How Well Can You Hear Audio Quality?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
My opinion, based on lots of listening, is that the audibility of compressed formats like MP3 depends on a number of factors, including the bit rate of the MP3 and the type of music that is being encoded.

As a drastic example, I live near Sacramento CA. Our local NPR type station (KXJZ) has several different programs that come on during different times of the day or week, e.g. news, classical, jazz, acid jazz and new music, etc. The content is GREAT. Each "program" is available with extended playtime (e.g. 24/7) via free audio streams. About a year or so ago, some loser at the station made the executive decision that all online streams should go out at 96kpbs! That is totally fine for a "news" program that consists of people talking, but I absolutely cannot stand to listen to e.g. the jazz program at that low of a bitrate. Any time there is complex harmonics (e.g. piano, trumpet) it is totally obvious that something sounds "wrong", I mean in a way that makes it irritating - I honestly have to turn it off after no more than a minute or two. It's terrible! I no longer use an analog tuner (my system is either internet streams or ripped music) in my system, so I just stopped listening to all but the news stream now and then. Several times I have been contacted to donate to the station, or even to participate in an online focus group about the music streams. I have replied multiple times that the horrible sound quality forced me to stop listening to their streams, telling them that the low bit rate is the problem (I can see this as part of the stream info in my player) and BEGGING them to please go to 192k, saying I would even PAY for that. Nope! They still broadcast crap. So I stopped donating money.

In contrast, I can find plenty of free streams at up to 320k that are totally listenable and there is nothing about the sound that would make me guess that it is compressed. With jazz music anything at or above 256k seems fine to me. With classical music I have even found MP3 streams at 128k that sound great (but not all at this rate).

My home system streams audio to a number of loudspeaker systems. At one time I was trying to decide what sample rate to use, and was streaming audio as MP3 format and so I did a lot of listening to see just what bit rate was needed to sound good. My experiences are summed up in the last paragraph. Later I figured out how I could stream PCM audio, so I no longer use compression. But the time I spent checking into bit rates and formats (also tried streaming AAC, etc.) has made me feel comfortable with compressed formats as long as they are meeting some minimum rate for the type of music in the stream.
 
Last edited:
Disabled Account
Joined 2017
This test is completely invalid. I also got 2/6 while listening through my regular setup with Frugalhorns.

The comparison should be between two sources not three. It confuses the mind too much to be comparing between three sources of audio. I have to juggle the information that I'm processing between three differing soundtracks? Are you kidding me? The human mind cannot do that not in a million years and it would take hundreds of hours of listening and then even if you are lucky. Its difficult enough comparing just two sources.

I KNOW that I can hear the difference and I know the differences between 128/320 MP3 and WAV because I've listened to my own rips and compared them and I know exactly what to look for. However 320KBps MP3 is far too close to WAV to make a guaranteed decision on the differences and the difference between 320KBps MP3 and 128KBps MP3 is again too close to make a guaranteed decision. But even saying that I can tell the difference between two format bitrates and types. I'm not a dumbarse.

I've ripped and encoded MP3s with 128-160-192-320 and VBR and can tell the difference between all of them and that is the reason why I prefer VBR at about 200KBps avg because I know that it sounds almost indistinguishable from the original wav files. I've put a lot of time in doing these tests on myself aswell. But I am so well focused on the task that I can tell you the difference between 128-160-192-320 and VBR and WAV.

However making a decision between 3 sources leaves it too much up to chance that you'll hit the wrong answer because I'm trying to remember how three soundtracks sound and differ between one another. Much like winning the lottery.

The test is fundamentally flawed and for shame on NPR for setting it up, its designed to fool and disenfranchise everyone into believing that they cannot tell the difference between any file format and/or bitrates. It is a fatal test and should be burned in a fire then nuked from orbit.

I guarantee you that if the test was between two choices you and I would've gotten high if not perfect scores.
 
Last edited:
Sent to this to a buddy. He said:

"The test is rigged. If i listen to a Katy Perry song first, my ears will be irreparably damaged."

I must confess I chose three 320K mp3s and only two WAVs. :eek:
The 128K files sound pretty bad and were easy to eliminate in most cases, but I did pick one--the a Suzanne Vega clip. Sheesh.

In my experience, the most complex or "full" arrangements are the easiest to hear a difference, as the sheer quantity of information exceeds the 320K bandwidth. The Coldplay tune was the easiest one for me in this case.
 
Disabled Account
Joined 2017
in my own testing 128kbps actually sounds really good. Sonic-wise I would prefer it over 160 and higher bitrates if I didn't care about losing the definition because bitrates between 128 and 320 sound more like a half-assed attempt at reproducing the sound. Whereas 128kbps cuts off that level of detail at exact the right spot to make it not matter so much where the loss of detail begins.

However in my testing I couldn't tell the difference between wav and 320kbps mp3, it was at least exceedingly difficult. 128kbps mp3 is a 'sweet spot' which sounds quite good, to the point that I would prefer 128 over 160. But I would prefer 320 over 128 mainly because it adds more detail. The other bitrates between 128 and 320 are just about useless.

Which is just bizarre.

Now if you want to really blow your mind then have a listen to VBR OGG or 64kbps OGG, it beats 128kbps MP3 and gets real close to 320kbps MP3.
 
Last edited:
Fraunhofer spent many years developing the mp3 codec and that included live panels of listeners; each time they were able to tell the difference, they went back to the drawing board and refined the algorithm. Then another year in a contract with the National Hockey League (NHL) who used mp3's for stadium sounds (organ music, timeout music) ... yes, through stadium PA systems ... before they released it.

So, the mp3 codec was specifically designed to fool the ear. Despite that, some untrained people can hear the difference, and the 80% of the population that can appreciate music (10% have physical limitations; another 10% simply don't have the receptors in the brain that normally respond to music) can be trained to hear the difference.

That's without considering test anxiety where someone who truly knows the material still get failing marks (as proven conclusively via studies of college students).

Use your own system, create files of your choosing, and create a playlist with the random order enabled ("shuffle" in Apple speak). Be in a relaxed state of mind. You can go back and see the recents to determine which are which after you take your test.

So, those that score low would not be able to enjoy music as much as those that score high and quality audio equipment would be wasted on them?

Not true. Appreciation of sound reproduction isn't about brief snippets of highly compressed highly processed pop music ("highly compressed" as in near-zero dynamic range in the lossless version as well as the lossy version). Some tracks do sound OK on a high bitrate lossy compression scheme but the "shivers down your spine" tracks need realism and that "you are there" illusion to work. Fear not.

For the record I was 2/6 via the speakers in my Macbook Pro and 5/6 via my Ultimate Ears In-Ear-Monitors (IEMs). I took the first test at least six months ago but decided to revisit it, mostly due to this thread, for the second time yesterday, so I definitely couldn't remember which were the "right" choices.

I also have a bit of an issue with the "test" for "audio quality" as it's presented. There is only one track that has any pretence to audio quality.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.