Well, for this old guy it’s been bout 25 years since I could tell the difference between a 256 mp3 and a .wav file


https://www.npr.org/sections/therecord/2015/06/02/411473508/how-well-can-you-hear-audio-quality

Bud

Bud,

That was a really interesting exercise!

It turns out that, for the most part, I can't hear the difference between a 320 kbps mp3 and a wav. When I listened to the samples, the best that I could do was to narrow it down to the two best sounding ones. Then I couldn't choose between those two. The final decision was nothing more than a coin toss. Interestingly, out of the 6 times, I only chose wav once and the rest were 320 kbps mp3s.

That doesn't surprise me because I already know that I can't hear too much above around 13,000 Hz. I also know that one of the ways to increase overall mp3 quality is to filter out the high frequencies so that data bits are not needlessly used up in recording information that many people won't even hear. If my understanding is correct, this would mean that a 320 kbps mp3 that has been set to compress only those frequencies less than (say) 15,000Hz will probably sound the same to me as a wav since the bits of data in the mp3 have been used to maximise storing my 'listenable' frequencies of music.

Thanks for sharing the link.
Noel

Guys I am one step short of deaf. I have hearing aids, but I lost one at the barber shop so I only ware one in my right ear. When I mix, its a mess since I can't hear a broad range of frequencies... but I got 4 out of six right for the uncompressed wave. I can hear it. It is cleaner and clearer.

I have zip hearing above 8k. And a severe loss from 2k on. Beyond 8k mixing for me is a visual exercise. Thank you Izotope!

Bud

.

3 out of 6. I'm gradually losing high frequencies - probably genetic.

This is funny! I recently submitted something for a songtradr opportunity that made final selection that I didn’t even listen to! I was in a public place and had forgotten to bring my headphones! It just “looked” right. Lol!

This is a really neat exercise! I managed to pull a 5/6, getting Coldplay wrong. I still swear the claimed high quality audio sounds a bit muddy, but maybe that's just how it was mixed - or my headphones need replaced. Or my ears :P.

Thanks for sharing !

At 72 years young I picked 1 wav file while all the rest were 320 MP3s. After playing in bands for years, many standing next to a Leslie, I think that is pretty good.

I tend to be a very methodical person.

The first time, I took the test I got 3 wav, 2@320, and 1@128.

The next time, I did not. What I got would be irrelevant to the fact that I got a different answer.

The third time, again, a different answer.

In all honesty, for the most part, it felt like I was guessing. I was trying, and felt I was hearing various things...but then felt like a "shot in the dark."

So, I can't hear quality that well. lol That was what I suspected.

Thanks Bud!

That was fun! I did it once last night. Got the first four and missed the last two. I was confident in the first three, four was a little tougher but I could hear the difference and the last two I felt like I was guessing. I was really pleased with that since I was listening through cheap $4 mini headphones on my little android tablet in the living room while my husband was watching TV and the noisy a/c in there was on. I was like alright! Maybe the secret is cheap headphones.

I've had mild high freq hearing loss in my right ear since a blow to the side of my head when I was a kid. What bothers me most is when tinnitus flares up in that ear. It's really obnoxious noise to mix through. It's been so bad at times that I seriously thought about quitting music. After lots of googling, I discovered quite a few suffer from it and continue making their music by learning to work around it. So I've learned to get ear rest among other things. At times it's still hard to deal with if I have a nasty flare. It's like the more you focus on it the worse it gets and ignoring it isn't easy. I have to encourage myself to keep on keeping on. And discipline myself to take breaks when I'm mixing something and my ears are tired - easier said that done. The good days are great when it's not there or so low I don't notice it. The bad days suck. I'm still hanging in there. Maybe someday I'll figure out how to do it by sight.

I've spent too many years in high ambient noise jobs (paper mill and construction) and refused to wear hearing protection so I have lost most of my high-end hearing. When working with audio, I EQ to compensate. I often rely on BIAB's default settings and they have been very good. If I want to pull the soloist or particular instrument out front more, I just guess.

Well, I was feeling quite discouraged when I first took the test, I wasn't hearing much difference whatsoever!

I'm in my mid 60s so some high end loss due to age, plus I played in a band for many years which didn't help.

Then I read the post from Josie so instead of listening through my studio headphones I listened through a really cheap set of phones and could get most of them. I think maybe because the cheap set really brings out the high end where the mp3 artifacts live?

Noel, can you post a link about this? I don't know if most mp3 codecs are smart enough to neclect coding data above 15kHz.

I got 4/6. I felt the one that was the hardest was the acapella one. It was easier for me to identify the ones that had a busier sound to the song.

I seem to recall a brick ceiling for MP3 at 16k .. let me do some digging.
Be right back

That brick ceiling seems to have been at the 128-192k conversion rates.
320 gets a bit higher.

https://thesession.org/discussions/19642

Funny how we're all different Ember. It was exactly the opposite for me. I couldn't hear any difference at all with Neil Young and Coldplay. The 128K Mp3 sounded just as good as the wav to me. But then I got the girl singer, Katy Perry, Jay Z and the classical piano one because they're cleaner, simpler mixes. I was listening for overall presence not specific freq's. The ones I got were obvious to me. I played these through my fairly high end stereo system with a pair of Altec Model 14's, not cans. I never mix with cans I only use them as a test later.

What this really shows is unless you're all set up in a good listening room, no distractions and using good equipment you would never hear any difference at all. This required silence and concentration and who really listens that way? Unless I'm playing mixing engineer I don't listen critically, I'm listening for relaxation and enjoyment and I'm actually surprised that I would be perfectly happy with the 128K Mp3's for that.

It's millennials who could really tell the differences if they actually cared. Sadly I doubt many would. Back in my day me and all my friends were all into high end and expensive stereo equipment. We could tell the differences between turntable cartridges, mid range speakers and high end ones, running tape at 15ips vs 7.5 and all that stuff.

Bob

Bob,

I pretty much agree with you on this. For the most part, to the degree possible for each person, musicians have trained ears - generally know what to listen for. Most audience members do not and really don't care. They just want you to play their favorite song - technical stuff totally irrelevant.

And yes, about the listening and recording gear back in the '70s, into the early '80s. There was always the discussion of what was better: Sansui, Pioneer, Fischer, etc. I had a complete Marantz system - 7 or 8 pieces. Even had a nice Gerard turntable at one point.

Alan

Hmmmmm. Not so sure this test is all that valid since the samples are streaming.

It should say "how well can you tell audio quality once it has been crapped out by streaming."

I can say that in a non streaming format I can tell a huge difference between a .wav and an MP3 and a large difference between a 256K MP3 once it has been posted to soundcloud and a 320K MP3.

Once a 256k has been posted to soundcloud it sounds "watery" to me. A 320K is better but still leaves a lot to be desired, but again, that is when it is streaming, not so much if it is playing off your computer or another medium.

For some reason I cannot stand .wmv files and am in the process of re-ripping a whole bunch of CDS I accidentally ripped as wmvs a while back. I can't stand to listen to them. I find them quite worthless as a matter of fact, and I am not sure why. They just don't sound right to me. Very thin.

I got most of these right, but again, I don't think "testing" audio quality from a streaming application makes a whole lot of sense.

It's like saying:

Lick these two apples wrapped in cellophane and tin foil and tell me which one is rotten.

Sure they are “streaming.” But they are being buffered so that they can play at a given quality w/o being downloaded. Note that the wav file takes longer to buffer and begin playing as one would expect from a larger file. And that’s a way to cheat on the test by noting that difference! I would not think NPR engineers would add any compression nor would your computer. You should be playing them exactly as if you downloaded them. FWIW!

Ummmmm, I am not so sure man.

I don't know the exact science of it dude, but the .wav files I hear when I play them on my computer or home audio do not sound anything at all remotely like the .wav files I hear streaming from NPR in this scenario.

They are 1000 times richer and more dynamic.

I have no idea why but they just are.

The important part for me is not what I hear on an NPR site "test"--it is what I hear in my own studio, and I know for a fact that I would not even think about posting a 128K MP3 anywhere because even here it sounds horrible, and once it is streaming it is not listenable. Learned that immediately.

Anyway, not worth a rumpus of any sort, just doesn't sound remotely like anything but a streaming watered down file to me, all of them.

I know what I hear is all. But I think everybody should just listen to what they want to listen to with the ears they have and say whatever they want to say because all of this is so subjective--in the end, there is no real absolute truth in this zone, just perceptions and opinions.

Moving right along!! Peace and joy to all ears!!!

David, to get the best quality sound from soundcloud you should upload WAV files.

Soundcloud automatically encodes anything you upload to (I think) 128K MP3 for streaming playback, so if you upload an MP3 it will get encoded twice.

From article https://help.soundcloud.com/hc/en-us/articles/115003452847

"Please be aware that we transcode all tracks to various codecs which are optimised for streaming playback. We recommend uploading uncompressed or lossless audio files to ensure this process results in the best possible quality."

Yep. We’ve never uploaded anything but wav’s to SC from the get go. Double compression is artifact heaven.

Thanks guys. Good advice on soundcloud. I found out many things there by trial and error before I read the instruction manual. Hey, I'm a guy.

My original point was about stuff I found out years ago when I was even more ignorant than I am now. That is, in sum:

A.) There is a massive difference between a 128k MP3 and a 320k MP3 whether you upload them or not.

B.) If you upload a 128k it will sound horrible.

C.) Yes you should upload .wavs when you can. Know that now.

D.) I am not at all convinced that a streaming anything will sound as good as the original, or something that is downloaded. Streaming stuff always sounds weak and watery to me. Always, in all circumstances. It just does. But it may just be me, and there may be something really wrong with my head, and myself in general, but that is what I hear. Streaming does not sound as good as downloaded or an original file to me, ever.

Peace out gents. Gotta go feed the machine and pay the mortgage.