IPB

Welcome Guest ( Log In | Register )

19 Pages V  « < 10 11 12 13 14 > »   
Reply to this topicStart new topic
Jplay - just another scam? YES IT IS!
Takla
post Apr 2 2013, 02:15
Post #276





Group: Members
Posts: 169
Joined: 14-November 09
Member No.: 74931



QUOTE (Wombat @ Apr 1 2013, 17:42) *
Are you seriously asking for some abx log? Just replaygain both files and listen. The voice coming in on second 9 sounds completely different!


Are claims of difference only required to be substantiated if asserted by persons that the peer group doesn't much like? TOS#8 does use the word "must" but in practice there often seems to be an invisible rider "unless you're confirming our expectations or bashing someone we don't like".

I would be surprised if Wombat's assertion was proven to be false, but perhaps applying the terms of service in a non-discriminatory manner would prevent threads veering off topic (it broke my remote desktop and ate my hamster, therefore it is evil and evil stuff sounds bad... the same ...or different ...or something) and also reduce the evident temptation to just join the queue, pick up a rock and throw it (he said jehovah!).
Go to the top of the page
+Quote Post
greynol
post Apr 2 2013, 02:25
Post #277





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



QUOTE (Takla @ Apr 1 2013, 18:15) *
I would be surprised if Wombat's assertion was proven to be false

Wombat, please provide an ABX log in order to validate Takla's suspicion.

I won't dignify the rest of the ironical post.


--------------------
Placebophiles: put up or shut up!
Go to the top of the page
+Quote Post
Wombat
post Apr 2 2013, 02:48
Post #278





Group: Members
Posts: 978
Joined: 7-October 01
Member No.: 235



Did anyone even bother to listen these samples besides dhromed and has the ability to qualify it?
Everybody should be able to abx this with his PC-Speakers as i did now, not even Headphones needed.

I needed to cut 1.38 seconds from the foobar sample and 92ms from the Jplay sample to have similar beginnings to prevent me from cheating. Replaygained foobar test:
CODE
foo_abx 1.3.4 report
foobar2000 v1.1.16
2013/04/02 03:42:31

File A: C:\Brezel Jplay\testfoo.flac
File B: C:\Brezel Jplay\testJ.flac

03:42:31 : Test started.
03:42:44 : 01/01  50.0%
03:42:54 : 02/02  25.0%
03:43:04 : 03/03  12.5%
03:43:14 : 04/04  6.3%
03:43:24 : 05/05  3.1%
03:43:37 : 06/06  1.6%
03:43:46 : 07/07  0.8%
03:43:55 : 08/08  0.4%
03:44:03 : 09/09  0.2%
03:44:10 : 10/10  0.1%
03:44:14 : Test finished.

----------
Total: 10/10 (0.1%)


This post has been edited by Wombat: Apr 2 2013, 02:49
Go to the top of the page
+Quote Post
dhromed
post Apr 2 2013, 09:48
Post #279





Group: Members
Posts: 1286
Joined: 16-February 08
From: NL
Member No.: 51347



QUOTE (sabrehagen @ Apr 2 2013, 00:06) *
If then, JPLAY's result has a difference in volume, we may never be able to get accurate tests.


The difference is in EQ, not merely volume. In fact there is so much EQ going on that it's next to impossible to accurately level match them.


Wombat, do you agree the jplay sample has much reduced bass and treble?
Go to the top of the page
+Quote Post
Nessuno
post Apr 2 2013, 10:25
Post #280





Group: Members
Posts: 422
Joined: 16-December 10
From: Palermo
Member No.: 86562



QUOTE (Porcus @ Apr 1 2013, 20:51) *
QUOTE (Nessuno @ Apr 1 2013, 20:19) *
all the "tricks" they say they play to the system to gain resources, like run Jplay as a service, modify scheduler's queues or priority, shutdown other services etc...


That fine-tuning is something I could actually wish for ...

In order to improve SQ? And even if, have you lost your PC Admin passwd?


--------------------
... I live by long distance.
Go to the top of the page
+Quote Post
Porcus
post Apr 2 2013, 10:51
Post #281





Group: Members
Posts: 1842
Joined: 30-November 06
Member No.: 38207



QUOTE (Nessuno @ Apr 2 2013, 10:25) *
QUOTE (Porcus @ Apr 1 2013, 20:51) *
QUOTE (Nessuno @ Apr 1 2013, 20:19) *
all the "tricks" they say they play to the system to gain resources, like run Jplay as a service, modify scheduler's queues or priority, shutdown other services etc...


That fine-tuning is something I could actually wish for ...

In order to improve SQ? And even if, have you lost your PC Admin passwd?


In order to prevent stuttering/dropouts on old cheap hardware. Hardware-upgrading a silent PC could easily cost more than 99 currency units. fb2k does offer provisions to set priority to the playback chain, but for the computer I just tried to get working, that is actually not enough.


--------------------
One day in the Year of the Fox came a time remembered well
Go to the top of the page
+Quote Post
Porcus
post Apr 2 2013, 11:05
Post #282





Group: Members
Posts: 1842
Joined: 30-November 06
Member No.: 38207



QUOTE (sabrehagen @ Apr 1 2013, 23:31) *
How do I calibrate volume levels other than turning the volume knob on my DAC to what sounds about right?


In foobar2000, the ABX comparator will align volumes. (Likely less precise if the files are EQ'ed very different, though.)


--------------------
One day in the Year of the Fox came a time remembered well
Go to the top of the page
+Quote Post
bennetng
post Apr 2 2013, 11:44
Post #283





Group: Members
Posts: 224
Joined: 22-December 05
Member No.: 26587



QUOTE (Porcus @ Apr 2 2013, 17:51) *
QUOTE (Nessuno @ Apr 2 2013, 10:25) *
QUOTE (Porcus @ Apr 1 2013, 20:51) *
QUOTE (Nessuno @ Apr 1 2013, 20:19) *
all the "tricks" they say they play to the system to gain resources, like run Jplay as a service, modify scheduler's queues or priority, shutdown other services etc...


That fine-tuning is something I could actually wish for ...

In order to improve SQ? And even if, have you lost your PC Admin passwd?


In order to prevent stuttering/dropouts on old cheap hardware. Hardware-upgrading a silent PC could easily cost more than 99 currency units. fb2k does offer provisions to set priority to the playback chain, but for the computer I just tried to get working, that is actually not enough.

How old and cheap the hardware you are using? The last time I experienced stuttering/dropouts (when I believed I have correctly configured my system) was a Pentium II 400, 192MB RAM PC, playing APE files with "Extra High" compression mode in foobar2000, with SSRC turned on.

When I upgraded my system to P4 2.8G and 512MB RAM in 2003 the task above was not a problem anymore.
Go to the top of the page
+Quote Post
dhromed
post Apr 2 2013, 11:52
Post #284





Group: Members
Posts: 1286
Joined: 16-February 08
From: NL
Member No.: 51347



QUOTE (Porcus @ Apr 2 2013, 11:05) *
In foobar2000, the ABX comparator will align volumes.


Yoou mean the "use Replaygain" checkbox? I thought that meant "use the existing RG tags", not "scan both files and adjust volume"


QUOTE (Porcus @ Apr 2 2013, 11:05) *
([RG is ] Likely less precise if the files are EQ'ed very different, though.)


That is a good experiment that I will try later.
Go to the top of the page
+Quote Post
Propheticus
post Apr 2 2013, 12:22
Post #285





Group: Members
Posts: 219
Joined: 10-September 11
Member No.: 93615



Ok so i ran the ABX test with the pretzel logic files three times: Replaygained individually, Replaygained as 1 album, and without. The results were easily recognizable. So there indeed seems to have been some processing and/or a level mismatch.

CODE
foo_abx 1.3.4 report
foobar2000 v1.2.4
2013/04/02 13:06:04

File A: Z:\Downloads\testSounds\foobaroutput.wav
File B: Z:\Downloads\testSounds\jplayoutput.wav

13:06:04 : Test started.
13:07:13 : 01/01 50.0%
13:07:29 : 02/02 25.0%
13:07:41 : 02/03 50.0%
13:08:21 : 03/04 31.3%
13:08:54 : 04/05 18.8%
13:09:21 : 05/06 10.9%
13:09:37 : 06/07 6.3%
13:09:54 : 07/08 3.5%
13:10:10 : Test finished.

----------
Total: 7/8 (3.5%)

CODE
foo_abx 1.3.4 report
foobar2000 v1.2.4
2013/04/02 13:11:26

File A: Z:\Downloads\testSounds\foobaroutput.wav
File B: Z:\Downloads\testSounds\jplayoutput.wav

13:11:26 : Test started.
13:11:45 : 01/01 50.0%
13:11:56 : 02/02 25.0%
13:12:07 : 03/03 12.5%
13:12:15 : 04/04 6.3%
13:12:24 : 05/05 3.1%
13:12:35 : 06/06 1.6%
13:12:50 : 07/07 0.8%
13:13:03 : 08/08 0.4%
13:13:08 : Test finished.

----------
Total: 8/8 (0.4%)

CODE
foo_abx 1.3.4 report
foobar2000 v1.2.4
2013/04/02 13:14:07

File A: Z:\Downloads\testSounds\foobaroutput.wav
File B: Z:\Downloads\testSounds\jplayoutput.wav

13:14:07 : Test started.
13:14:22 : 01/01 50.0%
13:14:31 : 02/02 25.0%
13:14:39 : 03/03 12.5%
13:14:45 : 04/04 6.3%
13:14:58 : 05/05 3.1%
13:15:06 : 06/06 1.6%
13:15:17 : 07/07 0.8%
13:15:26 : 08/08 0.4%
13:15:28 : Test finished.

----------
Total: 8/8 (0.4%)
Go to the top of the page
+Quote Post
dhromed
post Apr 2 2013, 12:38
Post #286





Group: Members
Posts: 1286
Joined: 16-February 08
From: NL
Member No.: 51347



I have a hard time believing you actually missed one.
Go to the top of the page
+Quote Post
phofman
post Apr 2 2013, 12:40
Post #287





Group: Members
Posts: 283
Joined: 14-February 12
Member No.: 97162



QUOTE (Porcus @ Apr 2 2013, 11:51) *
In order to prevent stuttering/dropouts on old cheap hardware. Hardware-upgrading a silent PC could easily cost more than 99 currency units. fb2k does offer provisions to set priority to the playback chain, but for the computer I just tried to get working, that is actually not enough.


The dirt-cheap way of upgrading the hardware in this respect is to increase size of the playback (DMA) buffer in the playback chain. Any HW from the last 20 years wil easily handle 192/24 long-latency playback (no DSP/resampling, of course).
Go to the top of the page
+Quote Post
Propheticus
post Apr 2 2013, 12:49
Post #288





Group: Members
Posts: 219
Joined: 10-September 11
Member No.: 93615



QUOTE (dhromed @ Apr 2 2013, 13:38) *
I have a hard time believing you actually missed one.


Haha me too. It was my first time ABX'ing though. So I was still figuring out how to best compare. (i.e. A,X,B,Y or A,B,X,Y .. setting a good starting point and listening a few seconds of each seems to hit the spot)

This post has been edited by Propheticus: Apr 2 2013, 12:51
Go to the top of the page
+Quote Post
greynol
post Apr 2 2013, 15:25
Post #289





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



Anyone finding the jplay sample more revealing of fine details in the music that is not present in the other sample? We were told they really were representative of the way they sounded during the initial DBT conducted by sabrehagen. wink.gif


--------------------
Placebophiles: put up or shut up!
Go to the top of the page
+Quote Post
lvqcl
post Apr 2 2013, 16:42
Post #290





Group: Developer
Posts: 3329
Joined: 2-December 07
Member No.: 49183



I played Pretzel_Logic_Excerpt.wav with foobar2000 1.2.4 and recorded the sound with rectest.exe using WASAPI loopback device.

Bit-identical.
Go to the top of the page
+Quote Post
krabapple
post Apr 2 2013, 17:08
Post #291





Group: Members
Posts: 2181
Joined: 18-December 03
Member No.: 10538



QUOTE (Wombat @ Apr 1 2013, 13:42) *
QUOTE (Arnold B. Krueger @ Apr 1 2013, 18:29) *
Someone did a listening test?

Are you seriously asking for some abx log? Just replaygain both files and listen. The voice coming in on second 9 sounds completely different!

QUOTE (Arnold B. Krueger @ Apr 1 2013, 18:29) *
Without the file being played, it is hard to say which is more accurate.

It is pointless to judge what is more accurate because we know already by some posts that the captured VLC playback is way to much off. I still donīt see the poster having fixed that playback issue.



Folks, Arny's measurements are for foobar vs Jplay and the samples are foobar vs Jplay...why does VLC keep coming into it?

This post has been edited by krabapple: Apr 2 2013, 17:15
Go to the top of the page
+Quote Post
db1989
post Apr 2 2013, 17:18
Post #292





Group: Super Moderator
Posts: 5275
Joined: 23-June 06
Member No.: 32180



If you read previous posts, the test was originally conducted on Jplay vs. VLC, the latter of which seems to alter output in at least some default cases. I am sure you can find discussion describing the switch from VLC to foobar2000. Valid results describing double-blind comparison of Jplay to any bit-perfect second player are still missing, not very surprisingly.
Go to the top of the page
+Quote Post
Porcus
post Apr 2 2013, 17:27
Post #293





Group: Members
Posts: 1842
Joined: 30-November 06
Member No.: 38207



QUOTE (phofman @ Apr 2 2013, 12:40) *
QUOTE (Porcus @ Apr 2 2013, 11:51) *
In order to prevent stuttering/dropouts on old cheap hardware. Hardware-upgrading a silent PC could easily cost more than 99 currency units. fb2k does offer provisions to set priority to the playback chain, but for the computer I just tried to get working, that is actually not enough.


The dirt-cheap way of upgrading the hardware in this respect is to increase size of the playback (DMA) buffer in the playback chain. Any HW from the last 20 years wil easily handle 192/24 long-latency playback (no DSP/resampling, of course).


... tried this while fb2k is also monitoring a media library of a hundred thousand items?

(I'm trying to get a Dell Mini 9 (Intel Atom-based, SSD, totally fanless) working. It takes six seconds performing a simple search. Longer buffer even caused stuttering at track change with some fb2k versions around 1.2. Certainly not much hope of getting foo_hdcd working, that's for sure ... but now I am drifting way off topic.)

This post has been edited by Porcus: Apr 2 2013, 17:32


--------------------
One day in the Year of the Fox came a time remembered well
Go to the top of the page
+Quote Post
krabapple
post Apr 2 2013, 18:04
Post #294





Group: Members
Posts: 2181
Joined: 18-December 03
Member No.: 10538



I've run an EQ comparison of the two uploaded samples versus Pretzel_Logic.wav as the reference audio. The two samples (3 Copy.wav , which was foobar, and 4 Copy.wav , which wasJplay) were first scanned with replaygain, which found Jplay to be 3.6dB lower than foobar. The .wav samples (and the reference .wav) were then scanned using Audition 1.0's Frequency Analysis tool, using a Blackmann_Harris FFT widnows size of 4096. . Data (average level in dB at each frequency sampling point) were dumped to Excel, and the difference between sample and reference was calculated for each channel of each sample. Difference values for Jplay were adjusted upward by 3.6 dB as per replaygain. Difference values from 20Hz-22 kHz were then plotted in Excel. Showing the left channel only.

This is a difference graph, where the Pretzel_Logic.wav file is the reference. A bit-perfect output of the reference would be a straight line overlaid on the X axis. Output that was merely raised or lowered in overall level compared to reference would show a straight line above or below the X axis. Clearly neither of these is a bit perfect output of the original, nor are they simply raised or lowered in overall amplitude. Both have considerable, and similar EQ added. The EQ is not quite identical for both, as the two curves do not completely overlap even with level-matching...however, this could be because the two samples are not exactly the same snippet of the reference. If someone wants to edit them to match, and post them, I will re-run the analysis.

In any case I conclude that saberhagen has not configured his system, or his players, for bit-perfect output, based on the shape of these curves. And foobar2k certainly can be so configured on all my PCs -- I wouldn't be able to stream DTS or AC3 from them otherwise.


The graph can be viewed here:
http://www.hydrogenaudio.org/forums/index....howtopic=100229

This post has been edited by krabapple: Apr 2 2013, 18:14
Go to the top of the page
+Quote Post
phofman
post Apr 2 2013, 18:17
Post #295





Group: Members
Posts: 283
Joined: 14-February 12
Member No.: 97162



QUOTE (Porcus @ Apr 2 2013, 18:27) *
... tried this while fb2k is also monitoring a media library of a hundred thousand items?

(I'm trying to get a Dell Mini 9 (Intel Atom-based, SSD, totally fanless) working. It takes six seconds performing a simple search. Longer buffer even caused stuttering at track change with some fb2k versions around 1.2. Certainly not much hope of getting foo_hdcd working, that's for sure ... but now I am drifting way off topic.)


Well, clogging your computer with maintaining library of that size is hardly a question of playback hardware requirements.
Go to the top of the page
+Quote Post
bennetng
post Apr 2 2013, 18:19
Post #296





Group: Members
Posts: 224
Joined: 22-December 05
Member No.: 26587



sabrehagen,

I would like to assist you to do a recording test, please provide the information below:
1. Your OS version
2. The recording software you are using

QUOTE
I was using a fresh install of foobar, having never used the software in my life before. No DSP, no nothing, just a stock install. I connected the output of my DAC to the line in on my on board sound card and recorded it there.

3. Does it mean onboard sound digital output> DAC digital input> DAC analog output > onboard sound analog input?

http://www.hydrogenaudio.org/forums/index....howtopic=100230
Please do the recording test again, using the same procedure as your Pretzel Logic test with the files I uploaded above. I need the files recorded in both 44khz and 48khz, that means you need to record 1644sweep.wv in 44khz and 48khz, and record 1648sweep.wv in 44khz and 48khz, using foobar without Jplay and foobar with Jplay.

This post has been edited by bennetng: Apr 2 2013, 18:28
Go to the top of the page
+Quote Post
Gecko
post Apr 2 2013, 18:26
Post #297





Group: Members
Posts: 937
Joined: 15-December 01
From: Germany
Member No.: 662



Thanks for the graph krabapple! Given the magnitude of difference between JPlay and foobar, it is not surprising that people are able to ABX.

Maybe some gain-aware loudness function from the soundcard/driver was active?

The difference between the original and the JPlay sample is also clearly audible (replaygain for level matching, manually aligned).

CODE

foo_abx 1.3.4 report
foobar2000 v1.2.2
2013/04/02 19:08:26

File A: E:\tmp\4_Copy_jplay.wav
File B: E:\tmp\Pretzel_cut.wav

19:08:26 : Test started.
19:08:39 : 01/01 50.0%
19:08:43 : 02/02 25.0%
19:08:48 : 03/03 12.5%
19:08:55 : 04/04 6.3%
19:09:01 : 05/05 3.1%
19:09:09 : 06/06 1.6%
19:09:18 : 07/07 0.8%
19:09:25 : 08/08 0.4%
19:09:35 : 09/09 0.2%
19:09:46 : 10/10 0.1%
19:09:50 : Test finished.

----------
Total: 10/10 (0.1%)
Go to the top of the page
+Quote Post
dhromed
post Apr 2 2013, 20:28
Post #298





Group: Members
Posts: 1286
Joined: 16-February 08
From: NL
Member No.: 51347



I can't abx the jplay sample (4_Copy) and the Pretzel logic excerpt. Maybe my ears are waxed up.

I manually edited the RG values to +3 and -7 respectively, at which point they sounded pretty much the same to me. A true RG scan came up with +8 and -0.49. Not bad for earballing it, I guess? smile.gif

The difference between 4_copy and 3_copy is much greater than 4_copy and Excerpt.

The Excerpt has its phase inverted-- at least that's what I see in Audacity.

QUOTE
the two samples are not exactly the same snippet of the reference.

the jplay sample lacks a drumroll at the start, and the very first portion of a crash hit, which could be crucial.

This post has been edited by dhromed: Apr 2 2013, 20:28
Go to the top of the page
+Quote Post
krabapple
post Apr 2 2013, 21:03
Post #299





Group: Members
Posts: 2181
Joined: 18-December 03
Member No.: 10538



QUOTE (Gecko @ Apr 2 2013, 13:26) *
Thanks for the graph krabapple! Given the magnitude of difference between JPlay and foobar, it is not surprising that people are able to ABX.


You mean the differences in bass and midrange frequencies, compared to the reference, when the samples are level matched? Because the Foobar and Jplay samples are otherwise pretty objectively similar from about 3kHz on up, when level matched this way. And as I noted, we still need to rule out the effects of the two samples not being from exactly the same section of the reference track. I expect that effect to be small -- the measurements are an average over the whole length of the section -- but I won't rule it out.



Go to the top of the page
+Quote Post
Gecko
post Apr 2 2013, 21:27
Post #300





Group: Members
Posts: 937
Joined: 15-December 01
From: Germany
Member No.: 662



Yes, basically I mean the difference between the red and black line. Eyeballing the graphs, there's a ca. 6dB drop in the 1kHz range over at least 2 octaves and ca. 3dB boost in the bass. IMO that's quite a lot and as others have demonstrated, this is audible. I'm too lazy to time align the samples so no claims about audibility from me.
Go to the top of the page
+Quote Post

19 Pages V  « < 10 11 12 13 14 > » 
Reply to this topicStart new topic
2 User(s) are reading this topic (2 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 26th July 2014 - 14:10