who's making money off this? i can't say i full understand the problem with test protocols and how it affects results but even with all the flaws, did it not show that "despite" all the flaws and inadequate equipment (like my own soundcard) some participants could still perceive a difference?
no smoking guns! no absurd proclamation where made! so why all the backlash?
if the test protocols are too strict or the equipment requirements to specific i think it would have shrunk the already small number of participants, no?
i got to test my own ability and discovered my soundcard is crap.
no smoking guns! no absurd proclamation where made! so why all the backlash?
if the test protocols are too strict or the equipment requirements to specific i think it would have shrunk the already small number of participants, no?
i got to test my own ability and discovered my soundcard is crap.
Grampa's ears ain't that useless for sines. Toshiba L645 and $12 Philips IEM.
Code:
foo_abx 2.0.5 report
foobar2000 v1.3.12
2018-12-05 18:01:33
File A: sig1.wav
SHA1: 621ecd158928499510d523dc5a85abd019cb1d44
File B: siga.wav
SHA1: 640783db9340fde87db66db7127031862c1d8814
Output:
WASAPI (push) : Speakers (Conexant SmartAudio HD), 16-bit
Crossfading: NO
18:01:33 : Test started.
18:02:20 : 01/01
18:02:39 : 02/02
18:03:06 : 03/03
18:03:34 : 04/04
18:04:11 : 05/05
18:04:42 : 06/06
18:05:05 : 07/07
18:05:24 : 08/08
18:05:41 : 09/09
18:05:57 : 10/10
18:06:14 : 11/11
18:06:31 : 12/12
18:06:42 : 13/13
18:06:59 : 14/14
18:07:11 : 15/15
18:07:24 : 16/16
18:07:24 : Test finished.
----------
Total: 16/16
Probability that you were guessing: 0.0%
-- signature --
6d82a62d90f1fa79d60de220a7eb3bd4e27ceb5e
The question here is if foobar2000 is a bad player. If yes, the ABX tests using different equipment/systems are not reliable.
Maty - please stop with your endless Foobar post unless you are ready to show some proof. I've asked you for it, Pavel has asked you for it. You have ignored those requests. Until you have some proof, stop posting the same thing over and over again. It is to no one's benefit.
No it is not. It yields perfect data, bit-perfect data and repeatable results in both 16-bit and 24-bit data. The problem, if any, is in users or their systems or inability to set drivers and foobar properly.
Myself is also JR 64bit user and fan but think you should relax a bit because JR can also output less good objective sound stream without the user knowing about it, agree PMA that every particular PC setup needs attention and measurements to get it right before using our ears.
I stop participating in your tests because I have already made it clear that I consider this player unreliable.
If I have been able to optimize again and again my Windows to improve the sound quality something I should know, I say.
Less Roon, JPlay and HQplayer I have tried all the software players available in windows and linux (of which I have proof).
- The End -
If I have been able to optimize again and again my Windows to improve the sound quality something I should know, I say.
Less Roon, JPlay and HQplayer I have tried all the software players available in windows and linux (of which I have proof).
- The End -
Last edited:
That sounds like another cop out, like others that slag off the test but miss the point entirely
Instead of waving hands, could you show a loopback measurement with sig1 sine played once in foobar directly and then through ABX plug-in? The way I have it attached. And make sure you have set a proper Fs and resolution in Direct sound for your soundcard in the Control panel. If you do not do it, then there is no advice.
Attachments
Was conclusions there based only on subjective tests or did you take time for every each setup string to measure hole chain, and also is any pre and post tweaking backed up by objective data so as to be sure one is not listening to any flaws/distortion that can end happen unintended please our ears.
Yes, under -120dB, which is most probably a difference/fluctuation in FFT analysis.
I have prepared the same sine files in 44.1/16 format, to be compatible with avocado/apricot, so if you want, please try again.
http://pmacura.cz/sines44.zip
They are the sine1 and sinea resampled to 44.1kHz with noise-shaped dither.
I have prepared the same sine files in 44.1/16 format, to be compatible with avocado/apricot, so if you want, please try again.
http://pmacura.cz/sines44.zip
They are the sine1 and sinea resampled to 44.1kHz with noise-shaped dither.
That's not the impression I picked up from your other posts on this topic
So much misinformation - I don't know if it is intentional or just not interested in reading the documents linked to many times before to inform yourself before posting such misinformation?
I'm sorry if you haven't read the guidelines on how to approach blind testing posted many times before or that you don't understand that your test was invalid which you would have discovered early on if following the recommended procedures in those guidelines - before running your invalid test. As PMA says, he considers only one test valid so what have you learned?
Is that what this listening test set out to do? What therefore was the point of the listening test? What would have been the outcome if PMA wasn't challenged to post sine wave test signal - would your & others test results have been considered valid?
Last edited:
The null-hypothesis is represented by H0 = 0.5 (perfectly random guessing implicated), as it is a difference test the alternative hypothesis (quite unspecific) would be H1 = p > 0.5 .
So the usual analysis looks at the observed data and calculates the probability to reach such a result per random guessing. If this probability is low we take it as a hint that occurence of the observed data is less compatible to our null-hypothesis, so there might be a real effect.
Replications are needed.
Of course, you are right, if there is a real effect it could well be that it is equipment dependent and listener dependent too.
But, as we have the actual test results from every listener we already have some data to evaluate the possible divergence.