It’s still my preferred format. Everything can play it. At 256kbit or better it sounds fine for usual listening.
Even at 160kbps, maybe 1/1.000.000 people can recognice a FLAC vs MP3 trying 10 times (continuous) using expensive headphones and players, 320kbps is overkill, I prefer a FLAC and just encode to Opus.
Right now Opus is better and can be played in web browsers, smartphones, YouTube and Netflix are using that for awhile.
I 100% do. I think mp3 is a good compromise of sound and space. It’s also the format I’m used to. Just like how people swear by physical record. If I’m at a get together and hear mp3 quality, I’m at home.
That being said, I have my absolute favorites in flac for my iPod 5th gen video I rebuilt. The 5th gen’s dac, Wolfson, is a solid little dac for the day and age. Got Rockbox loaded up and I’m ace, but I’ve hard saved all the Apple firmware for every model in case the time came to sell them. Old iPods could be an investment someday and I own every gen in multiples.
Still care about MP3- it’s the bog standard, the thing EVERYthing supports. Like the shitty SBC codec on Bluetooth. I’ve still got tons of MP3s and they aren’t going away anytime soon.
Everything I get new though is high-res FLAC.
Opus is better than MP3 in every way. File size is either better or the same, and audio is better even at lower bitrates. But realistically, most streaming services don’t provide HD audio, so it really doesn’t even matter.
249 webm audio only 2 │ 1.58MiB 49k https │ audio only opus 49k 48k low, webm_dash 250 webm audio only 2 │ 2.09MiB 65k https │ audio only opus 65k 48k low, webm_dash 251 webm audio only 2 │ 4.14MiB 128k https │ audio only opus 128k 48k medium, webm_dash 233 mp4 audio only │ m3u8 │ audio only unknown Default 234 mp4 audio only │ m3u8 │ audio only unknown Default 140 m4a audio only 2 │ 4.20MiB 130k https │ audio only mp4a.40.2 130k 44k medium, m4a_dashThis is YouTube music, which generally serves the split audio from a YouTube video as a song. Most of them I checked either don’t have audio above 130Kbps or don’t even provide MP3/Opus anyways.
Youtube Music doesn’t just serve the audio from a video. They do serve the audio from a video if nothing else is available, but they also get releases directly from the publishers/distributors.
The difference in sound quality is definetly noticeable.
Youtube Music doesn’t just serve the audio from a video.
Yes it does. You don’t even need to take my word for it. Look up any song by any artist and find their official video for that song. Take this one as an example: https://youtu.be/kPa7bsKwL-c
Analyze it with yt-dlp or something similar;
249 webm audio only 2 │ 1.51MiB 50k https │ audio only opus 50k 48k low, webm_dash 250 webm audio only 2 │ 2.00MiB 67k https │ audio only opus 67k 48k low, webm_dash 251 webm audio only 2 │ 3.92MiB 130k https │ audio only opus 130k 48k medium, webm_dash 233 mp4 audio only │ m3u8 │ audio only unknown Default 234 mp4 audio only │ m3u8 │ audio only unknown Default 140 m4a audio only 2 │ 3.90MiB 129k https │ audio only mp4a.40.2 129k 44k medium, m4a_dashYouTube already has access to the audio for that song without any additional effort because of how YouTube works. I’m sure publishers can provide higher quality audio, up to 256Kbps but that option isn’t even enabled for users by default. By default you’re listening to “normal” audio or 130Kbps: https://i.xno.dev/Ow2eC.png
The reason why YouTube Music works is because they already have access to a huge library of music through music videos and the like. They save a ton of time and money by doing things this way and it makes perfect sense that they do…
As I said, some of the music is just the audio of a video, but they also get a lot of releases directly from the publishers. They are both on YT Music and the difference in quality in between them is noticeable.
I have my audio quality set to high in that options menu btw.
It’s less supported, and for me mp3 is largely enough. Can fit a lot of them on my 20€ 128GB usb key…
Podcasts are almost exclusively mp3. There is no need for lossless fidelity on those. And when you are subscribed to 200 podcasts like I am a small file size matters. And when listening at 2.5x speed lossless is a complete waste.
All my podcasts appear to use the AAC spoken audio profile? It’s much smaller and cleaner than MPEG layer 3 audio.
Apple broke metadata compatibility with a recent update. The podcast producer I know with an explicit AAC feed decided to just redirect to the MP3 feed. Unrelated to that, they also increased the MP3 bitrate for better audio quality. The increased file size doesn’t really matter that much compared to 15 years ago and people without unlimited data can just set their automated syncs to WiFi only.
Well, most of my music collection lies as mp3. I care about metadata and all of them have tags. I would love to convert my collection to opus but first I need FLACs and an easy way to move over metadata, since vorbis is different than ID3tag. Do you know a streamlined way for this?
For FLAC you have torrents, no legal way to have that. For tags I use https://beets.io/ but it’s not moving tags, it’s detection and looking up on a database on the internet.
For Flac you have digital market places and CDs you can obtain from store fronts and private sellers like flea markets or shops like ebay or discogs.
Or torrents and DDL.
I actually used discogs a lot in the past. They can be quite expensive at times. Though this will be a mix of everything since not everything can be obtainable legally, at least for my archive.
If I juat want one song and flac isnt expensive to buy digitally I’ll buy it.
But if they want somethibg like 3€ per song I’d bail and pirate it.
Discogs is only if I really want it the CD and it’s out of sale. Else it’s usually less expensive to buy it from the official store.But if I had to choose between discogs and ebay, I’d prefer discogs due to more information about the release and condition.
It’s useful because it’s ubiquitous. Everything that can take in music files supports it.
Is MP3-encoded audio of the best possible quality? No, of course not. But for most people it’s Good Enough, especially if you do most of your listening in a noisy environment. MP3s are to lossless formats what CD was to vinyl for so many years.
I’d argue you’ve got that backwards; CD is to vinyl what lossless is to .mp3. That said, I know what you mean.
A lot of people cant tell the difference between MP3 @320Kbps and a fully lossless FLAC.
A lot of people cant tell the difference between MP3 @320Kbps and a fully lossless FLAC.
MP3 has some disadvantages over more modern formats, regardless the used bitrate. It’s been a long while since I was very interested in audio formats, so I may not be up to date on some newer developments but unless anything major changed, MP3 can’t do truly gapless playback between tracks (used in live albums), for example.
Aren’t there unofficial extensions to mp3 for gappless playback? IIRC you can tag tracks as gappless and many audio players will make them so.
Aren’t there unofficial extensions to mp3 for gappless playback?
Yes and no.
IIRC an MP3 track is divided in fixed-length frames and unless the actual audio matches perfectly with the end of a frame, it’s not possible and that’s why cross-fading plugins for audio players were invented. The padding data is there either way but can be documented in the metadata section of a file.
Last I checked (and that was years ago, so I may be wrong) this approach was never perfect and prone to breaking. It’s an inherent flaw with the format where some form of workaround exists.
That said, for most use cases this is irrelevant.
Audio playback is such a low-demand process, surely a player (e.g.VLC) can spare a thread to line up playback of track 2, a few seconds before track 1 ends? It knows the exact length of the track, why can’t track 2 be initiated when the audio level in track 1 drops to zero (or minus infinity dB) in the last frame?
Workarounds in a specific player don’t negate the fact that the format has limitations.
All people. 320kbps mp3 is completely audibly transparent under all normal listening conditions. It’s a low-tier audiophile meme to claim otherwise but they will never pass a double-blind test.
From what I understand, vinyl and CDs can both output in a range greater than human ears can detect, so the medium isn’t as important as the mastering and the gear being used to listen to the recording.
CDs can, by a very narrow margin, reproduce sounds beyond which the human ear can detect. There’s a theorem that states you can perfectly reproduce a waveform by sampling if the bitrate is double the maximum frequency or something like that, and CDs use a bitrate such that it can produce just above the human hearing range. You can’t record an ultrasonic dog whistle on a CD, it won’t work.
It’s functionally impossible to improve on “red book” CD Digital Audio quality because it can perfectly replicate any waveform that has been band-passed filtered to 20,000 Hz or thereabouts. Maybe you can talk about dynamic range or multi-channel (CDs are exactly stereo. No mono, no 5.1 surround…Stereo.) It’s why there really hasn’t been a new disc format; no one needs one. It was as good as the human ear can do in the early 80’s and still is.
The Nyquist limit?
You need sampling at twice the frequency as a minimum to extract a time domain signal into the frequency domain. It says nothing about “perfect” especially when you’re listening in the time domain.
There is a lot of data in the time domain that impacts sound/signal quality. As others have said though, it probably doesn’t matter without high quality equipment and a good ear.
It’s also good to note that you can train your hearing. A musician or producer or audiophile are going to hear things and qualities you don’t. It’s edge cases though, and generally irrelevant to regular listening.
You definitely can hear the difference between MP3 320 and lower mp3 bitrates though.
You need sampling at twice the frequency as a minimum to extract a time domain signal into the frequency domain. It says nothing about “perfect” especially when you’re listening in the time domain.
Yes it does. You can use a higher frequency, but that does not change anything except increase the maxiumum frequency possible. Even with perfect ears and the best equipment, there is no audible (and mathematical) difference to be had.
Everyone who claims otherwise should watch Monty’s explainer videos. I know they are quite old at this point, but everything he explains is still perfectly valid. If that does not convince you, nothing will.
It is my admittedly limited understanding that we really can’t do better at digitally recording an audio signal than how red book audio does it, such that the microphones, amplifiers, ADCs etc on the recording end and the DAC, amp and speakers on the playback end are going to be much more significant factors in audio quality.
The original idea behind the superiority of vinyl was that the ambient audio was being recorded directly to the media. Of course, this wasn’t even true when it was first made, as they were using magnetic tape by then to record in analog. However, there is still some merit to the idea that an infinitesimal amount of quality is lost when translating sound waves to digital data.
Most of the actual differences between cd and vinyl, though, can be chalked up to the loudness wars ruining the mixes on cd.
Vinyl is lossy in that any dust or scratches on the record can be heard in the output, so this is only true if you’ve got an absolutely pristine vinyl.
This is what we were all told for years and years- that it was impossible that anyone could hear anything in vinyl that was supposed to be there but that couldn’t be reproduced with digital at cd quality. Then DVD came out And people could genuinely hear the difference from CD quality audio even in stereo. It turns out that dynamic range is limited by the audio sampling rate and the human ear can easily detect a far greater range CD audio supports.
I am very slightly annoyed that people haven’t moved onto Opus which gives you better compression and quality than MP3. MP3s are still useful for any older devices that have hardware decoding like radio sets, handheld players, etc. Otherwise, every modern device should support Opus out of box.
Hilariously, x264 has the same problem where there are direct upgrades with H.265 and AV1, but the usage is still low due to lack of hardware accelerated encoding (especially AV1), but like everyone uses FLAC for the audio which is lossless lol.
I think SW Republic Commando sounds were stored in Vorbis. Back then.
I use Opus when I rip something. It’s been a long time since the last case. I’ve left FreeBSD for Linux and returned back to
LinuxFreeBSD again since then.I think SW Republic Commando sounds were stored in Vorbis. Back then.
Unreal Tournament also used Vorbis starting from either 2003 or 2004.
Yep with the music compressed to hell, still sad about it after all this years
But think about the 5 MB they saved!
I just use ogg vorbis and vp9 in webm container, also webp for images. No proprietary nonsense in this house. AV1 sucks on my hardware, but yes eventually.
I use it to (re)compress audiobooks, podcasts and such, they still sound very good at 32 kbps.
Fun fact, Opus has been supported by a hobby OS like MorphOS for years, my ancient hardware doesn’t break a sweat playing it.
I have boatloads of MP3s and at least they can pretty much be played by all imaginable software and hardware imaginable, and since the patents have expired, there’s no reason not to support the format.
MP3s are good enough for its particular use case. Of course, newer formats are better overall and may be better suited for some applications. (Me, I’ve been an Ogg Vorbis fan for ages now. Haven’t ripped a CD in a while but should probably check out this newfangled Opus thing when I do.)
I have thousands of mp3s so I’d say they still matter. As far as audio quality goes I doubt my ears, at least at my age, can tell the difference between them and a lossless format.
Anyone telling you they can hear the difference between a 320kbps MP3 and lossless audio is full of shit, anyway. It’s still a great format for keeping file sizes small, though I prefer ogg these days.
I’m in the same boat: can’t hear any difference.
But, I have GBs of 320k MP3s… is it worth converting to Ogg ?
I’m a big fan ogg opus, but I wouldn’t convert between lossy formats
I listen to mp3 all the time. Back in the Napster days I collected a ton of music, but moreover I’m a fan of Old Time Radio from the 30s and 40s, so I accumulated around 10,000 of those shows. More than I’ll ever have time to listen to. Audiophiles may deride the quality level, but I don’t believe in letting perfection be the enemy of good. And even if “computers” - whatever that even means anymore lol - drop support for mp3, there will always be software that plays it as long as there are people with big collections of files they don’t want to take the trouble to convert to something else.
I freaking love old time radio, that stuff is great!
That sounds fascinating. If I were interested in those shows, where would I start? Are there at least some that are easily listenable to on the open internet?
Check out the many OTR Gold podcasts that have the serialized shows as episodes.
yt-dlp uses m4a but sometimes I like my library to be mo3 just for nostalgia :(
… I’m out of the loop. Why don’t people care about mp3s?
Its mostly been superseded by AAC, Opus and FLAC.
Mhmm I haven’t heard of the first two. I still listen to mp3s that I got from the 90s.
There might be things that are better these days in the technical sense. But there is always value in having something “good enough” that is freely available and compatible with nearly everything that has speakers to use to keep those technically better yet more expensive options in check.
Yeah my car plays the 11,000 MP3s from a SDcard inside the armrest compartment.
Found the fellow Volvo user? :)
2013 Ford, but I know they shared some technology for a while.
Afaik Ford Focus == Volvo V40/V50 in those years, basically with a different chassis and insignia :)
And Mazda 3. The platforms are the same but engines and interiors a lot different between the Fords and the Mazdas at least.
the randomizer in my car sucks so it’s the same 100 of those 11,000 songs. :-/
So not a much different experience than Spotify
At least the shuffle is partly coherent.
Yeah WTF is up with that? My car does the same thing with a USB drive full of songs. It will literally play the same “shuffled” sequence over and over every time you drive. I have to take out the drive and change the files on it sometimes to make it actually Shuffle the songs’ order and that’s too much BS
I read the manual for my cars radio. It has a max file size limit of like 256 songs or so per folder. But it can also accept 256 folders.
So if your cars is anything like mine you can probably play your songs just by splitting them up into more folders.
no such luck for me there. the music is in /artist/album directories. I had considered flattening it all out to see if that makes a difference.
The randomizing n my Focus ST is good, but when I tell it to shuffle play it always starts with 1 of 2 different songs, every time.
