Anybody Out There Who Hears a Difference Between 320 kbps MP3 and Red Book CD? What Differences Do You Hear?

[JJB70]

I agree with those who can discern a difference if they listen with the objective of discerning a difference but that for normal enjoyment of music the difference is too subtle to have much impact and is pretty much unnoticeable. There is a difference but it is a subtle one and my advice would be that if you can't hear it then don't look for it.

 

[Patrick1958]

For my smartphone i convert my music to AAC LC 512 Kbps for 44.1Khz or 576 Kbps for 48Khz. I use the Nero codec. To my ears transparent, can not distinguish if from RBCD or Flac.

On My DAP's i use original RBCD, Flac, HR or SACD.

 

[Phronesis]

So if someone needs focused attention and maybe some training to detect a barely perceived difference between 320 and CD in most (not all) trials of a blind test, maybe we can agree that the difference is ‘audible’ but ‘subtle’. That leads to the question of whether the difference matters for enjoyment of music. If I listen to 320 some days and CD some days, will I consciously or subconsciously enjoy the music more on the CD days? Would brain imaging confirm that my brain responds differently on the CD days? I don’t even have a guess about the answers, but just in case it matters, if I find something I like on Spotify, I’ll usually preferably listen to it using Tidal hi-fi.

 

[SIY]

Yeah, that "night and day" stuff is nonsense; "subtle" is exactly correct. I don't have training, but with concentration and the right source material, I was able to distinguish between 320 and FLAC 16/44 double blind. No way I could do that without direct comparison and rapid switching.

 

[Phronesis]

Yeah, that "night and day" stuff is nonsense; "subtle" is exactly correct. I don't have training, but with concentration and the right source material, I was able to distinguish between 320 and FLAC 16/44 double blind. No way I could do that without direct comparison and rapid switching.

I've had the experience of consistently perceiving night and day differences between gear for months, and those differences were consistent with what many others consistently reported. But when I did sighted testing with matching of volume and music segments, using short music segments and very quick switching, those differences disappeared - I could no longer consistently perceive any difference at all. I was shocked.

I didn't even bother taking the next step of blind testing because I concluded that any differences which were really there were subtle at most, at least for my ears/brain. I'm inclined to think that those subtle differences don't matter much for enjoyment of music, or are at least swamped by differences in recordings, transducers, rooms, mood, etc.

All of this said, I'm not ready to conclude that a negative result on a blind test shows that a listener would perceive no difference between A and B in normal listening, so I leave room for the possibility of a subtle difference which could be perceived more subconsciously than consciously. But I do think that a null result on a decently-conducted blind test (or even sighted test with the kind of controls I noted above) would provide very strong evidence that there's no night and day difference.

I guess our auditory perception was evolved for purposes other than making fine distinctions in audio gear!

 

[vert]

People should be left to use whatever they want, and left to decide for themselves, without interference from corporations, whether differences between 320 kbps and CD quality are "negligible". In 2018 storage space can be expanded easily and affordably on most phones, except iPhones.

 

[Phronesis]

People should be left to use whatever they want, and left to decide for themselves, without interference from corporations, whether differences between 320 kbps and CD quality are "negligible". In 2018 storage space can be expanded easily and affordably on most phones, except iPhones.

I agree. Though it would be nice to better pin down which differences are audible, to what degree, for what percentage of listeners, etc., so that people can make informed decisions rather than being misled and misleading each other based on misperceptions due to biases, agendas, etc. I often see people stretching their finances to buy gear 'upgrades' which likely make little or no difference.

 

[Hugo9000]

The key difference is obvious to an untrained ear, and is certainly not negligible.

CD: Music plays seamlessly
MP3: pop or click or brief inserted silence throughout an opera recording or other music that has tracks programmed within a musical selection

The lack of gapless playback with MP3 alone is enough to make it garbage to me, even if the codec allowed for full fidelity otherwise to the source music file. So I don't even bother listening for other anomalies with MP3. It should never have gotten out of the gate for anything like classical or the odd pop/rock album with seamless transitions between tracks or songs.

Ogg Vorbis, on the otherhand, was designed to allow gapless playback. So if I had to have something lossy for space or other considerations, that's what I'd choose.

 

[bravomail]

I did my own testing of Flac, mp3 320k CBR, opus 320k, AAC 320k ABR in Foobar2000 with Deep Purple composition Hush. Opus went down in flames with it cutting both low and high freqs. AAC won subjectively. u can try your own testing.

 

[AnalogSteph]

The lack of gapless playback with MP3 alone is enough to make it garbage to me, even if the codec allowed for full fidelity otherwise to the source music file. So I don't even bother listening for other anomalies with MP3. It should never have gotten out of the gate for anything like classical or the odd pop/rock album with seamless transitions between tracks or songs.

Gapless playback of MP3s has basically been a solved problem for many years. On the encoding side, it takes LAME 3.90.3 and up, which is literally 15 years old at this point. Playback wise, foobar2000 can do it, as does Winamp since 5.2, not sure who else. In terms of DAP firmware, Rockbox definitely gets the job done, and in fact I've had some cases where it was playing gapless while Foobar wasn't.

So I haven't had any such problems with my own encodes for years, or even most files obtained from more or less dubious sources for that matter. Unfortunately, my experience with files obtained from Amazon has been all over the place - apparently they started out with plain LAME rips back in the day, which work fine, but the transcoding setups they've got going these days generally don't. It's maddening to think that a commercial supplier would be unable to tackle a problem that's been a non-issue for any average Joe ripping their own CDs for eons.

Opus went down in flames with it cutting both low and high freqs.

I'd double-check the levels then and make sure they are precisely matched for all formats. There might be some level reduction going on somewhere.

 

[bennetng]

I'd double-check the levels then and make sure they are precisely matched for all formats. There might be some level reduction going on somewhere.

Yeah, it is especially true when encoding 44k materials since opus working in 48k internally.
https://hydrogenaud.io/index.php/topic,114125.msg940028.html#msg940028
https://hydrogenaud.io/index.php/topic,113655.0.html

The excessive >1.0 float values in lossy files will either be clipped by ASIO/WASAPI exclusive API or compressed by CAudioLimiter APO when using shared mode.
https://blogs.msdn.microsoft.com/ma...audiometerinformation-and-full-scale-signals/

To be honest, ABX logs are not very meaningful to me as they can be easily faked. I like something harder to fake.
https://hydrogenaud.io/index.php/topic,111736.msg930087.html#msg930087
It's about someone claimed that he can judge sound quality of lossy encodings based on bitrates and spectrograms. People who are interested can read from page 1.

 

[JS Hoover]

It often results in sounding like a "gurgling" effect the worst on: (mostly) whenever transient, out of phase stereo signal reflections bleed into the opposite channel (as if someone was shaking a tambourine in a telephone booth underwater) or whenever the music gets "too busy" (which, in that case: it's always like there's a split-second lagging in the "speed" of the decay of the previous note/sound).

If you compare (for example) a generic YT upload of a single track to someone having uploaded an entire album containing that same track --- the uploaded version of the entire album, in my experience, will have more lossy problems to the point of being unlistenable (no matter what the physical source the original came from). However, if somebody did a good job uploading individual tracks one-by-one even still a little under CD quality: it's, actually, quite easy to tweak them "into analog-ishness"; by copying them onto vintage reel to reel tape stock as old as from the pre-"high output" blanks of the 1960s (which will give a slight "dulling effect" to any residual digital brittleness and create a more pleasing "fattening" of the sound). I like to amplify the phone through an a/v receiver to get active preamplification (to add midrange eq and any bass if needed) from headphone out into the rtr line in (via a very good -I think- home made "Y" patchcord of: 1/4" Switchcraft stereo phone plug soldered to a pair of old Monster Inter-Link RCA's with one of the ends cut off both).

 

[ernestcarl]

I remember doing my own tests a long time ago... and could perceive a difference with 256kbps files vs the original. But I could not for the life of me decide if one or the other were better. Probably some clipping in the highest frequencies and only at certain precise passages. Most of the songs sounded the same. Considering that I am not getting any younger, why bother? I am not an archivist. The music collection I personally already have will likely die with me as few of my family share my taste in music anyway. Virtual archival copies of the original masters already exist elsewhere so no loss either way. But one may have a legitimate reason for wanting the best of the best so I leave that choice to others.

*I use 256kbps AAC when compressing original audio files and 320kbps with mp3. Hard drive space is cheap. A single 30-40GB blu-ray disc is humongous in comparison to a single music album. I have yet to archive my BD collection, but am thinking... No. Too much work.

 

[Sgt. Ear Ache]

People should be left to use whatever they want, and left to decide for themselves, without interference from corporations, whether differences between 320 kbps and CD quality are "negligible". In 2018 storage space can be expanded easily and affordably on most phones, except iPhones.

That's fine. But let's avoid jumping on the forums and making audiophool claims about how such and such a cable or dac doesn't sound awesome compared to some less expensive item because "you have to use the highest possible resolution source to discern the differences." Is it within the realm of possibility for a person with a lot of expertise to discern differences? Maybe. But not under any normal, real world listening conditions. The distinctions are beyond "subtle." A better word would be "negligible."

And nobody, anywhere, ever, is discerning any difference between high bitrate lossy and lossless IN THE CAR! Come on...

 

[somebodyelse]

Gapless playback of MP3s has basically been a solved problem for many years.

Technically you're correct - the PJB100 was doing gapless MP3 with track labels and navigation 20 years ago after all, and I don't remember the Squeezebox ever having problems with it. It's not strictly part of the standard, but the solutions have been well known for a long time. Sadly most of the commercial world doesn't think it's important, and even big name DLNA devices glitch with properly encoded files. If your device plays other formats correctly but glitches on MP3 it's easy to incorrectly conclude, as @Hugo9000 has, that the problem is the format. Gapless playback isn't a requirement in the UPnP or DLNA specs, and many (most?) devices don't do it properly. IIRC it needs support from both the server and renderer ends, making it even less likely to work. This is assuming correctly encoded files, which you point out is far from a certainty.

 

[Midwest Blade]

320 MP3 to 16/44 CD is pretty close to my ears, when I compared the top level of Spotify to Tidal. I do notice a difference when comparing Pandora to Tidal.

I have been with Pandora for many years, tried Spotify for a brief period and then went to Tidal. Have now dropped Tidal and went to Amazon HD while still keeping Pandora as my wife likes to use it for general listening. I struggle to hear much difference with any hi-res files and am quite content with 16/44.

 

[ahofer]

When I have tested myself, I found I could often (more than 50%, less than perfect) tell the difference with orchestral and string quartet music, but failed miserably with a lot of pop and rock. Unlike a Stereophile writer, I’d have trouble articulating the difference. But I am extremely familiar with the live reference in the case of chamber music and strings particularly. Pop is a studio creation-there is no live reference.

 

[q3cpma]

What's the point of focusing on an obsolete codec, especially on ASR?

 

[JJB70]

What's the point of focusing on an obsolete codec, especially on ASR?

Because it still sounds fine, there are still an awful lot of MP3 files out there and although cheaper memory negates the advantage of smaller file sizes it is still useful in some cases.

 

[splattened]

Her voice had a lot of natural distortion. I can't see how this would not affect comparisons

But who can REALLY tell the difference and validate their claims. This should be an open-and-shut scientific case. Something amiss here. Is it lack of DBT trials or 'I can hear it' protests with no substantive back-up from that camp?
'Round and 'round we go ..................................... .

320kbps? A trained listener? We need a Cochrane's review on this topic re general listening. Science wise, not many think it is more than a narrow-field indulgence.

Next time you are at a bus-stop say you can smell fish. You will get some who can smell fish, too.

ASR, opinions are rife.

I realize this thread is a year old, but I'm having a hard time understanding your above posts. You seem upset that there are cases where people can successfully ABX lossless and high bit-rate lossy encodings but it's not clear why.