[Baxide]

Here you go: human auditory memory

Cheers

Thanks for the link. It only goes to further debunk your use of it as a quote in this subject. You wrote audio memory. The article refers to Auditory Sensory Memory. I hope you can distinguish between the two.
Even if you don't, the article clearly states that it lasts up to 10 seconds and the tests were done with audio tones. Testing a piece of music surely does not rely on listening to a tone of no longer than 10 seconds. You are listening to variations in frequencies and over a far longer period. As matter of fact, just preparing to listen from one music piece to another will take several seconds to get adjusted, and concentrated listening for comparison purposes will be well outside the 10 second span. So the auditory memory effect is of zero significance in this case.

 

[Makiah S]

  Ahh, but how is saying "accept objectively that FLAC is better than (well-encoded 320kbps) mp3" any different from saying "accept objectively that 24/96 is better than 16/44"? In both cases, there is technically more information in the latter, but that isn't the point. The real question is whether there's an audible difference.

 
24/96 and 16/44 are both formats of lossless, mp3 and Flac are lossless and lossy, lossy has always from the beginning comprimised quality for convience, and when you have a system sufficent enough, the differance is pretty obvious, and most noteably it's in the highs. Heck I discover lossy transcodes in my own collection at least twice a month, as I'm listening to a song I notice the treble... lacks extension and sure enough it's a 320k mp3 that's been transcoded to a Flac
 
the question of can you hear the differance between lossless and a higher bit rate lossless is a more candid subject, but trying to prove that lossy is audibly the same as Lossless, is pointless. It's been proven and established for years and last I checked there hasn't been a revision to the LAME encoder in years
 
now if you want to debate the audible differances between q10 Vorbis and Flac, that is another valid conversation, as the vorbis encoder focus more on quality than it does "shirnkage' in the case of vorbis [lossy] vs Flac [lossless] there's is a chance that the lossy file may be on par with the lossless one, but no one uses vorbis sadly, and mp3 and aac are still the biggest lossy formats

 

[pfurey89]

I'd bet very few ears can tell the differences, truly. 

 

[ab initio]

Thanks for the link. It only goes to further debunk your use of it as a quote in this subject. You wrote audio memory. The article refers to Auditory Sensory Memory. I hope you can distinguish between the two.
Even if you don't, the article clearly states that it lasts up to 10 seconds and the tests were done with audio tones. Testing a piece of music surely does not rely on listening to a tone of no longer than 10 seconds. You are listening to variations in frequencies and over a far longer period. As matter of fact, just preparing to listen from one music piece to another will take several seconds to get adjusted, and concentrated listening for comparison purposes will be well outside the 10 second span. So the auditory memory effect is of zero significance in this case.

You're ability to make stuff up as you go is mildly amusing. Please refer to the links in my signature regarding the proper way to argue one's position. Name calling and unsupported assertions that the other person is incorrect are at the bottom of the pyramid.

My previous post that determining whether there is audible difference between two sources is best accomplished via careful DBT and quick switching is the current and best supported stance concerning scientific evidence. I proceeded to supply you with a peer reviewed source that was on the topic which i mentioned in my post.

If you take a contrary stance to my statement regarding ABX testing, provide scientific evidence to back your claim, or admit that your opinion is unsupported and contrary to the known body of evidence.

You prentending that you dont understand my argument or asserting that my argument is something other than what I've said is counterproductive to furthering the discussion.

Cheers

 

[cjl]

I agree that rolled off highs are typical of low bitrate MP3s, but good quality encoders do not use a low pass filter on high-bitrate encodes (LAME doesn't use one at 256 or above). I would be extremely surprised if you could hear the difference between a LAME-encoded 320kbps MP3 and a FLAC. As for 16/44 and 24/96 both being lossless? Sure, but only in the sense that data is never lost after the original sampling has been done. 24/96 contains far more data than 16/44, so I think the analogy holds up just fine.

 

[cjl]

In fact, here are the two samples from the original set from the start of this thread. One of the two copies of each has been converted to mp3 with LAME (320 CBR), then converted back to wav. I will be extremely impressed if you can tell me which is which (or even just show a successful abx in foobar between two of the files):
 
https://dl.dropboxusercontent.com/u/40020825/music%20test%20files/MP3_vs_wav.zip

 

[Baxide]

You're ability to make stuff up as you go is mildly amusing. Please refer to the links in my signature regarding the proper way to argue one's position. Name calling and unsupported assertions that the other person is incorrect are at the bottom of the pyramid.

My previous post that determining whether there is audible difference between two sources is best accomplished via careful DBT and quick switching is the current and best supported stance concerning scientific evidence. I proceeded to supply you with a peer reviewed source that was on the topic which i mentioned in my post.

If you take a contrary stance to my statement regarding ABX testing, provide scientific evidence to back your claim, or admit that your opinion is unsupported and contrary to the known body of evidence.

You prentending that you dont understand my argument or asserting that my argument is something other than what I've said is counterproductive to furthering the discussion.

Cheers

I don't buy your suggestion for one moment. Have you ever tried to tune a musical instrument? The tuning fork output level is nowhere as close to that of the instrument that is being tuned. However, it is still possible to tune the instrument with certain accuracy.
Having two audio passages under test at the same level raises into question as to what exactly that means. A 16 Bit signal will have a different level than a 24 bit signal. So where exactly would the two of them be judged to be at the same level? Even the VU and Peak meters in our studio couldn't figure that one out, let alone for anyone to make that call purely by ear. You could set the two audio tracks at a point where both just about lighten up the clipping indicators, but a quick listen afterwards is more than likely to show that the overall playback of one audio passage is at a different amplitude than the other. Even overlaying them on a scope screen shows up differences in maximum amplitude between various bit rates.

 

[castleofargh]

  I don't buy your suggestion for one moment. Have you ever tried to tune a musical instrument? The tuning fork output level is nowhere as close to that of the instrument that is being tuned. However, it is still possible to tune the instrument with certain accuracy.
Having two audio passages under test at the same level raises into question as to what exactly that means. A 16 Bit signal will have a different level than a 24 bit signal. So where exactly would the two of them be judged to be at the same level? Even the VU and Peak meters in our studio couldn't figure that one out, let alone for anyone to make that call purely by ear. You could set the two audio tracks at a point where both just about lighten up the clipping indicators, but a quick listen afterwards is more than likely to show that the overall playback of one audio passage is at a different amplitude than the other. Even overlaying them on a scope screen shows up differences in maximum amplitude between various bit rates.

this post is really confusing.
what I understand is that you're saying a 24bit track turned into 16bit will not have the same levels(by level I guess you're talking db levels?). that I clearly don't get. a sound that is recorded as -5db below the max output(0db) will still be at -5db when turned into 16bit.
I guess I'm not understanding what you're trying to say at all.(but then again, as the king of "lost in translation", the problem might just be on my side).

 

[Makiah S]

These threads are so pointless, let's assume I correctly ABX the files, then we spend another 12 pages disproving what I did or I get to abx another 14 set of files till some one has sufficent proof to go back to their orignal belief 
 
and lossy encoders REMOVE data, 24 bit is simply a great sample size, both 16bit and 24bit encodes are lossless in that nothing is removed, lossy REMOVES data, again you cannot compare lossy to lossless. With 24 and 16 bit, both are samples taken from a complete, where as lossy is a sample of a sample. it removes byes from an already incomplete capture 
 
Have fun abx testing for teh next few weeks you guys lol

 

[mikeaj]

  this post is really confusing.
what I understand is that you're saying a 24bit track turned into 16bit will not have the same levels(by level I guess you're talking db levels?). that I clearly don't get. a sound that is recorded as -5db below the max output(0db) will still be at -5db when turned into 16bit.
I guess I'm not understanding what you're trying to say at all.(but then again, as the king of "lost in translation", the problem might just be on my side).

 
It's not just you. I spent a few minutes rereading that and also came away confused. Is there any playback gear that outputs different levels with 16 bit samples as opposed to 24 bit samples (not counting undithered and levels close to the least significant bit)?
 
 
Also, what does tuning instruments have to do with anything? I've tuned instruments, tuned orchestras (the meaning and process of tuning here being different, of course), tuned to instruments, etc. and am stumped. I guess the point was about output levels being different? But the process and aim of listening is different there, and even disregarding that I'm stumped.

 

[sonitus mirus]

  These threads are so pointless, let's assume I correctly ABX the files, then we spend another 12 pages disproving what I did or I get to abx another 14 set of files till some one has sufficent proof to go back to their orignal belief 
 

 
Who cares what any of us think?  Aren't you even a bit curious to know the results for yourself?  The ABX test isn't going answer anything for anyone except you. 

 

[castleofargh]

  These threads are so pointless, let's assume I correctly ABX the files, then we spend another 12 pages disproving what I did or I get to abx another 14 set of files till some one has sufficent proof to go back to their orignal belief 
 
and lossy encoders REMOVE data, 24 bit is simply a great sample size, both 16bit and 24bit encodes are lossless in that nothing is removed, lossy REMOVES data, again you cannot compare lossy to lossless. With 24 and 16 bit, both are samples taken from a complete, where as lossy is a sample of a sample. it removes byes from an already incomplete capture 
 
Have fun abx testing for teh next few weeks you guys lol

 I'm the first to come an say that 16/44.1 PCM is all we will ever need as music listeners. but it doesn't change the fact that strictly speaking 24/96 has more information. most of those will be redundant information, or even useless information when it comes to recreate the audible analog signal, but still saying that 24/96 is superior to 16/44 is every bit as true as saying that 16/44 is superior to mp3. that is if we focus purely on quantity of information contained(or better, the quantity each could contain).
 
the purpose of abx is not to prove or disprove this. it is to take another approach and see if for us as an individual(the good way to approach subjectivism), mp3 or 16/44 or 24/96 would really matter in practical everyday uses on our own gears. 
we can all get out of an abx with a different opinions and the results don't have to make us change religion. there are no wrong answer and we can all listen to whatever file we like, in the end it doesn't matter. the purpose of an abx is to get out of all the "ready made opinions" we have all seen and heard numerous times, and go make an opinion for ourself instead of repeating the last most convincing argument we came upon. I fail to see how an abx can be anything but a good thing for myself.
 
#mylife
I tried some macgyver style abx at the beginning of mp3 and decided at the time to go with ogg because it was audibly better(really clearly audibly better). a few years after I tried mp3 again and found out it had improve a great deal(post 2003). so I went and used mp3 cbr320 in all my daps(ok not on the minidiscs ^_^). and not long ago somebody talked about vbr being as good as cbr, so I went and did some abx again because I was not gonna trust some guy on a forum. I couldn't conclusively tell anything, except on 2 classical files I know perfectly and where I really heard there was a difference but still I couldn't conclusively pick between cbr and vbr(and in fact I kind of felt like I prefered the vbr... pretty surprising for me). so now I'm using vbr max on all my daps. every times abx helped me decide by myself, which is a great deal to me, as myself is pretty much the only guy that matters in my life ^_^.
 
that has nothing to do with my beliefs about lossless vs lossy, I archive all my music in 16bit flac on my computer and would never suggest to only keep mp3s.

 

[Makiah S]

   
Who cares what any of us think?  Aren't you even a bit curious to know the results for yourself?  The ABX test isn't going answer anything for anyone except you. 

I've done it so many times, already. With a friend of mine, on another forum like a year ago. I'm so tired of having this discussion, abx testing than having people spend hours trying to tell me it was a fluke, or that I passed on sheer luck... 

 
 I'm the first to come an say that 16/44.1 PCM is all we will ever need as music listeners. but it doesn't change the fact that strictly speaking 24/96 has more information. most of those will be redundant information, or even useless information when it comes to recreate the audible analog signal, but still saying that 24/96 is superior to 16/44 is every bit as true as saying that 16/44 is superior to mp3. that is if we focus purely on quantity of information contained(or better, the quantity each could contain).
 
the purpose of abx is not to prove or disprove this. it is to take another approach and see if for us as an individual(the good way to approach subjectivism), mp3 or 16/44 or 24/96 would really matter in practical everyday uses on our own gears. 
we can all get out of an abx with a different opinions and the results don't have to make us change religion. there are no wrong answer and we can all listen to whatever file we like, in the end it doesn't matter. the purpose of an abx is to get out of all the "ready made opinions" we have all seen and heard numerous times, and go make an opinion for ourself instead of repeating the last most convincing argument we came upon. I fail to see how an abx can be anything but a good thing for myself.
 
#mylife
I tried some macgyver style abx at the beginning of mp3 and decided at the time to go with ogg because it was audibly better(really clearly audibly better). a few years after I tried mp3 again and found out it had improve a great deal(post 2003). so I went and used mp3 cbr320 in all my daps(ok not on the minidiscs ^_^). and not long ago somebody talked about vbr being as good as cbr, so I went and did some abx again because I was not gonna trust some guy on a forum. I couldn't conclusively tell anything, except on 2 classical files I know perfectly and where I really heard there was a difference but still I couldn't conclusively pick between cbr and vbr(and in fact I kind of felt like I prefered the vbr... pretty surprising for me). so now I'm using vbr max on all my daps. every times abx helped me decide by myself, which is a great deal to me, as myself is pretty much the only guy that matters in my life ^_^.
 
that has nothing to do with my beliefs about lossless vs lossy, I archive all my music in 16bit flac on my computer and would never suggest to only keep mp3s.

and yea, I have everything as 16bit flacs, I also listen to my EDM in 320k mp3s, Since I EQ edm who cares about the extra resulution I'm going to throw so much DSP at it I might as well save some space, in addition  vbr is quality wise as good as mp3, but I've had players in the past struggle with vbr, hence forth I stick to 320k cbr 

 

[cjl]

Well, if you don't want to prove that you can hear a difference, that's fine with me. The main point of posting those files was to show that, contrary to your claim, high bitrate MP3 does not roll off or significantly affect the high frequencies in any way, aside from a pretty sharp cutoff at 20kHz (which isn't audible). If anyone is curious, here's the spectrum analysis showing the difference between the MP3 (bottom) and wav (top).
 

 

[Makiah S]

There is also signifcantly less "white noise" or that blueish saturation, that coloring is the decay of your higher frequcancies noise, some of that decay or exiting notes are removed, and it's visable in the mp3,