solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Mar 12, 2014 20:08:36 GMT
|
|
|
Post by hifidez on Mar 13, 2014 8:49:59 GMT
The test is uploaded There are 112kbs, 128, 160, 192 and 320 kbs files including the original file. Also there are test folders with each 5 files. In every test folder the MP3 rate and original files are present and you have to determine (by ear, no cheating) which of the 5 files is either MP3 (in the rate as per the folders name) or original. All files are WAV files so you can't see directly if a file has been compressed and later expanded again or original. There is a results folder that tells you which files are the original and MP3 ones so you can check if you were right. So the test is private really as nobody needs to know what their threshold is unless they post it themselves. Thank you. Will download later. Looking forward to the challenge :-) Derek
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Mar 13, 2014 8:59:53 GMT
the 'fun' think about this test is that care has been taken while encoding to MP3 and a GOOD resampler has been used that doesn't add things (most do, some even a LOT). Also the playback conditions are the same so you only listen to the differences brought on by decoding.
IF the playback software is not decoding MP3 correctly or, for those that feel processor load is degrading the sound, than this test could be better than just comparing an MP3 to a WAV as the playing field has been 'levelled' playback-wise.
Could construct a similar test using electronic music or very complex music.
Those that have suggestions or can upload a specific song I will gladly make a 'test'.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Mar 13, 2014 16:55:43 GMT
Have had some feedback from the Vikings and may have a listening hint:
Tonal balance etc. and detail retrieval should be exactly the same for all the files. Possibly the attack of cymbals can differ for MP3 files. The music should sound excatly the same when switching back and forth.
The differences should be/is said to be in 'air around soundsources - realism - smoothness - placement of instrments' - 3D soundstage - etc. (acc. to those who hear) Personally I can't hear a difference between the files because I don't listen out for the right clues, yet can assure you they are there and quite measurable. I usually listen out for technical deficiencies and these files have none.
listen relaxed to it and try to asses which one sounds 'best' or 'more real' instead of listening out for differences in sound/distortion or whatever. Gut feeling thing rather than technical listening. Maybe this could help during the 'blind test' phase, maybe it won't.
For those who got the answers correctly (and didn't cheat by looking/poking/analysing) put the known files in foobar ABX tool and gather some 'evidence'.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Mar 16, 2014 11:33:53 GMT
The test results explanation and info about the used files can be found HEREwww.mediafire.com/view/7rp97tfd497mf0p/file-test_info.pdfShould anyone have missed the opportunity or want to try it but with (192/24) music they know very well and may be more revealing than this one please contact me or post here.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Apr 19, 2014 15:27:56 GMT
On request I uploaded the MP3 test again and will be available for download till April 26th. www.filemail.com/d/erjomozxcjzgxceFor those that missed it and want to find out the audibility threshold of lossy MP3 compression. There are 112kbs, 128, 160, 192 and 320kbs files including the WAV original file. Also there are test folders with each 5 files. In every test folder the MP3 rate and original files are present and you have to determine (by ear, no cheating) which of the 5 files is either MP3 (in the rate as per the folders name) or original. All files are WAV files so you can't see directly if a file has been compressed and later expanded again or original. There is a results folder that tells you which files are the original and MP3 ones so you can check if you were right. So the test is private really as nobody needs to know what their threshold is unless they post it themselves. For those wanting to give the differences between 16 and 24 bit a try (but using high quality classical music) give this test a try: archimago.blogspot.nl/2014/04/internet-test-24-bit-vs-16-bit-audio.html
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Jul 9, 2014 19:40:58 GMT
|
|
Rabbit
Administrator
Posts: 7,087
|
Post by Rabbit on Jul 10, 2014 11:50:48 GMT
Those types of discussion always seem to rile audio people, Frans. I don't personally have a problem with mp3 but many hifi types basically take the mick if you dare mention the format. I don't really like the 'snobbery' element since to me, the music is way more important than worrying about the format. I find it difficult to hear massive differences between CDs and mp3. I only rip at 320 just for my own peace of mind. I certainly don't find mp3 offensive as so many audiophiles claim. I don't like the derisory comments made about it really either. I liken it to me finding out what music people listen to and then launch an attack on them over their lack of taste or whatever. More important is the music, not just the gear but so many if us get tied up completely with the equipment to the point that it becomes an obsession. I guess I've been like that for years although I've calmed down a lot recently with regard to headphones in particular. I've never forgotten the look on a dealer's face when I tried to get hold of one of the first MP3 players. He was absolutely horrified that I would stoop to such lows, given my job. The thing is, he'd never heard one, ever. However, there he was, running it down. As far as hifi snobbery goes, you're either an armchair critic (whatever that means) or a deaf electronic engineer!! That just about covers both ends of the spectrum - armchair suggests you just casually listen and don't do any diy, and deaf EE means you can do all the diy you like but you can't hear anything. That leaves the middle ground of critics who know better than the other two categories. The ...... Whingers? Or perhaps the 'Pick 'n Mix' mob who choose what they want to believe?
|
|
solderdude
Administrator
measureutternutter
Posts: 4,881
|
Post by solderdude on Jul 10, 2014 15:19:35 GMT
The part about MP3 being able to encode 24 bit and leave the resolution untouched was something I didn't know. I always thought it wasn't able to go below 16 bit but it obviously can describe 24 bit (hires ?) files just as well. Of course, as he explained, it's the player software/hardware and DAC as well as operating systems that f' it up.
I am not able to tell MP3 (320 in highest quality) apart from a WAV myself so I don't really care about hi-res audio either. That is unless the master is audibly better in which case I relentlessly encode it to 320k MP3 to save space on my 32GB uSD card and can fully enjoy all the sound quality that better master has to offer. Of course my DAC, amps, and even DAP can do 96/24 (DAP even 192/24) and have headphones that measure to 30kHz (the boundary of my mic) and are likely to go even further but even with those the 'magic' some report I can't detect when listened BLIND. When I know hi-res is playing I am convinced it sounds better but when listening out for the SAME things blind statistics tell me I am guessing 100%
I sat down and did some (semi) blind tests with an album which I saw had ultrasonic content and one converted to MP3 myself (nothing downloaded, God knows who may have f'ed it up) and cannot tell the difference though when you get it right are always thinking... see I DID hear it but statistics say otherwise. Funny how the mind works...
Before someone asks how you can do a semi blind test without someone else being present without using something like ABX tools ... Make 2 different files and copy both a few times so you end up with a folder of say 20 songs of which half is file-A and half is file-B. Now random play that folder... have a listen and see if you got it right. (make notes) hit the next button and the random choice will either let you listen to A or B.
Of course IF you are using a NOS DAC or another DAC or player that alters the sound depending on the sample frequency you can still do the test BUT you must upsample the MP3 again to the higher bitrate.
I am 100% sure I can tell MP3 320 highest Q apart from 192/16 very reliably IF I am using an audiophile DAC with 'gentle' or non pre-ringing filtering but in this case you aren't telling the files apart but the sample rates. This may very well be why many 'audiophiles' can tell red-book/MP3 apart from hires files using their uber expensive DAC's with soft knee filtering.
|
|
Rabbit
Administrator
Posts: 7,087
|
Post by Rabbit on Jul 10, 2014 16:40:15 GMT
It's really quite complex and I'm not certain whether people can tell the difference really, or whether they spot some kind of 'clue' that helps. For instance, I had a couple of tracks to compare a year or so ago and I got the hi res. however, for me to spot it was easy and not really based on what I'd say hi res actually was. For some reason, the hi res file had a stronger bass. That's not hi res exactly, but a tone difference. Whether that's one of the effects of hi res I'm not certain. More like a coloration and who's to say which was more correct?
That's one thing I have noticed with better dac's or high res files - they seem to have a stronger bass, but then, I may be imagining it. There doesn't seem to be a massive difference in the quantity of information you hear. I tend to detect tonal differences much more readily, but because a sound is 'fuller' doesn't mean it's necessarily hi res.
Many people 'think' that the fullness is a result of hi res, but if you get a lower res file (320) and tweak the bottom end slightly, you end up with the same sound tbh which isn't hi res at all.
True hi res should have stronger special clues and transients I guess as major features. I don't really hear any massive differences easily myself, other than tonal clues.
Of course, pride comes into it and there seems to be almost a kind of 'fear' in not being able to identify hi res files. That's why not many are prepared to try the blind types of test, since it is so hard to do and there's a certain amount of qudos that comes from being able to spot the correct files. I still wonder what the reality is because I have a hard time differentiating 320 from wav as well.
Another thing that puts me off is also the price of some of these hi res dac's. Especially given the audio differences and we get the old quote of ......... Diminishing returns.
Well, maybe I'm getting to the point of ignoring those diminishing returns since the premium you pay for those tiny returns can be quite enormous. When the gear starts to cost so much, I become a little suspicious about claims made in order to sell.
The T40 is damned close to way better headphones once tweaked, without the 'diminishing' returns so I have become rather more skeptical since you did that pair for me.
In the 70's, I was quite happy to use Senn hd414's. Not really hi res, but I had several pairs and they were used by me all of the time. I had no problem with inferior image or lower res problems of even cassettes. I just listened. I also had some Stax headphones, but hey, I stuck with the 414's!!
We seem more hung up on resolution nowadays than ever. Not the music really.
|
|
Dave
very active
Posts: 480
|
Post by Dave on Jul 10, 2014 19:42:03 GMT
One of the very few things that Frans and I have in common is that we both sometimes think we can tell when we are listening to hi-res or lo-res files but statistics seem to suggest that it isn't always the case. I have many hi-res files and even more lo-res (normal) files on my PC that I often play randomly and on some hi-res listening the statistics indicate that I am picking up differences - not sure what they are though - whilst on others I have to admit that it's frequently a guess and statistics support the proportion you'd expect from guessing. I don't feel quite so bad now about my crap hearing if Frans and Ian have similar experiences . For anyone interested, I play hi-res and lo-res files randomly using foobar and before the track ends I take a look at foobar to check my impression of the file resolution. Dave..
|
|
Rabbit
Administrator
Posts: 7,087
|
Post by Rabbit on Jul 10, 2014 20:34:10 GMT
That's the thing Dave, when someone says to me.... listen to this hi res version of ...... I hear it. But then again, I've been told what it is so it's slightly biased. Left to my own devices, I'm hard pushed to tell anything much between 320 and WAV. That's when hifi guys will tell me that my gear isn't up to it. Of course, I have never told people much about what I use so they are guessing. There is a reason for that ..... there will always be people who tell me I've wasted my money or there is much better for the money. Also, I wouldn't want to make people jealous. (That's supposed to be a joke ...... perhaps ) I have been told quite a few times that my gear can't possibly be up to the job when I have mentioned mp3, when in fact, the writers haven't a clue what I listen on. (Other than Ipods!!) My lowest res stuff is actually 320 MP3 going through various X-Cans.The rest of my gear is in the other house and there's not enough room in the new place for it, unless I'd like to be stared out by these big floorstanders that sit in the other place. However, since they must be such low resolution, maybe I'll shove them in a skip!!! (Along with the dozens of headphones that sit around in different places) I think it is actually quite hard to tell what the differences are with good rips. Poor rips or very low bitrate rips show blemishes sure enough and are dead easy to spot, but 320 has me in troubles more often than not. Maybe I should get a hifi now then? I must admit, there is something really alluring to having a digital player in the front room. Lovely and neat and the quality does seem ok to me. Perhaps a digital player going to some high quality active speakers would suit the new place. Where's techboy when you need him eh?
|
|