[ClieOS]

I'll urge you all to listen to the files first before pulling out the software to analyze them, as the point is really on whether you think you can hear the difference. In any case, I have generated differentiated clips (*one clip minuses another) and will post them up. All in good time.

 

[KSB1978]

There are numerous unconfirmed anecdotes that software/firmware changes sound quality.
I experienced a sound change when I was playing with Jriver player on my HTPC and their updates.
I certainly thought FW2.0 and 2.02 sounded different on the Fiio X5.
 
I don't know what the differences could be, but all conversations with developers and stuff indicate these effects are uncontrolled/unintended side effects. The changes I have perceived in any of the above certainly aren't huge. It's subtle colouration type stuff. Probably smaller change than cables (am I opening another can of worms?).
And as someone has pointed out (or raised the question?), some people are more sensitive to changes in SQ than others... or it could be that some people listen out for stuff like this deeper than others... whatever, some people tend to perceive it more than others.

 

[KSB1978]

I've finally got to update the FW to v2.2 from v2.02. The first thing I noticed was it's louder. I always listen around volume setting on 66.
I am now listening on 58-60. Is this a normal thing? or is it an inadvertent side-effect? I am not sure. But sound level is certainly one of the most obvious and widely perceptible sound quality change.

 

[ClieOS]

  I've finally got to update the FW to v2.2 from v2.02. The first thing I noticed was it's louder. I always listen around volume setting on 66.
I am now listening on 58-60. Is this a normal thing? or is it an inadvertent side-effect? I am not sure. But sound level is certainly one of the most obvious and widely perceptible sound quality change.

 
...or it can be all in your head. Actually it isn't that hard to check for loudness difference as long as you have a AC meter.

 

[KSB1978]

Definitely not "in my head". I work in acoustics and my listening level is very particular.

Are you disregarding because of your personal experience?

 

[ClieOS]

Definitely not "in my head". I work in acoustics and my listening level is very particular.

Are you disregarding because of your personal experience?

 
No, just providing alternative theory and the general understanding that human hearing is usually not that reliable.

 

[cjl]

  Would audacity analyze the difference in the waveforms for you and output the difference?  Anyway, quick glance, they are indeed identical.

The easiest way to do that is to invert one of the files and sum them. All the identical content will cancel, leaving you with the difference. In this case though, the files aren't perfectly time-aligned, so before you could invert one and so the sum, you would need to align them as perfectly as possible (and a slight misalignment could cause significant "differences" to appear that don't exist in reality). The misalignment doesn't seem to be an integer number of samples either, which makes it rather difficult with these particular files (I just spent a bit trying to get them to null out, but I can't get rid of the last bit of misalignment - I can get a -30dB null or so, depending on which two files I pick, but I can't get much better than that with just my quick bit of effort...). It's also possible that there are enough differences that it is impossible to get a true null, though I'm not willing to state that conclusively just yet without playing around quite a bit more...

 

[ClieOS]

... and what is the initial subjective opinion, if you may?

 

[cjl]

I can't hear a difference, for what it's worth. I haven't spent a lot of time trying to listen for any sort of fine details or differences though.

 

[Roly1650]

The easiest way to do that is to invert one of the files and sum them. All the identical content will cancel, leaving you with the difference. In this case though, the files aren't perfectly time-aligned, so before you could invert one and so the sum, you would need to align them as perfectly as possible (and a slight misalignment could cause significant "differences" to appear that don't exist in reality). The misalignment doesn't seem to be an integer number of samples either, which makes it rather difficult with these particular files (I just spent a bit trying to get them to null out, but I can't get rid of the last bit of misalignment - I can get a -30dB null or so, depending on which two files I pick, but I can't get much better than that with just my quick bit of effort...). It's also possible that there are enough differences that it is impossible to get a true null, though I'm not willing to state that conclusively just yet without playing around quite a bit more...

Audio Diff Maker will take the time, sample rate drift and level alignment struggle away for you, within reason, it obviously can't correct gross mismatches. It also does the phase inversion and subtraction. You get the difference as a .wav file you can load into Audacity or whatever so you can use the tools to analyse the files and play the audio. Audacity is actually better at looking at the difference file, the Audio Diff Maker results can be misleading viewed as raw numbers, the spectrum analysis in Audacity gives a much better idea about what may be audible. Do a same file comparison first to get an idea of system baseline.
Audio Diff Maker is free from Liberty.

 

[castleofargh]

I didn't have the strength to do them all in all combinations. but the only one that really stands out to me seemed to be the F one. I had 100% on abx with E-F and with B-F(about 10 tries each) and got around 50/50 on 2 or 3 other tries that didn't involve F.
should I say what main difference I noticed or is that for everyone to find out alone?

 

[cjl]

F? The file I downloaded only contains A-E...

 

[SilverEars]

  I didn't have the strength to do them all in all combinations. but the only one that really stands out to me seemed to be the F one. I had 100% on abx with E-F and with B-F(about 10 tries each) and got around 50/50 on 2 or 3 other tries that didn't involve F.
should I say what main difference I noticed or is that for everyone to find out alone?

Now I will be looking for it since you just provided an expectation bias.    I personally don't think a difference can be heard if the only difference in the waveform is only shift in time, unless there is phase in-alignment that is audible.  I guess the introduction of Audacity revealing the waveform is putting a damper on subjective differences to be heard.  Looking at the waveform creates a bias also.

 

[cjl]

  Now I will be looking for it since you just provided an expectation bias.    I personally don't think a difference can be heard if the only difference in the waveform is only shift in time, unless there is phase in-alignment that is audible.  I guess the introduction of Audacity revealing the waveform is putting a damper on subjective differences to be heard.  Looking at the waveform creates a bias also.

For what it's worth, just looking at the waveform in Audacity isn't conclusive. It's possible to have something 25-30dB down that is still audible, which really wouldn't show up much on a visible waveform. I won't rule that possibility out (I haven't had a chance to play around with DiffMaker yet). All the initial inspection with Audacity revealed was that there were no major differences between the files.
 
(That having been said, I still am confused about the reference to file "F", since as I said, the directory I downloaded only contains A-E)

 

[SilverEars]

 
(That having been said, I still am confused about the reference to file "F", since as I said, the directory I downloaded only contains A-E)

Look in the first post, he added F later on.