________
... zur Zeit tatsächlich ohne echte Wünsche ... d.h. wann sollte der ST-64 kommen ?!


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 21:27 
Offline
Benutzeravatar

Registriert: 22.02.2007 01:03
Beiträge: 441
Gigi,

Zitat:
You could easily prove me wrong, but you don't want to, right?


I’m afraid the experiment you propose won’t convince you as it is not double blind by design.
Otherwise it would have been quit interesting.
So if we can design an experiment which is blind for the prejudices of the both of us I’m willing to participate

_________________
the Rough Guide to the Hifidelio has become part of The Well Tempered Computer


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 22:39 
Offline
Benutzeravatar

Registriert: 22.02.2007 01:03
Beiträge: 441
Dirk,
You do have a point.
Now double blind is a bit of a problem.
If 1 out of 100 hears a difference and the other 99 don’t, does that mean there ain’t no difference? Or do we have 1 guy with excellent hearing and 99 normal people?
If 9 out of 10 hears a difference and 4 prefers system A and 5 system B, which one is best?

Famous is the Quad story. They where demonstrating somewhere and every body was impressed by their new line of speaker cables although some object against the colour (orange). The truth was they forgot to pack their cables so somebody went to a DIY shop and bought some powerline. Indeed, the Black & Decker orange coloured ones.

Personally I believe in the law of diminishing return.
Its easy to improve on low quality, its harder to improve on good quality and very hard to improve on top quality.
So you end up with the ‘cost no object’ systems. Maybe they are the best in the world but their price/performance ratio is horrible. You pay double for a (of course a highly subjective) 0.1 % improvement.

_________________
the Rough Guide to the Hifidelio has become part of The Well Tempered Computer


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 22:43 
Offline
Moderator
Benutzeravatar

Registriert: 21.05.2005 14:31
Beiträge: 2388
Wohnort: 52.429105,10.810525
Hi Vincent,

there is a solution for a double-blind test.

Me.

But before i explain it, let me explain, what the "Double-Blind-Test" normally means:

It is usually used with pharmaceutical tests. There it is a problem (or could be), if the doctor, which talks to the patient, "knows" which one is the real or the fake-medicine.

That is clearly caused by the interaction, even if he tries to "act" exactly the same, it can not been guaranteed, because he is not a machine.

And the Placebo-Effect is also a bit different, and not directly comparable to hearing-phenomens. A placebo-effect is measurable, statistically relevant more patient get better, while thinking "getting a new, strong, or whatever medicine".

At hearing, there is no such effect, which you can measure.


And the "double-blind-test":

If you dont know, which file comes with which file-format: the only thing, which can influence you, is the order of the files and its names.

There i come into the game:
I get the files from gig, rename them by "my order" - also not knowing, which files containing which sort of compress-effects, and then you get them.

You post your results, i post my "rename-order" - and gigi solves the test.

Ok, everybody can fake ... but that is the only risk.

Lets do it!

Space


Let

_________________
nope!


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 10.12.2007 22:52 
Offline
Moderator

Registriert: 27.01.2005 21:11
Beiträge: 2436
Spacesson hat geschrieben:
Ok, everybody can fake ... but that is the only risk.

At least i could not fake it. After the test (not only with vincent, the more participants, the better!) i would tell the exact commandlines and SW-Versions of the encoders, so everybody could prove how these files were produced.


Nach oben
 Profil  
 
 Betreff des Beitrags:
BeitragVerfasst: 11.12.2007 01:05 
Offline
Benutzeravatar

Registriert: 22.02.2007 01:03
Beiträge: 441
Space,
I don’t know if this is double blind but thanks for the offer anyway.

Lets try to formulate an experiment acceptable to all parties involved.

We all agree:
MP3 removes information by definition (lossy compression)
Lower the bitrate and more information is removed

We disagree:
- The difference between high quality MP3 and the lossless original is audible
- MP3 removes information in such a clever way that there is not audible difference with the lossless original when using a bitrate of 320
- Likewise at 192
- Anybody in for something lower?

We agree: below a certain level differences become audible

So what we need is a lossless track.
Everybody involved should have this track and know this is the reference
Then we need the same track in
WAV
MP3 320
MP3 192
MP ???

All MP3’s will be converted to WAV
This of course won’t restore the original (what is removed, is removed permanently) but at file format level there is no visible information about the ‘real’ format used.

Testing:
Compare the testing “WAVs” with the original.
Tell us which one is the lossless one

Testing tracks:
- a hit from the sixties, so a low fi recording
- a symphony orchestra
- pop music with a well recorded bass (not booming)

After the testing Gigi will provided the information enabling all parties involved to replicate the testing tracks.

_________________
the Rough Guide to the Hifidelio has become part of The Well Tempered Computer


Nach oben
 Profil  
 
Beiträge der letzten Zeit anzeigen:  Sortiere nach  
Ein neues Thema erstellen Auf das Thema antworten  [ 23 Beiträge ]  Gehe zu Seite 1, 2  Nächste

Alle Zeiten sind UTC + 1 Stunde [ Sommerzeit ]


Wer ist online?

Mitglieder in diesem Forum: 0 Mitglieder und 1 Gast


Du darfst keine neuen Themen in diesem Forum erstellen.
Du darfst keine Antworten zu Themen in diesem Forum erstellen.
Du darfst deine Beiträge in diesem Forum nicht ändern.
Du darfst deine Beiträge in diesem Forum nicht löschen.

Suche nach:
Gehe zu:  
Powered by phpBB® Forum Software © phpBB Group
Deutsche Übersetzung durch phpBB.de