The thing is, the “heard” difference between 320kbps MP3 and CD quality & CD quality and higher resolution formats are not “details” per se.
The audible difference can be described as sound stage size, instrument separation and atmosphere.
The problem is making these details audible needs a good system. “Good” doesn’t mean $10K+ here. Two high quality, large-ish two way bookshelf speakers and a good amp with enough punch (50W+ Yamaha or similar will do) plus a good source in a sizable room is enough.
There’ll be people who can’t tell any difference, there’ll be people who can “feel” it, and there’ll be people who can pinpoint differences. This is because the ear training and biological limits of said people.
I have a friend who can pinpoint a half note (natural vs. sharp) mistake in a 90+ people symphony from YouTube recordings, incl. the instrument. His natural ear resolution is around 1/9th of the tone. He always tunes his instruments via ear and verifies with a tuner. So, this is not impossible.
My ears are not that absolute, but I can divide music to layers and pinpoint details, for example.
Lastly, taking a “diff” of CD quality and 320kbps MP3 version of the same track will leave an audible residue.
There are other comments I left over the years here. Search them for more info. I’m on mobile. I have no practical way to link all of them.
This is the kind of snake oil that double blind studies in controlled reference environments have established is snake oil.
Like for example, the fact there is an "audible" diff is meaningless. The threshold of hearing is not linear nor frequency independent. This is called "masking" and it's exploited by lossy codecs to allow for better encoding as well as audio watermarking. You can add a noise that would be perceptible by itself to content that is entirely masked by the content itself. And the reverse is true, you can remove content without it being perceptible.
This is the first reaction I always get: "This is snake oil, and you don't know what you're talking about". While there are undisputed snake oils in audio and audiophile market, the difference I'm talking about, is not.
The idea of lossy codecs is they filter out the things you theoretically don't hear, yes. However, the presumption that you don't hear these when they are present is not completely true. Because they have a secondary order effects in overall sound.
The audible residue you claim that I don't hear when it's in the CD is the part of the sound which adds this instrument separation and soundstage expansion. Same is for higher sampling rates. While you can't pinpoint the difference with words, it shows itself as smoothness and "richer" sound.
Saying that you can't hear that difference is akin to saying "Human eye can't see faster than 30/60/X FPS anyways", which is not true.
When anyone presented with a lossy-encoded audio file produced with a state of the art encoder and not brick-wall mastered, will be impressed, yes. This includes me, too. However whenever you listen to the same file in lossless or, if present, higher resolution formats, with a sufficiently transparent audio system a couple of times, you start to notice the differences.
There are a couple of caveats in all of this audio business. First of all, you need to know how your audio system sounds and behaves to be able to discern differences. This requires time with the same system for a long time, to understand how it responds. In my case, I have the luck of having the same amplifier (An AKAI AM-2850) for ~30 years. I know how that thing responds to any genre of music, and I know how anything should sound at any quality level. Again, as I aforementioned, you need to do these ABX tests a couple of times back to back, esp. if you don't know the track, to be able to decode the details in sufficient manner. Digitalfeed's ABX test (https://abx.digitalfeed.net) understands this and makes you listen to the same thing 5-10 times according to your available time.
See, I'm an ex-orchestra player. I played in concerts, listened master recordings, and YouTube uploads of the concerts I played as well. I have also listened tons of CDs, MP3s of the same albums, etc. Some of the albums I listen have a captivating sound when I listen to them from CDs. MP3 versions of the same albums do not nail me to my chair, yet I can't leave the CD version of the same album to get a cup of tea. Both are ran through a Yamaha CD-S300 CD player with an iPod interface and MP3 playing capability over USB.
I can also write how CD tracking quality affects audio clarity, but this comment is long enough. In short, Yamaha's old CD-Recorder, CRW-F1 really improved sound quality by abusing Red Book standard by lengthening the pits of audio CDs. It reduced to capacity to 68 minutes, but it was worth it, esp. on lower end CD players.
> I can also write how CD tracking quality affects audio clarity, but this comment is long enough. In short, Yamaha's old CD-Recorder, CRW-F1 really improved sound quality by abusing Red Book standard by lengthening the pits of audio CDs. It reduced to capacity to 68 minutes, but it was worth it, esp. on lower end CD players.
Sorry this one part especially makes no sense. Digital is digital. Either it added more samples per second, or more bits per sample, or it's snake oil. There's a stream of bits that comes out of the reader. There's no residual information about the length of the pits.
EDIT, yeah sorry this is completely and utterly impossible that you are getting better "audio clarity":
"Yamaha tries to attract computer enabled audiophiles with the Audio Master technology. Audio Master promises reduced jitter and decreased error rates for audio recordings via extended pit and gap sizes on the CD-R. This is actually quite simply achieved by increasing the disc rotation speed vs. the laser clock frequency. In other words, Audio Master recording at 8x rotates the disc at 8.2x, thus creating the extended pit & gap lengths. This naturally reduces the capacity of the disc."
Literally they are just spinning the disc faster, reducing capacity and make it slightly less likely that errors will be read. If you're getting read errors on playback, that means your disc is dirty or your CD player sucks. It's the same bitstream, just read at a different linear rate.
If you honestly believe that this is an audiophile concern, I'd urge you to reevaluate a lot of your other beliefs, because they are clearly not all grounded in technical facts.
The problem is not the values delivered to the DAC. It's the aperture error of when the DAC was clocked. This affects what those sample values actually represent. If the clock is being extracted from the pattern of pits and there's a way to reduce the jitter, you will get a more accurate signal.
However, I'd hope that anyone that cares about fidelity has a CD player that does a little more to generate a DAC clock. NCO run by a software PLL or a hardware PLL with a good loop filter are things I've heard of, but control systems is not my specialty.
CRW-F1 didn't encode more information into the pits. It allowed lower end DACs to have more time to switch properly by giving them a slower signal stream within acceptable limits.
DAC's digital part is easy. What differs in quality is the analog part. If DACs were that simple, a 25 cent DAC would power every unit from bottom bin to top tier.
Before that Yamaha CD player I had, I used a lower end Sony CD-Player (I don't remember the model, sorry). Writing the same album, to same brand of CD-R, with the same speed in two different modes created two audibly different disks.
I sometimes challenged myself by writing in both modes, not marking the CD-Rs, and the audio difference was always audible. Even after weeks. 68 minute CDs were always had larger sound stages with more clarity and instrument separation. This is again on the same AKAI AM-2850 amplifier.
I guess this difference would be impossible to hear today, because higher end units have better tracking and better DACs. Also some of them use DAE and use multi-second buffers, so the "slower stream" is no longer present in the pipeline due to buffering.
A CD signal stream fed to a DAC is a 44.1 kHz 16-bit signal, period. All you did was force the drive to spin a bit faster to keep tracking, or let it fill its buffers more slowly if it spun at the same speed. The buffers, after error recovery, are what feed the DAC. Assuming an error-free read on two discs burned with the same data (regardless of "pit length", disc material, etc), you get the same bits in the buffers.
There's no "slower bitstream" for the DAC. That's provably nonsense and you can work it out from basic principals. The same bits would come out of the optical interface of a CD player, at the same rate either way. If the CD player has a built-in DAC, the same bits would get fed to that same DAC either way.
I'm sorry, but if this is truly what you believe, it really puts everything else that you said into question.
To give you the benefit of the doubt, I might say that the lenses or lasers on your CD players are filthy, and you're just hearing skips or noise from poor reads and that a slower-written, borderline-spec disk might just allow them the function better. Perhaps your player was interpolating or concealing frames [1] that it couldn't read correctly and failed to correct via ECC and you were just hearing a poorly reconstructed digital data stream.
This sort of confident incorrectness, ignoring the underlying technical architecture, is probably why people don't believe anything that an audiophile says.
Sure, I can understand how slightly changing the layout of the pits and lands can potentially reduce the error rate. But if I'm already getting playback with 0 C2 errors, there is absolutely no change in quality from the original signal.
Advanced Audio Master visibly reduces C1 errors in recorded media, to almost negligible levels. While C1 error can be corrected without any degradation theoretically, its result is up to the CD player's capabilities (and quality).
CDRInfo's tests back in the day showed dramatic improvements in C1 levels, see [0]. Considering some of the lower end CD players by leading manufacturers didn't even had 16bit audio decoding and used late stage oversampling, reducing C1 errors was/is a big deal in recorded media.
As I said in my earlier comment, this mode lead to clearly audible improvements in my older, low-end Sony CD player. I don't how how will it fare in my new Yamaha player due to technology improvements.
Sure, the C1 rates are lower here for one kind of media in one actual comparison, but C1 errors below a certain rate are entirely correctable errors. Meaning, you're not going to lose data when it gets reconstructed. You are not losing fidelity of the CD by having a lower C1 rate in the 20s to 40s; every bit decoded is being played back exactly as originally encoded. Otherwise reading data off a CD would always have significant amounts of corruption every time you installed software. Every time you opened a photo off a photo CD you'd get all kinds of JPEG corruption.
Assuming you have a properly functioning CD player, these errors have zero input in the quality of a CD being played. If you have a noticeable difference in audio quality between a CD with a max error rate of 24 or 32 C1 errors, you've got a tremendously faulty CD player that is complete trash.
Its funny too because in this table it shows the Mitsubishi media performed about identical or better in every speed while being significantly faster at recording. The 1x speed with AudioMaster on for Plasmon is the worst result, ignoring the time they burned 16x media at 44x speed. This table is also challenging to actually compare, because they show different speeds for the different modes (1,4,8x for AM, 4,16,44x for regular) so the only one we can really compare fairly is the 4x. Even then, at 16x speed AM off it had a lower average error rate than 8x speed with it on!
Don't get me wrong, burning a lower error rate from the get go is good, it implies the burn will possibly be more reliable over time as you get things like scratches and other imperfections on the disc. But arguing that a disc with an average of 2.1 C1s vs 1.1 is going to be noticeably different in the sound is absurd. And once again, even then this showed an improvement only in one of the two medias tested. Maybe its better in more media, maybe its worse, maybe there was just something odd with their burns and this is largely just noise in the overall results of burn results from this drive.
I'm still not convinced AM actually did anything but reduce your recording time. Your link doesn't say anything of actually increasing the audio fidelity in the slightest, just that in one straight comparison the C1 errors were lower. Practically every result in that table other than the 16x media being written at 44x speeds is already massively in the "negligible levels." Being below 220 is "negligible", and all of these burns (aside from the one using AM at 1x speed!) are well below that.
I have much of what you describe: 3 listening environments between near field computer speaker setup, reasonably high end home stereo and dac/amp with high end iems.
I have tried multiple times to discern flac vs 320 mp3 across genres. Every time I believe I can figure it out and I consistently fail to exceed 50% (pure chance) accuracy.
Makes me wonder what ultra-linear source gear or speakers would highlight the differences in real-world situations, if at all. But for my purposes I’ll happily accept the roughly 80% file size reduction for no audible difference.
I could hear the difference between 320 Kbps MP3 and lossless WAV in blind tests in my early 20s - though I needed headphones to do it, and it had to be a track I knew well, and I couldn't tell which was which. They were just different, and neither felt worse. This quickly converted me to using CBR 320 Kbps MP3 for everything, on the basis that it would be just as good as the original.
A couple of years later I found an ambient track that sounded messy at 320 Kbps, and that converted me to using flac instead. Disk space had got cheaper enough over that period that it made no meaningful difference.
THis was all years ago now (I'm in my 40s...), and I don't worry about it too much any more. Firstly, I still use flac, so it's identical to the original anyway. And secondly, even if I did use mp3, my aged ears probably couldn't tell the difference even if I turned the volume up to unreasonable levels.
I think you need to spend more time with the systems and the music you have. Because, at least for me, understanding the differences at the first shot is very unlikely.
Brain is interested in the low hanging fruit, i.e. the music and the melody itself, first. The music needs to became mundane or ordinary to be able to listen it deeper for more details. This is when differences can be heard more easily.
Lastly, you don't need perfect systems to hear differences, but understand how your systems respond to the music you're listening to. i.e., your music system's sound needs to be mundane to your brain too to be able to go from low hanging fruit to minute differences you were not able to hear before.
The audible difference can be described as sound stage size, instrument separation and atmosphere.
The problem is making these details audible needs a good system. “Good” doesn’t mean $10K+ here. Two high quality, large-ish two way bookshelf speakers and a good amp with enough punch (50W+ Yamaha or similar will do) plus a good source in a sizable room is enough.
There’ll be people who can’t tell any difference, there’ll be people who can “feel” it, and there’ll be people who can pinpoint differences. This is because the ear training and biological limits of said people.
I have a friend who can pinpoint a half note (natural vs. sharp) mistake in a 90+ people symphony from YouTube recordings, incl. the instrument. His natural ear resolution is around 1/9th of the tone. He always tunes his instruments via ear and verifies with a tuner. So, this is not impossible.
My ears are not that absolute, but I can divide music to layers and pinpoint details, for example.
Lastly, taking a “diff” of CD quality and 320kbps MP3 version of the same track will leave an audible residue.
There are other comments I left over the years here. Search them for more info. I’m on mobile. I have no practical way to link all of them.