Article: JRiver Mac vs JRiver Windows Sound Quality Comparison

[mayhem13]

Something the subwoofer and speaker DIY and commercial developers do on other forums. They do GTGs or get togethers with different gear for listening evaluations and blind comparisons. Some are hosted by members in their homes, some are hosted by mfgrs or dealers while others are hosted by vendors. They're the highlight of the year for some folks, traveling sometimes hundreds of miles to share their creations. I think the CA fold would be a great place to do these by region, maybe hosted by local Brick and Mortar shops as a way to get living souls into their shops. A little wine or some local micro brews, great gear and a place for CA members to meet face to face as well as with dealers and maybe guest speakers/ designers local to the area.

 

New York would be a great place to kick on of these off so if there's interest, I'll start a seperate thread in the General section for feelers.

[The Computer Audiophile]

I've long wanted to have CA meet-ups much like the head-fi members have been doing for years.

[stevebythebay]

I've long wanted to have CA meet-ups much like the head-fi members have been doing for years.

 

To pull this off we'd really need the info on member's locations, to see if a quorum could be had. And then, where the real interest would be.

 

Many high end dealers do sessions with customers. However, these are almost always synced with their retail brand engagement. It's a fine venue, but limited. And those few retailers who are selling into the computer source space rarely venture beyond DACs, or complete (e.g. Sonos) setups.

 

It's a bit surprising, though the whole area of ripping to disk is still a bit of a legal issue for the industry, though it goes on nonetheless. We'd be having "clandestine" sessions, it would seem...

[Miska]

??Huh? A shift is an 8 hour work schedule. The folks on second shift better darn well build the same spec'd engine or they'll lose their jobs.

 

When you go to analog domain, like the engine is. There are manufacturing tolerances. And even the safely guarded kilogram models have been drifting over the years (some µg). So in the end it's all about measurement resolution.

 

When we deal with symbolic presentation like mathematical formulas or replicating integer numerical values we can gain perfect representations. But when you end up with irrational numbers as result of calculations in real life, you have to decide to stop writing out the decimals at some point. How many decimals of pi are enough for you is your decision, but it won't be perfect.

 

So checking if you can transfer data correctly in digital domain is of course fun and elementary exercise, but when you want to see if the analog conversion of digital samples is always the same, finding the difference is only about measurement resolution. At least the thermal noise won't have full correlation on two subsequent runs. (two independent sets of thermal noise shouldn't have normalized correlation factor of 1)

[Audio_ELF]

Sorry I misunderstood "shift" I assumed (wrongly) you were saying something about gearboxes (I believe American's refer to a manual gearbox as a stick shift?). Please ignore my comments about the engine though car analogies usually seam very limited to me and virtually pointless.

 

I'm sorry you feel I was belittling you; I do agree there is a big "ahh yes but" in Mitch's testing; but at the end of the day it appears like you (and several others) were getting away with "yes but how does it sound" type comments when "ahh but how does it measure" comments are ridiculed and put down on threads that start with purely subjective observations.

 

Eloise

[Paul R]

I thought Mitch was quite clear that he could not hear any differences when he listened in the article? That's why this is totally un-arguable, even though one may disagree. Mitch both measured to the best of not insignificant ability, and then listened.

 

That he heard no difference at all is startling to me, but not arguable. He heard what he heard.

 

I can see where it is a bit on the edge, but I don't see anyone sneering at him, calling him names, or casting aspersions on his character, which happens more often than not in the reverse scenario. Mostly I see people complimenting him on doing a good job.

 

I can see the point you are making though. Slippery slope stuff.

 

 

-Paul

 

 

Sorry I misunderstood "shift" I assumed (wrongly) you were saying something about gearboxes (I believe American's refer to a manual gearbox as a stick shift?). Please ignore my comments about the engine though car analogies usually seam very limited to me and virtually pointless.

 

I'm sorry you feel I was belittling you; I do agree there is a big "ahh yes but" in Mitch's testing; but at the end of the day it appears like you (and several others) were getting away with "yes but how does it sound" type comments when "ahh but how does it measure" comments are ridiculed and put down on threads that start with purely subjective observations.

 

Eloise

[ted_b]

Paul, he only listened (subjectively) to both recordings on one platform, through foobar ABX (which only runs on Windows). If I'm wrong here, then delete all my Mitch questions and Audio_elf arguments. Sorry.

[Paul R]

Had to go back and re-read the article, and yes, you are correct. So we can say that the recorded output from the Mac and the PC, when played back on the PC, sounded the same.

 

Not the most reasonable test for saying that the two systems, JRMC on Windows and JRMC on Mac sound the same. Definitely some compelling evidence that they put out the same digital data when recorded though.

 

-Paul

 

 

Paul, he only listened (subjectively) to both recordings on one platform, through foobar ABX (which only runs on Windows). If I'm wrong here, then delete all my Mitch questions and Audio_elf arguments. Sorry.

[mitchco]

Thanks for everyone's comments. Btw, did folks download and listen to the files on whatever system and hear any differences? I see no comments about that.

 

Just to clear up a couple of points. I did indeed listen, casually over headphones, to JRMC on the Mac and then on Windows. Heard no differences. However, it was not a properly setup ABX test as I need some hardware switching to make that work proper. So I did not include in the results.

 

Setting up proper, controlled listening tests are difficult as levels need to be matched to 0.1 dB is one of many considerations. Another point is the interesting research by James JJ Johnston that indicates our short term hearing memory is accurate for about 1/4 a second. Hence the requirement to be able to switch reasonably rapidly between sources.

 

Finally, if the bits are identical, as the results show, and that anyone can replicate (anyone up for the challenge?), at the DAC whether it came from JRMC on the Mac or Windows, then how can there be an audible difference? It can't be from the music players...

 

I am looking for a listening/measuring test, which is repeatable and can be replicated by anyone, like I did with the tests above, which demonstrates an audible difference that anyone can hear. Let me know what you think that test might look like...

[acousticsguru]

I thought Mitch was quite clear that he could not hear any differences when he listened in the article? That's why this is totally un-arguable, even though one may disagree. Mitch both measured to the best of not insignificant ability, and then listened.

 

That he heard no difference at all is startling to me, but not arguable. He heard what he heard.

 

I can see where it is a bit on the edge, but I don't see anyone sneering at him, calling him names, or casting aspersions on his character, which happens more often than not in the reverse scenario. Mostly I see people complimenting him on doing a good job.

 

I can see the point you are making though. Slippery slope stuff.

 

 

-Paul

 

Curiously enough, I've an engineer friend who can't seem to hear any difference between any two things until one produces measurements that prove there is a difference - and all of a sudden, he hears the exact same difference everyone else has been pointing out all along. I know this sounds like a story I just made up, but it's not. It's kind of hard to take an otherwise intelligent person seriously when they behave like that. But then, I know a bunch of wine lovers who are well-neigh unable to answer a seemingly simple question such as if they like the wine that's being poured before its identity and Parker rating has been revealed. Same thing, as in either case seemingly insufficient information yields to seemingly superior information. The emphasis, needless to emphasize, being on "seemingly" either way.

 

Greetings from Switzerland, David.

[Paul R]

I'm sorry, but I just cannot see the same value you do with listening to identical files on the same hardware and OS.

[stevebythebay]

Curiously enough, I've an engineer friend who can't seem to hear any difference between any two things until one produces measurements that prove there is a difference - and all of a sudden, he hears the exact same difference everyone else has been pointing out all along. I know this sounds like a story I just made up, but it's not. It's kind of hard to take an otherwise intelligent person seriously when they behave like that. But then, I know a bunch of wine lovers who are well-neigh unable to answer a seemingly simple question such as if they like the wine that's being poured before its identity and Parker rating has been revealed. Same thing, as in either case seemingly insufficient information yields to seemingly superior information. The emphasis, needless to emphasize, being on "seemingly" either way.

 

Greetings from Switzerland, David.

 

And it's even worse with musicians who seemingly listen to music in ways non-musicians don't. Sound is sometimes valued secondarily to many other attributes of expression. And I suppose it's also quite different based on the instrument being used, whether bowed string or percussive piano, for example. And better still, if a person is intimately familiar with the recording, or better yet, was part of the engineering process, the closer they can get to a valid judgement.

 

Regarding your wine metaphor: the taste/smell of wine is heavily influenced not only by expectations but by what we're consuming in the process. In my case, and likely as not many others, we're at a wine tasting in the absence of a great meal in a great setting with close friends. All of these will impact on our perception of a wine that, in absence of these elements, would be pedestrian, at best.

 

At least for me, a great music reproduction system should fool me into believing I'm in the room with real musicians, in a real venue. Even passing by such a room should make me think there are actual musicians and singers in there. The more I'm fooled the better.

[acousticsguru]

And it's even worse with musicians who seemingly listen to music in ways non-musicians don't. Sound is sometimes valued secondarily to many other attributes of expression. And I suppose it's also quite different based on the instrument being used, whether bowed string or percussive piano, for example. And better still, if a person is intimately familiar with the recording, or better yet, was part of the engineering process, the closer they can get to a valid judgement.

 

Regarding your wine metaphor: the taste/smell of wine is heavily influenced not only by expectations but by what we're consuming in the process. In my case, and likely as not many others, we're at a wine tasting in the absence of a great meal in a great setting with close friends. All of these will impact on our perception of a wine that, in absence of these elements, would be pedestrian, at best.

 

At least for me, a great music reproduction system should fool me into believing I'm in the room with real musicians, in a real venue. Even passing by such a room should make me think there are actual musicians and singers in there. The more I'm fooled the better.

 

In that case I highly recommend trying something I used to show people when I still built loudspeakers: turn off or dim all electric light and light a candle - you'll swear the bits are not the same!

 

Greetings from Switzerland, David.

[Jud]

Thanks for everyone's comments. Btw, did folks download and listen to the files on whatever system and hear any differences? I see no comments about that.

 

Just to clear up a couple of points. I did indeed listen, casually over headphones, to JRMC on the Mac and then on Windows. Heard no differences. However, it was not a properly setup ABX test as I need some hardware switching to make that work proper. So I did not include in the results.

 

Setting up proper, controlled listening tests are difficult as levels need to be matched to 0.1 dB is one of many considerations. Another point is the interesting research by James JJ Johnston that indicates our short term hearing memory is accurate for about 1/4 a second. Hence the requirement to be able to switch reasonably rapidly between sources.

 

Finally, if the bits are identical, as the results show, and that anyone can replicate (anyone up for the challenge?), at the DAC whether it came from JRMC on the Mac or Windows, then how can there be an audible difference? It can't be from the music players...

 

I am looking for a listening/measuring test, which is repeatable and can be replicated by anyone, like I did with the tests above, which demonstrates an audible difference that anyone can hear. Let me know what you think that test might look like...

 

Before thinking about tests, I did want to briefly discuss as a preliminary matter your cite to James Johnston above regarding short term hearing memory. If short term hearing memory is indeed limited to about 1/4 second, then it seems to me one can't be talking about a requirement to switch "reasonably rapidly," but rather a two-fold requirement: (1) To be able to switch nearly instantaneously; and (2) to be able to hold in our memories, with absolute reliability (an absolute reliability that doesn't exist in normal human memory), the only accurate impressions of sonic differences available to us, those from the first .25 second after switching. Frankly, I don't see how that is possible, and thus what good ABX testing would do in terms of collecting scientifically reliable data.

 

On the other hand, if asked to distinguish in a blind test between a table radio and the New York Philharmonic within an hour or even a week of having last heard them, I don't think anyone here would have any doubt. In fact I'm not sure memory would be involved - don't you think you could tell which was which the moment you first heard one or the other? So perhaps what is involved here has less to do with short term hearing memory, which if James Johnston is correct will be unreliable in any implementation of ABX testing that does not instantaneously take the data of the sense impression (something like audio ABX testing in an fMRI scanner), and more to do with how large the difference is.

 

Having cheerfully cast doubt on the idea of any comparative listening test (including ABX testing) being scientifically reliable, and being also quite skeptical of the notion that we yet know how to measure everything that affects our perception of realism in sound (see for example, http://arxiv.org/pdf/1301.0513.pdf, on what we can hear versus what measurements based on Fourier transforms can show), let me go on to suggest some tests and measurements. But why am I doing this if it's not going to give scientifically reliable results? Because I think we ought to be willing to be humble enough to let go of the idea that we are "doing science" here. I think we are following hunches that interest us and having fun along the way. Following our natural curiosity and listening to good music - what could be better, right? Methodological discipline and skepticism are powerful tools that we ought to use to full advantage, but let's not take ourselves too seriously and forget the enjoyment that is the basis of the hobby we share.

 

In suggesting tests I'm going to have to go with what I'm familiar with, and I'm not familiar at all with JRiver. Thus it will be understandable if what I suggest isn't something you care to do. If you want to go ahead, I'd love to read about what you found, positive or negative.

 

My suggested test involves XXHighEnd (XXHE), which is Windows software. (The general user consensus at this point is that it sounds best on Windows 7 rather than 8.) You can perform the test on the trial version, though the licensed version opens up capabilities that I subjectively think might allow you to hear (and measure?) differences to a greater degree. XXHE involves some system-dependent setup; if you'd like to go ahead with this, feel free to PM me for setup suggestions.

 

XXHE is a memory player. One of its settings is "SFS" (Split File Size), showing, in megabytes, the size of the "chunks" of the music file that will be pre-loaded into RAM. The test is this: Have a friend or family member (or someone who's both ;-) switch between low and high SFS settings, say 2 and 430. (Sorry, this can't be done on the fly - obvious when you think about it - so you'll have to stop XXHE and hit Play again to switch.) This won't alter the bits. The only thing that will change will be how much of the music file is pre-loaded into memory at a time.

 

See if you can hear any difference. If you can, think about what tests might be good to try to measure the types of differences you are hearing.

 

It seems to me that if something as seemingly trivial as how much of a file at a time is pre-loaded into memory can have an audible (and measurable? - my guess is that PeterSt the developer has measurement techniques for this, but how much he'd be willing to divulge of his intellectual property in this regard is another matter) effect, then it can be conceded that major changes like an entirely different OS could have an effect as well on the sound of a software player.

[stevebythebay]

In that case I highly recommend trying something I used to show people when I still built loudspeakers: turn off or dim all electric light and light a candle - you'll swear the bits are not the same!

 

Greetings from Switzerland, David.

Ah yes, when bits become waves. Though I'll be in the dark, I'll feel it course through my bones...

[mav52]

I thought Mitch was quite clear that he could not hear any differences when he listened in the article? That's why this is totally un-arguable, even though one may disagree. Mitch both measured to the best of not insignificant ability, and then listened.

 

That he heard no difference at all is startling to me, but not arguable. He heard what he heard.

 

 

 

-Paul

 

That how I read it....

[Quest]

Thanks for everyone's comments. Btw, did folks download and listen to the files on whatever system and hear any differences? I see no comments about that.

 

Just to clear up a couple of points. I did indeed listen, casually over headphones, to JRMC on the Mac and then on Windows. Heard no differences. However, it was not a properly setup ABX test as I need some hardware switching to make that work proper. So I did not include in the results.

 

Setting up proper, controlled listening tests are difficult as levels need to be matched to 0.1 dB is one of many considerations. Another point is the interesting research by James JJ Johnston that indicates our short term hearing memory is accurate for about 1/4 a second. Hence the requirement to be able to switch reasonably rapidly between sources.

 

Finally, if the bits are identical, as the results show, and that anyone can replicate (anyone up for the challenge?), at the DAC whether it came from JRMC on the Mac or Windows, then how can there be an audible difference? It can't be from the music players...

 

I am looking for a listening/measuring test, which is repeatable and can be replicated by anyone, like I did with the tests above, which demonstrates an audible difference that anyone can hear. Let me know what you think that test might look like...

 

mitch, can you respond to my simple request to do the same test with other software on the same platform? for example, itunes vs jrmc.

 

Do let me know if this isn't feasible for some reason..

[ruskiru]

Here’s a very simple test. Take a single track, play it through JRIVER and then again on the same hardware through JRIVER (via JPLAY). Both send the same bit-perfect digital stream to the DAC but using different methods. They sound different. Quite different.

 

Then do a blind listening test and ask the listeners to identify which path the track was played through, path A (standalone JRIVER) or path B (JRIVER+JPLAY). The listeners will be able to tell with some certainty which playback path was used each time. Equally you could do the same test, JRIVER vs. Foobar. Make it easy for the listeners and choose the track carefully, as some recordings will sound more markedly different than others. It also depends on the rest of the audio chain being setup correctly, especially ensuring that the speakers are properly mounted, otherwise the listeners stand less chance of being able to hear differences.

 

I don’t think sound differences in the playback chain are subjective. You either hear a difference or you don’t. Whether you like the way a particular system sounds in subjective. The test above could be repeated using the same player on different hardware, MAC vs. PC for instance.

[ruskiru]

I guess my point is that you need to use human ears to judge differences in sound quality. I can't imagine looking at an audio waveform in Audacity on a computer screen can possibly show up subtle differences in sound quality. Neither can looking at 1s and 0s, as there are many other factors that influence the final analogue signal, such as jitter in the digital domain and RFI/EMI noise etc.

[Encore]

... the only accurate impressions of sonic differences available to us, those from the first .25 second after switching. Frankly, I don't see how that is possible, and thus what good ABX testing would do in terms of collecting scientifically reliable data.

 

Very good point, Jud. The James JJ Johnston link provided by mithco doesn't link to the source of the .25 seconds, but assuming that he is quoted correctly (which I don't doubt), this basically invalidates ABX testing as it is usually performed. Test signals would have to consist of only .25 second snippets of music. Maybe it would work, but I doubt it.

[Encore]

Very good point, Jud. The James JJ Johnston link provided by mithco doesn't link to the source of the .25 seconds, but assuming that he is quoted correctly (which I don't doubt), this basically invalidates ABX testing as it is usually performed. Test signals would have to consist of only .25 second snippets of music. Maybe it would work, but I doubt it.

 

In fact, 0,125 second snippets ...

 

Moreover, I should say that I don't doubt the .25 s claim if the test signals in question are very similar. As you also point out, Jud, when test signals are sufficiently different, our auditory memory may last weeks, months, and years.

 

Another interesting case where the .25 s claim obviously isn't valid is when it comes to voice recognition. Even on the crappiest table radio you can always recognize, say Paul McCartney's voice. But I should hurry to say that I don't know if the .25 s claim was meant to pertain to any of these situations--the JJ Johnston link in mitchco's post didn't link directly to the source (and I don't trust Ethan Winer as a source).

[mayhem13]

I guess my point is that you need to use human ears to judge differences in sound quality. I can't imagine looking at an audio waveform in Audacity on a computer screen can possibly show up subtle differences in sound quality. Neither can looking at 1s and 0s, as there are many other factors that influence the final analogue signal, such as jitter in the digital domain and RFI/EMI noise etc.

 

Agreed.....as there's so many functions in a complex waveform, our brains couldn't possibly interpret their electrical waveforms in a comprehensive way no less recreate them through audible memory for comparison.

 

That being said, properly conducted AB testing of sufficient subject numbers can with certainty determine if an audible difference exists. But when faced with these results, more often than not the results are discredited by the audiophile community....some of which a actually conduct these tests. So really if boils down to faith, and a dedication to what the audiophile community stands for......powerful stuff and often beyond reason.

 

One could also consider the power of passion....for what one holds dear. We as humans make horrible mistakes of an unimaginable frequency where matters of the heart are concerned.....and will continue to do so as we are still....um....human. But it's nice to have some science around to advance the cause and keep us grounded. Modern day heretics the objectivist may be, but without passion and art, life is lifeless.

 

Enjoy the music...whether it costs $100 or $100 cubed.

[Superdad]

Agreed.....as there's so many functions in a complex waveform, our brains couldn't possibly interpret their electrical waveforms in a comprehensive way no less recreate them through audible memory for comparison.

 

 

Bulls**t! I don't need a panel of people or an ABX switch to tell me that the differences I hear are real. My system is resolving enough and my experience in the pursuit of musical reality is more than plenty to be able to make decisions about "better 1?", "better 2?" It can be s/w players, upsampling filter settings, a resistor type, a capacitor type, a particular mastering of a recording, operating system versions, or whether the draperies should be open or closed a few more inches.

 

All of these constant long threads about if people can hear this or that change have become so tiring. Frankly I think those who doubt the ability of the human ear either: a) don't have a system--or room acoustic--evolved enough to resolve what are admittedly sometimes subtle changes; b) don't have enough of a connection to the feeling that well performed/recorded music can evoke; or c) are not listening at all and are just arguing based on what they think/wish should be perfect theory.

 

How do I reach the above conclusions? Easy. I just listen. With my ears, my heart, and my experience. It is not that hard--at least for me as I have been doing this my whole life. I am not an audio engineer (though I am close with many), I not a musician (though I have lived and been lifelong friends with several), I am just deeply connected to a wide range of music and to the pursuit of a truly accurate and moving recreation of it in my home.

 

I hope others here will ignore the rhetoric and listen for themselves. I have listened to most all of the software players and there is, to my ears and my mind, no denying that there are real and significant differences between them. Someday, someone will find a way to measure those differences, but now and always, I will use the only instrument that counts, my ears.

 

Peace,

ALEX

[Encore]

That being said, properly conducted AB testing of sufficient subject numbers can with certainty determine if an audible difference exists.

 

We actually agree, you know. Just not that the tests you refer to are (only) about testing sound differences. They are just as much testing the extent to which the test persons were able to listen to each option--A or B--without having their judgment influenced by previous trials.

[Encore]

Someday, someone will find a way to measure those differences ...

 

Yes, I think you're right. In fact, for the scientific curious, I think there's some gold to be dug here. A lot of new scientific findings begin with a body of anecdotal evidence large enough to warrant further inquiry, and frankly, I don't think you could trick so many audiophiles into spending such serious money on, say, power cables if it was just of matter of placebo. I'm pretty sure most of us, especially the those of us trained in science, would rather spend it on the electronics, and have only reluctantly acknowledged that we are not just fooling ourselves.

 

Peace