Page 1 of 3 123 LastLast
Results 1 to 15 of 31

Thread: MP3s sound like crap through speakers

  1. #1
    Senior Member millionrainbows's Avatar
    Join Date
    Jun 2012
    Location
    USA
    Posts
    14,798
    Post Thanks / Like
    Blog Entries
    137

    Default MP3s sound like crap through speakers

    That's right, I can tell a pronounced difference. MP3s were designed for listeners who listen to portable devices with ear buds, and need storage space.

    I remember the first time I heard a lo-rez MP3 on a stereo, and I heard all that
    "phasing" in the high frequencies, on cymbals and stuff. I thought to myself, "Why on Earth would anybody want to listen to this crap?"
    Last edited by millionrainbows; Jun-02-2019 at 22:11.

  2. #2
    Senior Member
    Join Date
    Nov 2011
    Location
    Hollywood U.S.A.
    Posts
    5,883
    Post Thanks / Like
    Blog Entries
    1

    Default

    You haven't done a careful, controlled listening test yet.

    It just so happens that I have just such a test that I would be happy to share with you. It contains a sample of particularly difficult to encode music encoded to Fraunhofer MP3, LAME MP3 and AAC at 192, 256 and 320... along with the original lossless sample. Ten samples in all in a variety of compression codecs and levels. If you would like to take the test, I will send you a FLAC file containing all ten samples. All you have to do is to carefully listen to all ten and rank them from best (1) to worst (10). You send me your rankings, and I'll tell you which samples were which. It's a fun test and it will allow you to directly compare using your own stereo equipment and ears.

    Would you like to try it and find out how much difference there actually is? Just let me know and I will send you a download link for the test file.

  3. Likes Merl liked this post
  4. #3
    Senior Member millionrainbows's Avatar
    Join Date
    Jun 2012
    Location
    USA
    Posts
    14,798
    Post Thanks / Like
    Blog Entries
    137

    Default

    OK, send the file link, and I will tell you if I think the nature of the material is suited for testing.
    Last edited by millionrainbows; Jun-03-2019 at 15:02.

  5. #4
    Senior Member
    Join Date
    Nov 2011
    Location
    Hollywood U.S.A.
    Posts
    5,883
    Post Thanks / Like
    Blog Entries
    1

    Default

    There are some basic ground rules that go with the test...

    You have to take it in good faith. The purpose is to learn something about your own threshold of audibility, not for argumentative purposes about how the test is conducted, whether the tracks are well chosen, etc. If you don't want to know the truth, don't ask to take the test.

    There will be ten samples in random order... Lossless, Fraunhofer MP3 192, 256, 320, LAME MP3 192, 256, 320, and AAC 192, 256, 320. You have to rank all ten samples from best to worst. If you don't rank them all, I won't give you your results.

    This is a listening test. No opening the file in a sound editing program. People have tried to cheat the test, but I have ways of telling if you do that. You're free to listen to the tracks however you want on whatever equipment you want as long as you listen in a normal manner... normal listening level, music playing through, not looped on tiny sections or gain ridden. The idea is to see if you can hear a difference under normal listening conditions.

    The material consists of a classical choral work and an orchestral work selected by a golden eared audiophile who chose these samples because they were particularly difficult to encode without artifacting. I've shared this test with dozens and dozens of people. I'll let you know the results after you take the test. It is a blind test.

    You have as long as you'd like to take the test, but listening longer is just going to make you tired of the clips. It won't allow you to hear the artifacting any better. You should be able to take the test in 2 hours.

    I can give you the test file in either FLAC or ALAC. Your choice. Let me know.

    By the way, if anyone else would like to take it, just let me know.

  6. #5
    Senior Member Larkenfield's Avatar
    Join Date
    Jun 2017
    Location
    Sedona
    Posts
    4,272
    Post Thanks / Like
    Blog Entries
    4

    Default

    It also depends on the specific bit rate. I would consider low rez to be 96 kbps, with 128 a close second. Who listens at 96 or even 128 any more? I find 256 to be acceptable even through audiophile headphones with the sound quality becoming deeper, richer and more detailed the higher up one goes in bit rates.
    Last edited by Larkenfield; Jun-03-2019 at 21:46.
    "That's all Folks!"

  7. #6
    Senior Member Larkenfield's Avatar
    Join Date
    Jun 2017
    Location
    Sedona
    Posts
    4,272
    Post Thanks / Like
    Blog Entries
    4

    Default

    The only blind tests I would give any credence to would be with music that one is completely familiar with backwards and forwards to notice the subtle changes. But even that can be a problem because the music memory for comparing two different tracks is extremely short for most listeners... By the time you've heard the second sample, you've already lost the memory of the first.

    Do your own listening tests. Start with a track that has a low bit rate of 64—then listen; then go up to 96—listen... 128... 256.. 320... and Lossless... and you can notice the noticeable and subtle differences yourself as the sound takes on additional details, richness and depth, not to mention that the sound floor gets quieter the higher up you go. Yes, it does. Then you'll know from then on that the higher the bit rate the better quality of sound even if you can't hear it at the conscious level and the intangibles that can only be sensed or felt though not necessarily heard at the conscious level. But the so-called sound experts don't want you to do that because they want to prove you wrong and tell you that there are no discernable differences in sound between certain bit rates when that's false upon closer examination. No one needs the "experts" when you can test different bit rates with music that's familiar to you and you can hear the subtle differences yourself. It's great ear training and you can take notes on the highlights of the music according to different bit rates because different instruments are sometimes more noticeable or subtly highlighted at different rates, such as the percussion instruments. From personal listening experience.
    Last edited by Larkenfield; Jun-04-2019 at 06:24.
    "That's all Folks!"

  8. #7
    Senior Member jegreenwood's Avatar
    Join Date
    Dec 2015
    Location
    NYC
    Posts
    2,339
    Post Thanks / Like

    Default

    Quote Originally Posted by Larkenfield View Post
    It also depends on the specific bit rate. I would consider low rez to be 96 kbps, with 128 a close second. Who listens at 96 or even 128 any more? I find 256 to be acceptable even through audiophile headphones with the sound quality becoming deeper, richer and more detailed the higher up one goes in bit rates.
    My favorite internet radio station streams at 32kbps.

  9. #8
    Senior Member
    Join Date
    Nov 2011
    Location
    Hollywood U.S.A.
    Posts
    5,883
    Post Thanks / Like
    Blog Entries
    1

    Default

    Quote Originally Posted by Larkenfield View Post
    I find 256 to be acceptable even through audiophile headphones with the sound quality becoming deeper, richer and more detailed the higher up one goes in bit rates.
    The differences between higher data rates don't involve deepness or richness or even detail. It involves the frequency and degree of artifacting. Low data rates are unable to reproduce certain sounds without artifacts... digital splats, warbling and distortion. At some point, increasing the data rate will reduce the number of artifacts to the point where there is no audible difference from lossless any more and it becomes audibly transparent. You can increase the data rate beyond that point, but it will just be redundant data filling out the file size- no difference in perceived sound quality. The impression that higher data rates sound richer and deeper is likely placebo effect. If you did a controlled test, those differences would likely disappear.

    Quote Originally Posted by Larkenfield View Post
    The only blind tests I would give any credence to would be with music that one is completely familiar with backwards and forwards to notice the subtle changes. But even that can be a problem because the music memory for comparing two different tracks is extremely short for most listeners... By the time you've heard the second sample, you've already lost the memory of the first.
    If you want to find the absolute limits of your hearing, you need to do multiple trials with a wide range of music comparing lossless and lossy that have been level matched, direct A/B switched and blind. That will give you a finite and clear line. I did this before I committed to a codec and data rate for my music server. It took about a week. Different codecs perform differently at different data rates. As you say, you need to compare various codecs and settings to know what you can hear and what you only imagine you can hear.

    The test I am sharing is different than that though. It has three codecs and three different settings, but they are housed in a single file to make it easy to listen to straight through- just like we generally listen to music in the home. It's a simple, casual test to see if there are significant differences that really matter, not to find the absolute limits of perception. There's a purpose for general tests too. Especially when audiophiles speak of "veils over sound", "detail", "soundstage" and "night and day differences" that are all likely chalked up to bias and placebo. Every human being is subject to bias. Blind testing is the way to eliminate that from influencing comparisons.

    The title of this thread is "MP3s Sound Like Crap Through Speakers". My casual test is a quick and easy way to determine if that statement is true or not. Crap should be obvious even on casual comparison, right?
    Last edited by bigshot; Jun-03-2019 at 23:19.

  10. #9
    Senior Member
    Join Date
    Jul 2012
    Posts
    7,282
    Post Thanks / Like

    Default

    Quote Originally Posted by bigshot View Post
    The title of this thread is "MP3s Sound Like Crap Through Speakers". My casual test is a quick and easy way to determine if that statement is true or not. Crap should be obvious even on casual comparison, right?
    Having very limited technical knowledge about the production of mp3s, all I can say is that most of the mp3 rips I have on my PC sound fine, thank you very much. I have come across mp3s that sound like crap, and had assumed that it was the bit rate that was the issue, but whatever, the difference was like day and night!

  11. #10
    Senior Member millionrainbows's Avatar
    Join Date
    Jun 2012
    Location
    USA
    Posts
    14,798
    Post Thanks / Like
    Blog Entries
    137

    Default

    Quote Originally Posted by bigshot View Post
    The differences between higher data rates don't involve deepness or richness or even detail. It involves the frequency and degree of artifacting. Low data rates are unable to reproduce certain sounds without artifacts... digital splats, warbling and distortion. At some point, increasing the data rate will reduce the number of artifacts to the point where there is no audible difference from lossless any more and it becomes audibly transparent. You can increase the data rate beyond that point, but it will just be redundant data filling out the file size- no difference in perceived sound quality. The impression that higher data rates sound richer and deeper is likely placebo effect. If you did a controlled test, those differences would likely disappear
    You don't understand the MP3 Codec: it's psychologically-based. How can that be "objective" if it is designed to "give the ear/brain" what it thinks it wants? That's truly high-grade snake oil.

    Remember: "bits is bits." Information has been removed! But I guess that old saying only applies when it is convenient.
    Last edited by millionrainbows; Jun-12-2019 at 16:08.

  12. #11
    Senior Member apricissimus's Avatar
    Join Date
    May 2013
    Location
    Malden, MA, USA
    Posts
    682
    Post Thanks / Like

    Default

    Quote Originally Posted by millionrainbows View Post
    You don't understand the MP3 Codec: it's psychologically-based. How can that be "objective" if it is designed to "give the ear/brain" what it thinks it wants? That's truly high-grade snake oil.

    Remember: "bits is bits." Information has been removed! But I guess that old saying only applies when it is convenient.
    Just because information has been removed, that doesn't mean that a) the information that has been removed is meaningful or important, and b) even if it is "important" in some sense, that we can tell the difference.

  13. Likes bigshot liked this post
  14. #12
    Senior Member Larkenfield's Avatar
    Join Date
    Jun 2017
    Location
    Sedona
    Posts
    4,272
    Post Thanks / Like
    Blog Entries
    4

    Default

    One of the most disappointing things in all of music are those who don’t mourn the beauty of the continuous sound wave, not one with square corners, not with those sound frequencies lost forever as if it really doesn’t matter and doesn’t make any difference, that you can’t hear the difference in sound – and I’ve never found that to be true. The best, most authentic sound will always be analog. Why? Because life is analog sound and some can tell the difference. Imagine if life was only conducted through MP3 at 128 kbps. How thrilling! And yet it’s accepted around here by those who cannot understand what’s been lost and give out misinformation on an almost daily basis.
    Last edited by Larkenfield; Jun-13-2019 at 08:17.
    "That's all Folks!"

  15. #13
    Senior Member
    Join Date
    Nov 2011
    Location
    Hollywood U.S.A.
    Posts
    5,883
    Post Thanks / Like
    Blog Entries
    1

    Default

    See Nyquist. There are no "square corners" in digital audio. That is a myth. A square wave is an illegal signal in digital audio, and it doesn't exist in real world sound.

  16. #14
    Senior Member
    Join Date
    Nov 2011
    Location
    Hollywood U.S.A.
    Posts
    5,883
    Post Thanks / Like
    Blog Entries
    1

    Default

    Quote Originally Posted by millionrainbows View Post
    You don't understand the MP3 Codec: it's psychologically-based. How can that be "objective" if it is designed to "give the ear/brain" what it thinks it wants?
    Humans do not perceive everything in the world. We perceive some things and we don't perceive others. Compressed audio is based on auditory masking. We are unable to hear two sounds at the same time in certain circumstances. Lossy codecs eliminate what we can't hear without affecting what we can.

    I have a test of three lossy codecs at three different data rates, along with a lossless sample. If you would like to take an objective listening test, I would be happy to assist you with that. This test helps you determine your threshold of transparency. Transparent means you can't hear a difference any more. Sound beyond the threshold of transparency all sounds the same, because any differences are beyond your ability to hear.

  17. #15
    Senior Member millionrainbows's Avatar
    Join Date
    Jun 2012
    Location
    USA
    Posts
    14,798
    Post Thanks / Like
    Blog Entries
    137

    Default

    Quote Originally Posted by bigshot View Post
    Humans do not perceive everything in the world. We perceive some things and we don't perceive others. Compressed audio is based on auditory masking. We are unable to hear two sounds at the same time in certain circumstances. Lossy codecs eliminate what we can't hear without affecting what we can.

    I have a test of three lossy codecs at three different data rates, along with a lossless sample. If you would like to take an objective listening test, I would be happy to assist you with that. This test helps you determine your threshold of transparency. Transparent means you can't hear a difference any more. Sound beyond the threshold of transparency all sounds the same, because any differences are beyond your ability to hear.
    Okay, if I can choose the input sources, with music I am very familiar with, with different 'densities' of information.
    I.e., I want to see if the Buzzcock's remasters sound any different.

    The first time I ever heard CD sound was in a mall, at Christmas. The recording was a solo penny whistle playing Christmas songs, and it was incredibly realistic.
    I also remember hearing snippets of movie sounds in Best Buy, advertising their surround-sound systems. I remember thinking how "artificial" it sounded when it stopped; having had previous experience with digital delays, it struck me as similar; a very brittle, artificial sound.

    This is the difference in "density" and how it affects perceived resolution.
    Last edited by millionrainbows; Jun-13-2019 at 16:34.

Page 1 of 3 123 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •