Head-Fi.org › Forums › Equipment Forums › Sound Science › 320 kbps MP3 vs. normal audio CD listening Sound quality
New Posts  All Forums:Forum Nav:

320 kbps MP3 vs. normal audio CD listening Sound quality - Page 30

post #436 of 516

ok...wher can you get 320kbps music other than from a cd directly to my android?? Ive tried to reaseach on google but i could find a solid website.

post #437 of 516

Both MOG and Spotify subscription music services offer a premium streaming bitrate that is 320kbps.  Also, Google Music will allow you to upload your own songs up to 320kbps which you can play back via the Android Google Music app.

post #438 of 516
Quote:
Originally Posted by sonitus mirus View Post

Both MOG and Spotify subscription music services offer a premium streaming bitrate that is 320kbps.  Also, Google Music will allow you to upload your own songs up to 320kbps which you can play back via the Android Google Music app.

 i live in canada, mog and spotify dont work in canada

on the other hand how the the google music work? do you need data when you use it? and can you download music from it?


Edited by twg1996 - 2/9/13 at 5:16pm
post #439 of 516

uhh.... i just plug my android into the com and transfer it that way.
or use a sync app and transfer from com to android through wifi.

post #440 of 516
Quote:
Originally Posted by Brooko View Post
 

 

Very few people actually submitted data from a true abx test though.  Of those who did, I think there were only two that actually reliably could tell the difference - and neither said it was 'night and day' ( a phrase I personally dislike - and one that seems to be thrown about readily among the misinformed IMO).  The test is easy to set-up as well > Foobar2000 + abx comparator plugin.  Two files - volume match using the automatic plugin tool,  Run the abx (minimum 15 iterations - 20 is better).  Produce the log.

 

It's funny - but when people are asked to do this - they either ignore it, or claim they can pass the test (one listen), and then promptly ignore any requests for proof.  If you believe the 2/3 - then we have an awful lot of golden ears here wink.gif.

 

IMO - most people will fail a test if they do it honestly - they're actually supposed to.  The funny thing is that the ones that refuse to at least find out the truth are in effect living in denial.  A very few (usually with well trained ears) may be able to tell the difference - but the vast majority of us won't.  Knowing the truth is in a way liberating.

 

Now I'm not advocating not using lossless, and converting all your files to lossy.  I use FLAC for everything at home - simply so that I have a copy of the original file (if I ever need to transcode, or if there are breakthroughs in compression standards).  But I know I personally can't tell the difference between AAC256 and lossless - so it makes decisions especially on formats for space limited portables really easy.

 

What I am suggesting is that the constant stream of people who claim to "easily tell the difference" is sad.  It's time a large portion of us stopped lying to ourselves - and actually find out instead of guessing.  There is no shame in having 'normal' hearing biggrin.gif

What's a good result for this test?

 

I was rather dismayed to score 10/20 using an Askman 24-96 FLAC copy of Outlandos D'Amour vs a 320 MP3 LAME taken from the FLAC.

 

However I was at 9/15 at one point and it really did seem like I could tell the difference. The way I did it was by listening to my heart really. One version excited me the other didn't. I got 3 wrong in a row before that point but that was when I wasn't listening to both x and y, just one or the other, I plan to repeat the test at some point.

 

But the thing is I don't know if I was really listening to bits of the song that were good to compare and my short term memory is a  bit pants.

 

As the test went on I was feeling really bored and my ears were getting tired, that could be why my results dropped off towards the end. Listening to the same bit of track 30 times in a row....jeez your hearts really not in it by then.

post #441 of 516
Quote:
Originally Posted by V-Moda M80 View Post
 

What's a good result for this test?

 

I was rather dismayed to score 10/20 using an Askman 24-96 FLAC copy of Outlandos D'Amour vs a 320 MP3 LAME taken from the FLAC.

 

However I was at 9/15 at one point and it really did seem like I could tell the difference. The way I did it was by listening to my heart really. One version excited me the other didn't. I got 3 wrong in a row before that point but that was when I wasn't listening to both x and y, just one or the other, I plan to repeat the test at some point.

 

But the thing is I don't know if I was really listening to bits of the song that were good to compare and my short term memory is a  bit pants.

 

As the test went on I was feeling really bored and my ears were getting tired, that could be why my results dropped off towards the end. Listening to the same bit of track 30 times in a row....jeez your hearts really not in it by then.


Well 12 of 15 would be a 95% confidence level of not guessing though only 15 trials is a bit sketchy.  Your talking about not caring by then is a part of the problem.  Doing 10 trials over 3 sessions would make more sense though it isn't convenient.  21 of 30 would also meet the 95% confidence level.  As you increase the number of trials it takes less to show you are not just guessing. If over time you did 100 trials you need only score 60 of 100 to meet the 95% confidence bar.

 

But just randomly streaks among 12 to 15 trials is not as uncommon as you would think.  They always leave you thinking, "hey I was almost there, must be something to it, I know I was really hearing something".  Most likely not the case, but if you string together several short sessions over time you can either regress toward the 50/50 level or show that you can hear something going on different.

post #442 of 516

To get a statistically relevant result - you'd need to be running at least 15 trials, and scoring 95% +.  Anything below that, and really you're guessing.  10/20 or 9/15 both show same result - you can't tell the difference.  Don't worry - it just means you are normal ;)

 

Best way to do it is try and run 10-15 at a time, with a rest between, and accumulate results.

 

EDIT : esldude's reply (above) nails it.


Edited by Brooko - 5/4/14 at 9:30pm
post #443 of 516

Thanks for the replies esldude and Brooko. Do either of you have any recommendations for songs that are good for this test? Or perhaps parts of songs that are markedly different in 320kbps MP3 LAME compared to FLAC but could only be detected by someone able to tell the difference? I do plan on repeating this test again.

 

 

Quote:
Originally Posted by esldude View Post
 

Your talking about not caring by then is a part of the problem. 

 

It wasn't that I was not caring, just that by then it was getting a bit tedious. I like to cycle my songs so that they stay fresh, doing the same thing time and time again is against my nature at least without a break in between.

 

For my next test I'll likely compile my results from doing sessions of 10 or perhaps 15 at a time.

 

I'm really interested by this, I thought I had really good hearing but I don't want to be lying to myself!

post #444 of 516
Quote:
Originally Posted by V-Moda M80 View Post
 

Thanks for the replies esldude and Brooko. Do either of you have any recommendations for songs that are good for this test? Or perhaps parts of songs that are markedly different in 320kbps MP3 LAME compared to FLAC but could only be detected by someone able to tell the difference? I do plan on repeating this test again.

 

 

 

It wasn't that I was not caring, just that by then it was getting a bit tedious. I like to cycle my songs so that they stay fresh, doing the same thing time and time again is against my nature at least without a break in between.

 

For my next test I'll likely compile my results from doing sessions of 10 or perhaps 15 at a time.

 

I'm really interested by this, I thought I had really good hearing but I don't want to be lying to myself!


By problem I meant the tedium of it.  Not that you didn't actually care, just music gets stale.  Completely normal reaction to such a test.  Which is why real research usually involves lots of people with each doing a handful of trials.  I have suggested it elsewhere, and it isn't the same with music.  Pink noise is easier to hear differences on than music.  More discerning if you will.

post #445 of 516
Quote:
Originally Posted by V-Moda M80 View Post
 

Thanks for the replies esldude and Brooko. Do either of you have any recommendations for songs that are good for this test? Or perhaps parts of songs that are markedly different in 320kbps MP3 LAME compared to FLAC but could only be detected by someone able to tell the difference? I do plan on repeating this test again.

 

It wasn't that I was not caring, just that by then it was getting a bit tedious. I like to cycle my songs so that they stay fresh, doing the same thing time and time again is against my nature at least without a break in between.

 

For my next test I'll likely compile my results from doing sessions of 10 or perhaps 15 at a time.

 

I'm really interested by this, I thought I had really good hearing but I don't want to be lying to myself!

 

Hydrogen Audio has some threads with references to the very few "killer tracks" that LAME has trouble with encoding correctly - http://www.hydrogenaudio.org/forums/index.php?showtopic=83905

 

As far as your testing goes - here's what I'd recommend you do next.

 

  1. Start with music you know really well.  Limit yourself to less than 10 tracks total.  Cover a wide range of genres, styles. etc.
  2. If possible start with something relatively high-res (say 24/96).  Compare to redbook. MP3 320, mp3 V0, aac256.  The whole aim of this is discovering your own thresh-hold for what is discernible.  Mine is ~ AAC 200, so I simply use aac256 for all my portable listening now.
  3. Once you know your thresh-hold, forget about what everyone else says - and simply encode your music for your own (now known) limits.  Celebrate your ability, rather than lamenting the ability not to be a "golden-eared listener".  Remember - you're actually normal in your limitations.  I use FLAC at home simply because there is no reason not to.  I rip my CD's to redbook.  It means I have a completely lossless copy for archiving - and I use the same archive for listening.  For my portable, I mostly use aac256 - simply because I know I can't tell the difference, and it allows me to carry as much music with me as possible - without having to buy an uber-expensive DAP. 

 

Once you know your own limits - then you can safely relax and not worry about music formats - but instead just enjoy the music itself.  And that is the whole key to the exercise IMO - enjoying the music.


Edited by Brooko - 5/5/14 at 3:32pm
post #446 of 516

I honestly preferred Flac (CD) to 320kbps with headphones, I did actually feel I heard new things, and the bass and highs really seemed to hit harder.

 

Although, for lower-end systems, I think 320kbps is actually superior due to the compression. Low end speakers & headphones don't really pick up the highs and lows, so when it's compressed, they're more audible. Now, for the simple fact flac take up 350% more space...it's not really worth it unless you've somehow got 500gb+ storage on your Mp3 player.

 

Although I'm not allowed to post links...

 

I listened to "Burn the Sky Down" by Emma Hewitt to compare.


Edited by Hapster - 5/5/14 at 3:42pm
post #447 of 516

Have you done a level matched double blind test? Your description of the differences between the 320 and FLAC don't sound like any compression artifacts or compression loss that I've ever heard, but it does sound a lot like either a slight volume mismatch (with the FLAC a bit louder) or a simple expectation bias. That having been said, it is theoretically possible to hear a difference (though an extremely subtle one) between 320 and lossless on some tracks, so it is possible that you are hearing a real difference.


Edited by cjl - 5/5/14 at 4:40pm
post #448 of 516
Quote:
Originally Posted by Hapster View Post
 

I honestly preferred Flac (CD) to 320kbps with headphones, I did actually feel I heard new things, and the bass and highs really seemed to hit harder.

 

It depends on what kind of codec you used. I believe there might be a slight edge for lossless over 320 Frauenhofer. But 320 LAME and 320 AAC should be totally indistinguishable.

post #449 of 516

Well, I wasn't the one who converted it so unfortunately I don't know what kind of codec was used.

post #450 of 516
Quote:
Originally Posted by Hapster View Post
 

I honestly preferred Flac (CD) to 320kbps with headphones, I did actually feel I heard new things, and the bass and highs really seemed to hit harder.

 

Although, for lower-end systems, I think 320kbps is actually superior due to the compression. Low end speakers & headphones don't really pick up the highs and lows, so when it's compressed, they're more audible. Now, for the simple fact flac take up 350% more space...it's not really worth it unless you've somehow got 500gb+ storage on your Mp3 player.

 

Although I'm not allowed to post links...

 

I listened to "Burn the Sky Down" by Emma Hewitt to compare.

 

Hi Hapster.  Thanks for your post.  I underlined a couple of things above, because the way you've worded them suggests that you haven't abx'd the two formats properly (level matched and blind).

 

It's entirely up to you if you want to spend the time doing it - and all it will help is your own personal knowledge.  I do think it's worth it.  Most of the people who invest the time to test themselves admit to it being an "eye opener".  If anything it lets you know your own limitations.  It'll probably also show that you are one of us (ie normal human beings).

 

Here's a link on how to do it.  The software is free.  All you need is access to a PC, the internet, and a CD.

http://www.head-fi.org/t/655879/setting-up-an-abx-test-simple-guide-to-ripping-tagging-transcoding

 

Important things are using the same source file for comparison, volume matching, and making it blind (eliminating placebo).

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › 320 kbps MP3 vs. normal audio CD listening Sound quality