IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
How to match volume for ABX 5.1 to stereo, What to do with rear speakers
ktf
post Apr 2 2013, 22:10
Post #1





Group: Members
Posts: 398
Joined: 22-March 09
From: The Netherlands
Member No.: 68263



Hi all,

A while ago I bought a cheap 5.1 surround set (Logitech Z906), mostly to check how much added value multichannel audio has. This I wanted to know, because I'm planning to spend a lot of money buying a nice set of speakers, so I wanted to know whether I should buy two or five speakers smile.gif

However, I've ran into problems. I was trying to 'AB' multichannel audio to stereo audio by just looking whether I could correctly guess which was A and which was B, so omitting the X part of the test, because I haven't installed any special ABX tools. For testing this, I took the multichannel and stereo mixes of the audio-DVD of The Lord of the Rings: The Return of the King (Complete Recordings), because it was the only multichannel source I had. I correctly identified 8/8 of some tracks and couldn't on some others, but I wasn't sure whether this has been because of the multichannel vs stereo part or because of level matching.

So my questions to you:
1) Is there an ABX tool with which I can do this (so it has to be compatible with multichannel audio) and which runs on Linux or Wine
2) How should I level match the stereo with the multichannel? I ask this because there are a few buttons on the remote of the surround set to modify the volume of the rear- and center channels, I suspect setting them to the default setting might not be neutral. Any thoughts?
3) Are there other tools that enable me to assess the added value without adding expectation bias?


--------------------
Music: sounds arranged such that they construct feelings.
Go to the top of the page
+Quote Post
dc2bluelight
post Apr 16 2013, 14:25
Post #2





Group: Members
Posts: 83
Joined: 16-June 11
Member No.: 91562



QUOTE (ktf @ Apr 2 2013, 16:10) *
Hi all,

A while ago I bought a cheap 5.1 surround set (Logitech Z906), mostly to check how much added value multichannel audio has. This I wanted to know, because I'm planning to spend a lot of money buying a nice set of speakers, so I wanted to know whether I should buy two or five speakers smile.gif

However, I've ran into problems. I was trying to 'AB' multichannel audio to stereo audio by just looking whether I could correctly guess which was A and which was B, so omitting the X part of the test, because I haven't installed any special ABX tools. For testing this, I took the multichannel and stereo mixes of the audio-DVD of The Lord of the Rings: The Return of the King (Complete Recordings), because it was the only multichannel source I had. I correctly identified 8/8 of some tracks and couldn't on some others, but I wasn't sure whether this has been because of the multichannel vs stereo part or because of level matching.

So my questions to you:
1) Is there an ABX tool with which I can do this (so it has to be compatible with multichannel audio) and which runs on Linux or Wine
2) How should I level match the stereo with the multichannel? I ask this because there are a few buttons on the remote of the surround set to modify the volume of the rear- and center channels, I suspect setting them to the default setting might not be neutral. Any thoughts?
3) Are there other tools that enable me to assess the added value without adding expectation bias?


ABX-ing stereo to 5.1 is pointless because it's a drastic change than anyone can hear. It's absolutely and always detectable. Even if you somehow did level match (difficult because of the mixes are completely different), you'll still score 100%.

Every time the channel count is doubled or more, the difference is obvious to anyone. Mono to stereo is unmistakable. Stereo to 5.1 is also, and 5.1 to 11.1 is also. Beyond that, the differences are more vague, but we don't have much good test material beyond about 10.1. Dolby Atmos seems to be pretty obvious for people in theaters, but there's no easy comparison to 5.1.
Go to the top of the page
+Quote Post
ktf
post Apr 16 2013, 15:07
Post #3





Group: Members
Posts: 398
Joined: 22-March 09
From: The Netherlands
Member No.: 68263



QUOTE (dc2bluelight @ Apr 16 2013, 15:25) *
ABX-ing stereo to 5.1 is pointless because it's a drastic change than anyone can hear. It's absolutely and always detectable. Even if you somehow did level match (difficult because of the mixes are completely different), you'll still score 100%.

I'd like to prove that. (TOS #8) I tried some non level-matched samples and I didn't find it that easy. For film there's probably a big difference, but for music it might not be.

This post has been edited by ktf: Apr 16 2013, 15:07


--------------------
Music: sounds arranged such that they construct feelings.
Go to the top of the page
+Quote Post
ChronoSphere
post Apr 17 2013, 00:49
Post #4





Group: Members
Posts: 519
Joined: 11-March 07
Member No.: 41384



I have to agree 5.1 vs 2.0 is easily detectable, provided it's a true 5.1 version of the song and not just "upsampled" from stereo. And even then... most stereo recordings have a lot less positional information than 5.1, because when they're doing a 5.1 version, they go out of their way to introduces some special effects. At least on the 5.1 music I own, the difference is 100% clear.

Does TOS#8 apply here though? Number of channels is not really something related to quality, at least I don't see it. To me it's more of a cool gimmick.
Go to the top of the page
+Quote Post
ktf
post Apr 17 2013, 01:23
Post #5





Group: Members
Posts: 398
Joined: 22-March 09
From: The Netherlands
Member No.: 68263



QUOTE (ChronoSphere @ Apr 17 2013, 01:49) *
I have to agree 5.1 vs 2.0 is easily detectable, provided it's a true 5.1 version of the song and not just "upsampled" from stereo.

I have only two 5.1 recordings, which both are classical music. With this kind of music you're supposed to hear everything in front of you like a regular performance, so the only difference would be the center channel (so better positioning) and reverb over the side channels. No instruments panned all the way to the side channels or something, which would probably be easy to hear. Because I do orchestral recordings from time to time, I was wondering whether it is more life-like if you add the center channel and pass the reverb over the rear channels instead of over the front. What are your surround-music sources?

Anyway, that was not my question. I want to do a proper comparison and am looking for tools to level match this stuff. Any ideas?


--------------------
Music: sounds arranged such that they construct feelings.
Go to the top of the page
+Quote Post
ChronoSphere
post Apr 17 2013, 01:45
Post #6





Group: Members
Posts: 519
Joined: 11-March 07
Member No.: 41384



Isn't applying replay gain enough? I just tried doing it from foobar and it did sound the same volume for me.
I am using 5.1 headphones though, so I can't really advise anything with respect to rear/center speakers. Maybe make linux fill up the speakers similar to foobar's DSP?

Here's the result of my test:
CODE
foo_abx 1.3.4 report
foobar2000 v1.2.3
2013/04/17 02:32:55

File A: D:\Music\Ar tonelico I\Ar tonelico Hymmnos Concert Side Crimson ~Tsukikanade~\06. みとせのりこ - EXEC_SUSPEND/..flac
File B: D:\Music\Ar tonelico I\Ar Tonelico Tiny Hymmnos Concert 5.1\03. Mitose Noriko - EXEC_SUSPEND/..flac

02:32:55 : Test started.
02:33:13 : 01/01 50.0%
02:33:31 : 02/02 25.0%
02:33:48 : 03/03 12.5%
02:34:17 : 04/04 6.3%
02:34:34 : 04/05 18.8%
02:34:57 : 04/06 34.4%
02:35:13 : 04/07 50.0%
02:35:32 : 05/08 36.3%
02:35:40 : 06/09 25.4%
02:35:49 : 07/10 17.2%
02:36:02 : 08/11 11.3%
02:36:21 : 08/12 19.4%
02:36:32 : 09/13 13.3%
02:36:50 : 10/14 9.0%
02:37:09 : 11/15 5.9%
02:37:16 : Test finished.

----------
Total: 11/15 (5.9%)


The song had some parts that sounded confusing (attempts 4-7/11-12) but the rest was crystal clear. The two tracks are different in the second part, so I only tested the first one where they are in sync. As for the source, those come from the soundtrack of the game.

edit: wasn't there somewhere a thread here on HA where a test file showing that positional information can also be stored in two channels was posted? I remember something about a barber, that file gave me the creeps cause it sounded so realistic... despite being "only" 2.0

P.S.: I should stop posting at 3am after working the whole day. My sentences stop making sense.

This post has been edited by ChronoSphere: Apr 17 2013, 01:51
Go to the top of the page
+Quote Post
ktf
post May 10 2013, 19:04
Post #7





Group: Members
Posts: 398
Joined: 22-March 09
From: The Netherlands
Member No.: 68263



I just found a way around this issue, while it might not be universally applicable: instead of doing an ABX-test, I tried just playing randomly either A or B and saying whether is was surround or stereo, after a listening a few (3-5 times or something like that) times. That way small volume differences shouldn't be too much of a problem, as there was about 10 seconds between each try.

Anyway, I tried two tracks, for one I scores 17/20, the other 15/20. If my math is correct, the chances of guessing it right are 0.1% for the first one and 1.5% for the second one. I think I recognized the surround files by the instrument placement and improvement of timbre.

Still, I'm not really convinced of the benefit of 5.1 over 2.0, but I'll let it rest for now, until I buy a new set of speakers.


--------------------
Music: sounds arranged such that they construct feelings.
Go to the top of the page
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 26th October 2014 - 02:05