Does a ~0.1db channel imbalance matter?
May 21, 2015 at 9:17 PM Thread Starter Post #1 of 24

trenzterra

Head-Fier
Joined
Jun 5, 2012
Posts
50
Likes
11
Hi all, I have a Galaxy S6 that seems to be about 0.1db louder on the right channel than on the left. To test this, I did RMAA analysis and found that the frequency response curve on the right channel is 0.1db louder. Is this the right way to measure channel imbalance? (I tried using a multimeter but was unable to get a reading no matter what... perhaps someone can shed some light how to use a multimeter to measure?)
 
I know RMAA isn't that accurate without professional, calibrated equipment. But I tested numerous other phones using it and they were generally more or less balanced (0.05db at most). So I am a little shocked that the Galaxy S6 shows so much variation.
 
So I did a couple of tests on mono mode with several earphones and noticed many of them leaned to the right on the Galaxy S6 whereas they were more-or-less balanced with my other equipment... So my question is, is this placebo, or do I have such sensitive ears to be able to tell channel imbalance with just a 0.1db difference? Or is there some other relevant aspect in the RMAA analysis that I am neglecting?
 
It's a tough shot getting the phone replaced though; I doubt the service centre will care about my request. It also does seem like a common problem. One user on XDA measured his right channel louder by 0.1db as well (whereas his Note 4 measured in at around 0.05db); a Korean review put the left channel louder by about 0.1db.
 
Any thoughts?
 
May 21, 2015 at 9:23 PM Post #2 of 24

money4me247

Headphoneus Supremus
Joined
Jan 27, 2013
Posts
6,453
Likes
3,945
try this to see if you can really hear a difference without expectation bias now that you measured one channel to be louder than the other.
 
for 0.5 dB difference: http://www.audiocheck.net/blindtests_level.php?lvl=0.5
lol edit: I scored 4/10. so 0.5 dB difference does not matter to me! :)
 
for 0.1 dB difference: http://www.audiocheck.net/blindtests_level.php?lvl=0.1
 
May 21, 2015 at 9:35 PM Post #3 of 24

trenzterra

Head-Fier
Joined
Jun 5, 2012
Posts
50
Likes
11
  try this to see if you can really hear a difference without expectation bias now that you measured one channel to be louder than the other.
 
for 0.5 dB difference: http://www.audiocheck.net/blindtests_level.php?lvl=0.5
lol edit: I scored 4/10. so 0.5 dB difference does not matter to me! :)
 
for 0.1 dB difference: http://www.audiocheck.net/blindtests_level.php?lvl=0.1


I got 1/10 even for the 0.5db test :\
 
I was just wondering if there's a test that places like the right channel 0.1db louder than the left? Or any software that I could use to adjust gain by just 0.1db?
 
May 21, 2015 at 9:48 PM Post #6 of 24

Shaffer

Headphoneus Supremus
Joined
Sep 7, 2013
Posts
2,740
Likes
272
It's probably a manufacturing variance. 0.1dB is nothing to worry about; it's likely within spec. 0.2dB, OTOH, is audible, especially if one is extremely familiar with the sound of the system.

Doug Katz tells a story of a Wadia A/D converter he was sent to audition. Almost immediately, he was blown away by its performance. He went so far as to call many of his colleagues and tell them about this miraculous piece of gear. Long story made short, he eventually measured the converter's output and found it 0.2dB louder than his reference. Level-matched to 0.1dB, he wasn't nearly as impressed with the sound.
 
May 21, 2015 at 9:49 PM Post #7 of 24

MindsMirror

1000+ Head-Fier
Joined
May 9, 2014
Posts
1,012
Likes
254
0.1dB is only 1% difference in voltage. You might expect a typical SMD component to have somewhere around a 1% tolerance, so it is easy to see that you could end up with a 1% difference in the channels.
 
To measure with a multimeter you must use AC voltage mode. Plug in a 3.5mm aux cable and play a sine wave. Measure the voltage across the sleeve and ring for the right channel, sleeve and tip for the left channel.
 
May 21, 2015 at 10:08 PM Post #8 of 24

trenzterra

Head-Fier
Joined
Jun 5, 2012
Posts
50
Likes
11
  0.1dB is only 1% difference in voltage. You might expect a typical SMD component to have somewhere around a 1% tolerance, so it is easy to see that you could end up with a 1% difference in the channels.
 
To measure with a multimeter you must use AC voltage mode. Plug in a 3.5mm aux cable and play a sine wave. Measure the voltage across the sleeve and ring for the right channel, sleeve and tip for the left channel.

 
Thank you. So is this 1% difference material to our ears?
 
Anyway, for RMAA which figure should I look at for channel imbalance? I see that while frequency response shows a 0.1db variation, the variation for intermodulation distortion + noise (swept frequency) is about 1.5 db. Is there any cause of concern?
 

 
May 21, 2015 at 10:44 PM Post #9 of 24

money4me247

Headphoneus Supremus
Joined
Jan 27, 2013
Posts
6,453
Likes
3,945
@Shaffer is correct. 0.1 dB volume difference between two components is considered acceptable when testing gear in a double blind testing set-up. Therefore, a 0.1 dB channel imbalance should also not be audible. I wouldn't worry about it at all.
 
May 21, 2015 at 10:51 PM Post #10 of 24

jcx

Headphoneus Supremus
Joined
Jul 24, 2002
Posts
2,371
Likes
368
your soundcard likely uses 1% resistors, and the ADC, DAC chips have R/L channel tolerance of that order too
 
so you really need to dig into measurement technique, calibration, or clever ratiometric techniques to get it all lined up to definitively talk 0.1 dB, ~ 1% accurate measurement 
 
May 21, 2015 at 11:04 PM Post #11 of 24

MindsMirror

1000+ Head-Fier
Joined
May 9, 2014
Posts
1,012
Likes
254
No, 0.1dB or 1% is imperceptible.
 
The frequency response graph will show the channel imbalance.
 
1.5dB difference in levels is noticeable. However, that difference is only in the IMD+N, not in the signal. The IMD+N should be below the audible threshold anyway, so while the left and right channels may produce different levels of IMD+N, if neither of them are audible it doesn't matter.
 
May 22, 2015 at 12:18 AM Post #12 of 24

trenzterra

Head-Fier
Joined
Jun 5, 2012
Posts
50
Likes
11
Thanks all for the replies and assurance!

your soundcard likely uses 1% resistors, and the ADC, DAC chips have R/L channel tolerance of that order too

so you really need to dig into measurement technique, calibration, or clever ratiometric techniques to get it all lined up to definitively talk 0.1 dB, ~ 1% accurate measurement 
I am using a sound blaster z. Any idea how the resistors perform? Sorry I'm not that well versed with the technical details. All I know is that for most of my other audio equipment they are evenly matched or 0.05db mismatched, usually with the right channel louder. So it could be that mY soundcard has a 0.05db error and my phone is contributing the rest?
 
May 22, 2015 at 3:13 AM Post #13 of 24

cdsa35000

500+ Head-Fier
Joined
Oct 14, 2014
Posts
530
Likes
110
They mostly use SMD 5% resistors, maybe your ears is imbalance?
Did you try swapping/reverse L/R headphone or loudspeakers test?

You can try burn-in all gear at loud volume 24+ hrs continuousloop with: DeMagic, so all electronics will be stressed and then settle down/stablized their % variance fluctuations.

[VIDEO]https://www.youtube.com/watch?v=8BTaK3TsuoA[/VIDEO]

You can use http://keepvid.com etc. to convert youtube to mp3/mp4 files etc.
 
May 22, 2015 at 6:48 AM Post #14 of 24

castleofargh

Sound Science Forum Moderator
Joined
Jul 2, 2011
Posts
9,361
Likes
4,746
  Hi all, I have a Galaxy S6 that seems to be about 0.1db louder on the right channel than on the left. To test this, I did RMAA analysis and found that the frequency response curve on the right channel is 0.1db louder. Is this the right way to measure channel imbalance? (I tried using a multimeter but was unable to get a reading no matter what... perhaps someone can shed some light how to use a multimeter to measure?)
 
I know RMAA isn't that accurate without professional, calibrated equipment. But I tested numerous other phones using it and they were generally more or less balanced (0.05db at most). So I am a little shocked that the Galaxy S6 shows so much variation.
 
So I did a couple of tests on mono mode with several earphones and noticed many of them leaned to the right on the Galaxy S6 whereas they were more-or-less balanced with my other equipment... So my question is, is this placebo, or do I have such sensitive ears to be able to tell channel imbalance with just a 0.1db difference? Or is there some other relevant aspect in the RMAA analysis that I am neglecting?
 
It's a tough shot getting the phone replaced though; I doubt the service centre will care about my request. It also does seem like a common problem. One user on XDA measured his right channel louder by 0.1db as well (whereas his Note 4 measured in at around 0.05db); a Korean review put the left channel louder by about 0.1db.
 
Any thoughts?

 
1/ rmaa is what it is.
2/ how do you know that the imbalance comes from the S6 and not from the input you used to measure it?
3/ 0.1db is what any audible test ever done asks before we can say that 2 gears are volume matched and that people won't be able to tell a difference just from volume level. so maybe start trusting people when they tell you it doesn't matter.
4/ I'm ready to bet that most of your headphones have more than 0.1db imbalance at many points in the frequency response. and then just the way you place a headphone/IEM can make changes of several db. if you hear a mono sound too much on one side, maybe start by looking into that instead.
5/ do you think your left and right ear are perfectly balanced? that can be checked by an audiologist, and you can ask him about what he thinks of 0.1db.
 
 
summary, forget about it! ^_^
 
May 22, 2015 at 7:13 AM Post #15 of 24

money4me247

Headphoneus Supremus
Joined
Jan 27, 2013
Posts
6,453
Likes
3,945
   
1/ rmaa is what it is.
2/ how do you know that the imbalance comes from the S6 and not from the input you used to measure it?
3/ 0.1db is what any audible test ever done asks before we can say that 2 gears are volume matched and that people won't be able to tell a difference just from volume level. so maybe start trusting people when they tell you it doesn't matter.
4/ I'm ready to bet that most of your headphones have more than 0.1db imbalance at many points in the frequency response. and then just the way you place a headphone/IEM can make changes of several db. if you hear a mono sound too much on one side, maybe start by looking into that instead.
5/ do you think your left and right ear are perfectly balanced? that can be checked by an audiologist, and you can ask him about what he thinks of 0.1db.
 
 
summary, forget about it! ^_^

1) Maybe rightmark audio analyzer
3) yup, 0.1dB measured difference is considered to be inaudible to the human ear
 
summary, what you think you are hearing is unlikely due to the Samsung Galaxy S6 measuring a 0.1dB volume difference. can be your ears, your source, or something else completely. quick first check I would do is ensure that your computer levels are balanced. Right click volume icon > properties > right click the playback device > properties > levels tab > balance button >>> check to see if the levels are matched. 
 
then you should check with mono test tones first to see if the channel imbalance is repeatedly identifiable. then eliminate aspects of your chains to isolate what component is causing the issue.
 

Users who are viewing this thread

Top