Explaining Audio File Formats
ฝัง
- เผยแพร่เมื่อ 18 พ.ค. 2024
- Digital audio file formats tutorial, including sample rate, bit depth, codecs, containers, MP3, AAC, WAV, AIFF, BWF, DSD, Monkey’s Audio, WavPack, AC-3, EC-3, AC-4, Ogg Vorbis, Opus and WMA.
If you find this video useful, you may appreciate these other ExplainingComputers episodes:
Explaining Digital Video: Formats, Codecs & Containers:
• Explaining Digital Vid...
Explaining Image File Formats:
• Explaining Image File ...
Explaining File Compression Formats:
• Explaining File Compre...
More videos on computing and related topics can be found at:
/ @explainingcomputers
You may also like my ExplainingTheFuture channel at: / @explainingthefuture
Chapters:
00:53 Digital Audio (sample rate & bit depth)
03:16 Codecs & Containers
05:03 Non-compressed audio formats
07:44 Lossless compression formats
10:24 Lossy Compression formats
13:28 Wrap
#AudioFormats #MP3 #WAV #AIFF #AAC #ExplainingComputers - วิทยาศาสตร์และเทคโนโลยี
The sound coming from my PC speakers was exceptionally bad as I started watching this video and I thought "ah, he's comparing and contrasting lossy and lossless formats, very smart!" In the end it turned out to be a problem with my headphone jack. Great video!
Years ago I compressed my CD collection to 320 kbps mp3s using CDEX to fill micro SD memory cards with hundreds of songs each. I can rarely hear the difference except in the softest passages, though the bit rate was reduced by more than 2/3, and the file sizes shrank by as much as 10x. I thought I knew it all about audio codecs. I’m unceasingly amazed at what I continue to learn from watching EC videos. Thank you!
I think a lot of that is that the common players don't really produce a different output, so their incarnation of 320k and CD sound is functionally identical as it arrives at the audio jack.
I can't listen to classical on CD, because it loses all the ambient sound that makes an orchestra "sound right" (this is preserved on tape or vinyl, but not on CD which is already a very lossy format). But most people don't notice this.
CDEX uses LAME as encoder, the best of the 3. I don't like Fraunhofer on 320 kbps. And Helix is the quicker variant.
There is a lot more to explain about using MP3 the right way. (CBR, joint stereo etc)
In the past i had a encoder that also makes 500 kbps MP3 that only plays on an few programs. Useless.
Old players, car radio's gives hiccups when they play 320 MP3
SACD makes a lot of difference, it's a real problem to buy these.
Audio BD is also a improvement to a CD when it's done properly.
..what happened to your CD collection? 😞
I hope you did not "throw away" the high quality cds!
@@dallesamllhals9161 I still have the CD collection. I also made digital clone images of every CD using CloneCd. Those copies are backed up across 11 different servers and NAS devices as well as several SLC USB memory devices.
Chris hitting us with another WAV of knowledge
:)
XD
Long live the FLAC
hear hear
..the skull n' bones "Flag"? 😛
Forever FLAC
TTA and TAK are fine as well!
WAV is better since its compatible with older computers and takes less processing power to play
I love how Apple always comes in like _Yeah, we're doing our own thing that'll be slightly different._
So true . . .
Sony did a similar thing with, I believe ATRAC (and not forgetting "memory sticks"), and that is why I will never purchase another Sony product or ever buy an Apple item. This despite both companies making some of the best hardware available
@@skf957 Circa 1998 I got an MiniDisc player for my car, and a Sony component MiniDisc recorder. I used the format for mix discs. Anyway, the issue I had.... ATRAC back then had - for me - a lack of lower frequency and noticeably less bass.
@@skf957 Apple, I can understand, but Sony? Even now? They aren't on a non-stop train to purposefully make things difficult like Apple. Every company tries something stupid then gets back on track. It all depends on if the customer allows it. Apple customers allow Apple to do whatever they want and that's why they are the way they are. They can do no wrong. But other company customers will actually speak up.
...and then they have the gall to call it "interchange format"
The interesting thing about 44.1kHz sample rate and 16 bit depth, which are still so often used today, is that they were set at the start of CD. In turn, they were chosen in part because of the bandwidth available on an old video cassette format called U-matic. The PCM digital audio data for early CD was delivered on U-matic video tapes recorded usually with the PCM1610/PCM1630 digital encoders. A long forgotten set of equipment whose specifications are still reflected in the data rates of today's digital audio formats. I have two PCM1630 and U-matic rigs set up here and use them to recover digital audio tapes which recording studios find in their archives, using a pure lossless digital route. Tapes which are 40 years old can still sound stunningly good today, though sometimes they do suffer from data errors.
That's Sony's side of the digital audio, and subsequently CD project, where they partnered with Philips, who where forerunning the optical disc side.
U-matic was Sony format and they also produced 16-bit digital decoders. Meanwhile Philips had been working with 14-bit digital audio they were planning for the CD as well, but Sony vetoed it.
That's why old Philips CD players proudly proclaim 4x oversampling, as they only had 14-bit DACs and had to accommodate them for 16-bit CD audio.
Although there are a variety of formats, what's surprising is how many of them have lasted for decades, with no sign of disappearing. Thanks Chris, good one.
Indeed! Some older formats do tend to hang around!
I remember aiff from my Amiga.
A neat thing with smartphones as mediaplayers is that format compatibility can remain pretty reliably wide.
The .SID format is of course superior for having a lot of music on the go.
Always go lossless if you're backing up/archiving your cd collection. That way if you ever want lossy versions to save on space/bandwidth you're basically future proofed to make lossy copies with whatever the best format is at the time.
I ripped all my CDs to flac for that very reason. A friend of mine spent ages using windows media player to rip his collection only to find when his Windows went tits up he was left with thousands of unplayable wma files that were drm protected.
Great advice, but I think most people don't really care about having a lossy format. I always rip and convert my music to MP3, even if I buy an album in FLAC, because I know it will play on every device. I've had too many experiences trying to play a FLAC file on a stereo that doesn't support lossless files. I eventually just gave up.
@@ReedsRedactions My old car is so noisy the audio quality doesn't really matter when playing music while driving. So the file format I use is _loud!_
@@ReedsRedactions How can you have problems playing flac lol
I like to advocate for a bigger use of Ogg Vorbis due to one curious fact: for some reason MP3 has a little gap of silence when looping, while Ogg Vorbis does not - this is why you see so many game devs use Ogg Vorbis for music instead of MP3 (and WAV for sound effects).
Also because despite the fact MP3 patents have expired long ago, Ogg Vorbis remains a free and open standard. Might be a moot point for some, but I stand firmly with it.
Serious music players support the delay and padding metadata in the Xing frame, which describe the duration of extra silence and it is then discarded. All lossy codecs have this property. A game engine is a closed system and they could easily either follow the standard or save this data in another format to allow looping.
VLC is the most serious media player around, but still produce a silent gap between each track. I read it was about to be fixed. Il really hope so.
Afaik Spotify uses ogg
Ogg Vorbis produces better results than mp3 at same bitrate, or smaller files at same quality. Long live Ogg.
Today most OGGs (container) use Opus (codec) instead of Vorbis (codec). It is simply the best lossy codec as of today in any aspect.
I only rip in FLAC these days. Then I take it to my Android phone and use PowerAmp's various features to get everything out.
Great informative video; keep it up.
In Poweramp we trust, probably the most worthwhile purchase I've made on the Google Play Store
While the content is obviously a grown up level of complexity, these videos give me a strong vibe of sitting cross legged on a primary school carpet watching a video tape on a crt tv on a trolley. It feels nostalgic and I love it
As a news publisher, I receive for publication audio, video and text on a daily basis. Very often I receive what are described as MP4 files, but are in reality M4A. This is due to a lack of knowledge on the part of the reporters who don't understand the difference, believing that an MP4 is a 'better' MP3.
It is one more mp better, isn't it? Haha
@@bertblankenstein3738 Exactly what Nigel Tufnel would say. And he probably did at some time or another
but it is better. mp4/m4a implies the AAC codec which is lightyears better than crappy old MP3.
@@martontichi8611 My point was, is that I get audio identified with an extension of mp4, but it should be m4a. The header data sez M4A bla blah. It affects so many sources that it's hard to believe industry wide stupidity. We *are* dealing with non-english speaking _news_ people.
An M4A is MP4. Apple decided to create a new file extension to indicate that the file has no video in it. You can rename it as your software requires.
I've grown up listening to rather horrible audio quality, records and cassette tapes and I gotta tell you hearing digital music streaming in my car is a total joy. And I have no idea what the specs of the audio are on Spotify, it just sounds wonderful. I really enjoyed this lesson Chris. Fascinating, thank you.
Spotify uses OGG vorbis AFAIK, and OGG eclipses MP3 when at the same bitrate.
If you're going to listen to compressed audio, you could do a lot worse
If you haven't heard the sound quality of CDs, you are missing out
Try tidal or apple hi-res, surely a difference but don’t play it over Bluetooth:(
Spotify is really low bitrate. CD (a FORTY year old technology at this point) blows it out of the water.
Spotify has been saying for years now that they will let people use a reasonable modern bitrate but don't hold your breath.
Tidal is identical to Spotify but with lossless audio, it's like digging mud out of your ears that you didn't know was there.
I Worked in an audio facility for film dubbing and knew pretty much everything here but I watched all the video no matter what. A little brush up can’t hurt especially coming from a very well respected source. Thanks!
Thanks for watching. :)
Being a DJ and tech enthusiast, this video is RIGHT UP my alley. Thanks for all the work chris. Its appreciated.
For long term storage lossless is the only way to go. Even if the lossless format becomes obsolete in the future you can always convert it to a newer lossless format preserving 100% of the audio quality, same can't be said of lossy files since each conversion from lossy to lossy degrades the audio quality. So always have a lossless backup of your music.
something important about opus is that it can have the same quality of mp3 at half the bit rate. for example, i use 64kb opus files on my phone, and they have the same quality as 128kb mp3 files. This brought my music library from 18gb to 9.
My lib is like 90gb😅
I actually do the opposite, simply because so many devices and programs can't read, sort or display cover art for opus while storage is dirt cheap. But yeah, if you use a HQ converter the resulting mp3 has around double the bitrate.
Maybe under half. Opus > Vorbis > MP3.
I would still - always - archive lossless. The days of me hand-wringing over storage and bandwidth are 20 years ago.
@@hansdampf2084 Same. Or so I thought. It's actually 461GB and I don't even use it; I listen to Spotify. So 461GB was from 5-10 years ago.
In the early days, I joined all my mp3s from each album in my collection into a non-compressed zip file, changed the ending to .mp3 and used that on my car's mp3 CD player so I didn't have to go through all the songs of each album to get to the album I wanted. Worked like a charm!
The fascinating part is, that is so wrong. It's the kind of wild swing that should never work like you thought it would. And I'm guessing you had no idea WHY that worked, or maybe even any reason to believe it would. But I would imagine it DOES work, and should be totally fine with most players.
MP3 is not much of a container format. It's actually just a stream of mostly independent frames. The only thing at all special about it is that several prolific pieces of software added custom headers and trailers to it, in order to support things like titles, album art, sample-accurate lengths, and indexes for easier seeking.
This is possible because an MPEG player is just searching for a frame header, which is a sequence called a sync pattern. If it reads a byte but doesn't see the sync pattern, it just moves on to the next byte and looks there, ad infinitum, until it finds one. Ergo, you can drop random data before, after, or between frames and it's still perfectly valid, as far as the MPEG reader is concerned. (At least until sync is lost to the point where the reader assumes the stream is hopelessly corrupted and gives up.) If the reader comes across metadata that it happens to know and understand, it can intelligently handle it -- parse it and benefit from it, or simply skip over it, if it can determine its length. That's why players that didn't support ID3 tags didn't mind them being there. They just ignored them, or supported them enough to know how long they were, and jump past them.
When you add an album's worth of MP3s into a ZIP file with no compression, you're basically just prepending a ZIP header and some table of contents garbage in front of the first MP3 file, which might have its own ID3 tags or whatever, and then you roll right into the data that was in the next MP3 file, and the next one, and so on. As long as the ZIP file is _uncompressed,_ the end result is not much different than if you had just squished the files together. Maybe there's a marker or CRC in between them, I dunno, I'm not familiar with the on-disk layout of a ZIP file. But it wouldn't matter. The reader would just see that as "a bad frame" and run through it until it found the next sync pattern.
Another way to accomplish the same thing is to use the command prompt with the copy command. E.g.: "copy /b track1.mp3+track2.mp3+track3.mp3 output.mp3" You would end up with the same result, minus the essentially useless, albeit benign, ZIP file overhead.
Needless to say, this would NOT work with WAV or MP4 files. (But some other MPEG formats, like VOB, are also fine being split and concatenated on frame/packet boundaries.)
@@nickwallette6201 Exactly right! Only I was shocked and stunned that it actually worked, I never thought it would. Though I did make sure they all had the same bitrates as that would have led to interesting results otherwise, probably.
Nowadays I'd use the command _cat ./*.mp3 > albumname.mp3._
@@dittikke Probably would've been fine with different bitrates too! Each frame has its format encoded in the header, and using different bitrates on a frame-by-frame basis is essentially how Variable Bit Rate (VBR) works. Not ALL players support VBR, so it's not a guarantee, but it's usually OK.
@@nickwallette6201 This was long before VBR was widely supported, early 2000s if I remember right. I did a weird transition from cassette tape to mp3-CD. Since all my "music" was recorded off the radio in mono, I also did the split-channel millisecond (nanosecond? I forget) shift thing to make it sound like stereo. So no, the quality wasn't exactly premium either. But I had my entire music collection on one CD, that was so wild.
This video is a great explanation of how all these file formats work and what they are. I've been working with many of them for years and still learned a couple of new things. Thanks, Chris!
This is one of those tutorials that I am saving to re-listen to, as my BHPS )(Brain Head Space Storage) Needs some compression to find contiguous area to absorb and process!
Greeting Leslie. I hope that all is well with you.
As usual, a very clear and understandable discussion of technical specifications. Keep up the good work!
The spirit of the 1990's ... when they chased 16-year olds for sharing their mp3 files with Napster. It was such a classic age.
Good old limewire.
That was the early 2000s, not the 1990s. And they weren't "chasing" anyone as P2P sharing of music wasn't illegal until later.
16-year-olds*
But you got Napster right in that comment.
@@encycl07pedia- _Money good, Napster bad!_
Metallica.
@@encycl07pedia- Thanks for the history lesson
@@encycl07pedia- Napster launched in 1999.
Bit rate affecting high frequency is an encoder thing, MP3 was designed for voice recording and there are lower roll offs.
AAC and Opus maintain the same high frequency range at low bitrate, the quality will suffer, a 160kbit will sound very close to 320kbit MP3 in high freq.
Other than high frequencies, quality mostly affect things like room reverb presence in recordings, aliasing like a vocalist matching the same note as a violin in background and mostly loud music. Aliasing sounds like a very light dial tone in tandem.
Unlike headphones, stereo speakers share the same air which introduces phase distortion, and the effects are amplified that causes annoyance not just noticable, e.g: vocalist with many instruments will sound like a phone loudspeaker until the instruments stop.
For AAC implementations, Fraunhofer and Apple differ in low bitrate, despite being compatible in most players, Apple's focusing on high freq.
For listening, audio going through a chain of Bluetooth, DSP, audio enhancments and speakers benefit from higher sample rate, example:
44.1khz content will have to be resampled to 48khz or higher.
Long chains are common in Hi-Fi and cars.
Most systems run at 48khz and higher as a preperation for correction and effects, since there is more content in mids and highs and these noticeably degrade at lower sample rate when processed multiple times.
Resampling isn't equal between operating systems, dedicated music players resample the content internally before sending to avoid issues. Playing 48khz content directly bypasses any imperfections.
E.g Windows 8.1 and below had a bad resampler, Pulseaudio defaults were very low; macOS, Android and iOS avoids resampling unless multiple applications are using audio or the DAC doesn't support the target sample
16 bit depth is the main problem when used with 44.1khz especially when using replay gain, most people can't tell between 24bit 44.1khz and 192khz.
Lowering the digital volume of a DAC or software mixer running at 16bit degrades quality quickly, hence why most DACs and players emphasise on controlling the analog output and not the digital gain, this isn't a problem when switching to 48khz with a resampler for the average ear.
A perfect way to forget about being at work!!! Chris is a saviour !
Looking back to the 90s, how many of us mistakenly compressed and replaced our entire CD collection into MP3 (16/44.1/128) to maintain that precious hard drive space? Now 128 kb/s seems so Napster.
Didn't know Donkey Kong had his own audio file format lol
Thanks for that very clear breakdown. One thing that podcasters (talking audio podcasts of course) had to balance was quality, bit-rate (and therefore) file size - especially going back 10+ years or so ago when they were just becoming popular. To keep our file sizes down for these almost exclusively 'spoken word' recordings many (of us) opted to drop the mp3 audio to 22050 kz MONO In fact after a few years we then decided to go better & moved up to 44100kz Stereo mp3 - only to get complains from some of our listeners! For health and safety reasons at work (although listening was allowed) they could only have an earbud in one ear to still be aware workplace dangers. Consequently they missed half of the two-person chat. Our solution - we switched to encoding in 'joint-stereo' mp3 :-)
th-cam.com/video/i6n-p9HdcOM/w-d-xo.html
my dude got 1M views? last time i watched u were 2k and never thought u would grow honestly. keep up. love the nerdy vibes
I amused myself once by taking a .wav file, taking its SHA-256 hash, compressing to FLAC and then uncompressing back to .wav The SHA-256 hash of the resulting file was identical so proving that it was bitwise identical to the original retaining all the information that was captured at the time of digital recording. Loss of information = loss of sound quality even if someone tells me my ears shouldn’t be capable of detecting the difference
Intresting...
Flac rocks !
Makes sense since it’s a lossless compression algorithm. No different to compressing a very large text file with zip.
So, you compressed and decompressed a file, and checked if any data was lost using a SHA-256 hash. Very nice, I never would have thought of taking the hash. My 7 year old brain would have just used dd to make a copy of the original, performed the experiment on the original, and used diff to check for equality.
Question - what was the file size of the original WAV, and how much did it shrink when compressed to FLAC?
@@Tom-uy4io I didn't actually check the compression ratio when I did this test but when I routinely convert .wav to .flac files I usually find that the flacs are around half the size or maybe a bit less.
I onced got into an argument with an audiophool who claimed that FLAC sounded better than ALAC. I did exactly the same experiment -- or, mostly the same, anyway. I took a source file, compressed with ALAC and FLAC, then decoded back to WAV. Then, to avoid any arbitrary differences in the WAV format headers (note), I used an audio editor to save the resulting decodes as raw PCM data. Then I used "fc /b" to compare them. "No differences encountered."
The guy's reply? "I don't care. I hear the difference, even if tools like that can't distinguish them." Sure ya do, bud.
(Note: If you write the bare minimum WAV file, the headers will _probably_ be identical too. But there are lots and lots of ways they can be different, while describing exactly the same audio data. This would ruin a checksum, and produce benign differences with a comparison utility, but the audio samples would still be a 100% match, so I dropped the headers before comparing to avoid that.)
FLAC is my prefered format these days. It preserves the 5.1 that I rip from DVDA and SACD. Plays back on my home theater and my car supports 5.1 as well.
My car too. 😊 How do you rip your SACDs? I have everything else covered from CD to UHD Blu-ray, but i thought SACD needed some special players hardware.
as a music producer, this is really helpful 😊
These are my favourite type of EC vids
Thanks. :)
As someone who creates music my go to audio file format is .WAV, gotta keep it as high quality as possible for the streaming services just in case the quality degrades a little on their platform 😎
Great summary of the different audio formats dancing around today; it's also worth noting that many "super high quality formats" (MQA, DSD or even FLAC at 24bits 192Khz) are nearly indistiguishable from a properly mastered, engineered and post-produces audio cd quality (16 bit 44Khz) file.This is easily confirmed with a double blind test; kind of an 80s Coke vs Pepsi challenge when you don't get a clue about brand, model or format of the subject being tested.It's made to avoid bias, which is a great problem in testing or "personal taste".
That's why it's important to separate music from audio quality.
And we don't even want to go into cable price snake-oil, which is a plague among audiophiles.
I never knew the whole "container vs codec" thing until today. Thank you for your concise and insanely informative videos!
Lovely video. Reminds me of the days 30 some years ago when I was taking computer science courses in college. Things certainly have progressed since then. Also you are a far better teacher than any I had in school. Thank you!!
Great video. I’ve lost count of the number of times I’ve tried to explain that MP4 files are containers and not a specific codec.
Would have been nice to mention that AAC can achieve the same quality at a lower bitrate than MP3.
I'm always a little bit smarter after watching an EC episode. Thanks, Chris.
Another really good video. My entire collection is a combination of primarily WAV and FLAC's. No lossy music for me. We have 20 TB HDD's, 6GHZ CPU's and dedicated hardware decoders or CPU's who can decode dozens of streams without breaking a sweat.
I remember when you couldn't listen to music on Linux because most of it was stored in a proprietary format and Linux was all about "open software." I'm glad that nonsense is behind us. But it would be great if the industry would settle on a single format in each category. It would make things much easier for everyone.
Holy acronym Batman..... this was both enlightening and confusing.... thank you Chris! Stay well, see you next time....(you had to remind me of what over 50 feels like...cheers 😞 )
Thanks for your support my friend. :)
I am obsessive with my music collection and years ago I moved to a home media server for better access to my collection. I decided to go with .mp3 for the file size and quality, but also the id3 tag structure that would allow me to make corrections to what many MP3 ripping softwares would make, such as selecting the wrong genre (in my opinion) or splitting an album between mutiple artists, because of one guest artist on the album.
As always, a really great video that I look forward to every Sunday morning!
I have archived my whole CD collection (way more than 100 cd's over many years) in the Flac format on my pc. Great video explaining audio formats.
Exactly the subject most relevant to my computer needs. This is my favorite computer channel on TH-cam. How about one on how to make a good amplifier from a Raspberry Pi.
Raspberry Pi by itself doesn't have enough amplification to drive a set of speakers. I've seen Raspberry Pi "hats" (accessory boards) that are DAC's (Digital to Analog) converters but no audio amplifiers that would drive a speaker. "Toids DIY Audio" had a nice feature about an amplifier sold by "Parts Express" that included onboard programmable DSP (Digital Signal Processing) that could be done from a PC to achieve premium sound for whatever speaker drivers you were using in your build. Hope this helps!!!
Build a Pi4 into a boombox. Better than using a laptop for entertainment. Power tool batteries, 5A buck converter. You'll want to power using 2 separate batteries. Android 13 because of higher over_voltage allowed for overclock. Accessories would cut out. I use mine for music, movies and tv anywhere.
Thank you. Another well organized, informative presentation.🙂
This is why you hit a million subscribers, Christopher - because you can take a complex topic and make it understandable and interesting. It's a great skill, keep up the great work :-)
well, you always post the most important subjects, explaining things that we deal with on a daily basis and that we probably didn't see in school. Nailed it, amazing job
Thanks :)
because i'm programmer i'm so frustrated thas the .ts file is recognise as audio file and not as typescript file
Thanks for a brilliant clear explanation, fantastic!
Thanks for the very well produce explaining video. I got into MP3 back in 94 when one of my cousins demoed how he was able to back up a CD audio that was about 600MB into a fraction of the size into his hard drive so he can listen it on his computer whenever.
I did not get into the other newer formats except FLAC since I didn't need to keep up with audio formats. Thanks to this video I am all caught up.
I deal a lot with videos and would like a similar video explaining the video codecs.
Thanks for this quick and very helpful overview! I have always been an audio guy at heart, working in radio broadcast engineering, as well as for companies who designed and manufactured audio equipment for broadcast and streaming.
Greetings Chris. Sounds like an interesting career.
YESSSSSSS, AUDIO FORMATS!!! THANK YOU, CHRIS! THANK YOU THANK YOU THANK YOU!! I LOVE YOU!!! 🥺🥺🥺
Over here in the colonies, we pronounce WAV as wave. I still can’t say the way the UK pronounces aluminum. 😂
And then it's funny when you find Brits insistent that aluminum is an Americanism that we just made up who then get mad when you point out that the term aluminum was created by the English chemist who first discovered the element.
Al-u-min-u-mmm
I’ve been looking more and more into all video file types lately. Time to see what audio has to offer. Great video as always 👍
Mr. Barnatt, I've been an audio nerd / tech most of my adult life, and I still learn something new. Thanks for the refresher between codecs and containers. I'm in my late 50s, and for me, yep, it's harder to tell the difference between totally lossless and lossy formats. Currently I'm downloading ALAC songs on Apple Music, which sounds indistinguishable from CD to my current hearing ability.
Very useful information. You always make easy to understand explanations!
Very informative; great work!
Great explanation behind some things i've always taken for granted, used the terms, and not know the details! Thanks!
Thank you for sharing your knowledge with us, I recently started my journey as an anime video encoder but wasn't really sure which audio codec to use, you helped me to clear my questions
Finally a video about Audio file formats love your video. Miguel
My whole music collection has been digitized and compressed into Ogg Vorbis format and takes up a lot less space than the alternatives while sounding great.
Mine, too, with the exception of a small percentage of downloads that aren't available in lossless formats. ~26k tracks (just under 150GiB at 192k) and still growing.
And while I am keeping an eye on its successor, Opus (very nice, both qualitatively and quantitatively), there's just no way I'm going to re-rip that many CDs. I prefer to keep things as uniform as possible. Plus there are a few idiosyncrasies in tagging that I don't like.
I have switched to it as the go-to audio codec for my video encodings, though.
Opus, the successor of Vorbis is better than Vorbis in every way.
• Quality better than Vorbis
• Encoding speed faster than Vorbis (libopus 1.5)
• Decoding speed faster than Opus (libopus 1.5)
In my wheelhouse now, this is concise and complete and I even learned some interesting trivia. Since I work with thousands of hours of production audio, I switched to ALAC and FLAC since the computation for compressing and decompressing is not a time factor with current hardware. Conveniently, each 3-hour reel-to-reel tape I do is 1 GB in ALAC. 3,500 reels will fit on a 5 TB drive with room to spare.
I use Opus and Flac.
Ahh Audio Formats I love it❤
As a musician almost exclusively using a computer to make my music, this is a great complement for my knowledge, thanks!
Wow, such a great and accurate video about audio formats! I’m a music producer and I use WAV for production. Most of the time I record in either 24 or 32-bit at 44.1 and 48kHz (sometimes higher if needed). Printing a mix is done in WAV 32-bit (takes up a lot of storage, though), only most of the time, 16-bit is enough and yes, I use dithering when printing a 32-bit project to 16-bit. I’ve subscribed to your channel.
Thanks for the sub -- welcome aboard! :)
Yes, I'm watching this video for fun
Cool! :)
I think the argument for lossy compression schemes like mp3's was more justified a few years ago when storage was more expensive. But nowadays, hard drives are incredibly cheap, so I see no reason to use lossy compression anymore. I can understand bandwidth costs, and therefore the need for internet-based radio stations to use lossy formats, but with paid-for music streaming services, [I use Qobuz] uncompressed is the way to go [though technically it is FLAC lossless]. Anyone with a decent internet connection is able to stream 16bit 44.1Khz content with ease, which is the minimum needed for high-quality audio.
I'm still learning every time I watch your videos. Thanks for sharing, Chris.
Greetings Brian. :)
was going to do some research on audio file formats since most music players i use doesn't support opus and hey what do you know my favourite tech explaining youtube channel posted exactly that recently!
wonderful stuff chris! keep up the great work :D
Greetings! Real human comment
Did you know that "The Mother of MP3" is Suzanne Vega? Her acapella song "Tom's Diner" was used as the basis for the compression algorithm by Karl-Heinz Brandenburg.
Suzanne is probably better know for tracks such as "Luka", "Marlena on the Wall" and "Blood Makes Noise".
This I did not know!
@@ExplainingComputers My job here is done! However do look into it (fact-checking and all) :-)
Personally, I buy CDs and rip them into Flac files. I love how your videos look and sound like the 90s. I was born in 2003, so its an experience I never had. Subscribed.
this is for me going to be very beneficial i have played with a few formats and never checked each one thank you for covering this information in brief is helpful
Would be nice to see tutorials on how to create audio in different formats - especially lossless
MP3 is still the king of audio file formats, because of its presence in the majority of audio players. My personal favourites are Opus and Flac. All my Beatles albums are in Flac 24/96 ripped from vinyls.🤣
If you Own Nokia Phone Back In The Day.
They always Used AAC Quality For Ringtones 👍
Given that MP3 is over 30 years old, it's surprising how well it still holds up today. It can still keep up with modern codecs and is perfectly usable, although I'd say that AAC is probably more prevalent in newer media.
@@cromulence there was a time when I used to have my mp3 files on 128kbs range. I can’t stand the sound of files lower than 192kbs.
@@dougr.8653 Agreed, 192 is really the viable minimum, but 128k files were my first foray into music files I could actually keep on my PC!
Happy Monday 😁
Even though we all like to pretend we are precious individuals who constantly rebel against 'the system'... we are a species of standardization and standards. We are for the most part much better off because of it. From standardized measures to audio file formats, we humans certainly like our agreed upon standards as well as constantly trying to improve upon them. Great video!
I'd be interrested in a algorithm comparison of a few of these, you could do that with explanations, compress-decompress and then compare the waveform or possibly even analysis of the raw data of the compressed files and seeing what edits do to a decompressed sound
FLAC for lossless and OPUS for Lossy. Simple!
If I hadn't seen Friday's member's only video, I would not have known that this video was edited in Linux with DaVinci Resolve and other production software. I must say, I'm not sure how keen I am on the animations you made for each bullet point on the slides. I much preferred the older animation you did, but I do understand that you were probably trying something new for you video production experiment.
Thanks for this. :) I too prefer the older transition, but had to run with what I could (relatively quickly) set up in Autograph in Linux. The old bullet animation will return. :)
@ExplainingComputers that's good to hear
Thanks for that very clear explanation. I was about to add this to my list of bookmarks for your other Explaining ... Formats series then see that you've listed them in the description.
Наконец-то!!! Нашелся человек, который сказал правду про MQA. Браво! А не пудрит мозги людям со степенью развертывания
My choices for files are ogg/vorbis and FLAC for current use, mp3 for legacy devices. Kia stuff can read ogg vorbis, maybe other cars can too. Opus is cool but good luck getting native car support for it, unless you know otherwise. It just doesn't offer much over vorbis to justify using it.
Oh and for movies, 8 channel vorbis works great with Kodi.
Opus:
=> Much better quality than Vorbis
=> Higher encoding speed
=> Higher decoding speed
=> Low latency
=> Even more efficient for small files like sound effects
=> Surround optimizations (The LFE channel gets lower bitrates since it only needs a very low bitrate)
@@EmergedFromReddit It might have better quality than vorbis *for a given bitrate*, but this isn't 2004 where storage is tiny. And not sure why decoding speed means anything when vorbis has been hardware decoded for over 15 years on SoCs for android and other chips. Latency also means nothing when listening to music. LFE is a thing but again, means probably little when dealing with gigabyte size files.
And again, opus means nothing to me unless my car supports it.
Atrac! Brilliant quality on Minidisc
Agreed -- I still use Minidisc. But sadly not a mainstream audio format in use today.
@@ExplainingComputers Replaced by Mobile phones. Jack of all trades, master of none.
Back in the early 2000s I transcoded my Vinyl collection to high bit rate MP3. The best we had at the time was 192k which, as far as I could tell, decoded back to identical sound when compared to the originals. A couple of years later, I re-coded them to be 320kbps, which still sounds just like the originals.
Since then I've stayed with MP3 format for one simple reason ... I can curate them with MP3-Gain to play at the same -16lufs level as streaming services use. So my entire collection now plays at practically the same loudness allowing me to start long playlists and enjoy them without constantly interacting with my music system.
It's a great solution for me... Good sound quality and no huge racks of stored LPs or CDs, no wear and tear, easy local storage and all nicely at the same volume.
It's pretty amazing what you can do with the right software.
I did exactly the same.
Except once I got computer with two 500 mb hard drives, I redid them third time into .WAV
@@jhutt8002
Yep. Wav files are pretty big, though. My current collection, which has grown rather a lot, occupies about 30gb on disk. The whole thing is now housed, along with 1,400 transcoded movies, in MP4 files, on a 4tb NAS drive in an external housing.
@@Douglas_Blake_579 500 mb... Yea, sorry bout that, I of course meant 500 gb!
They are large indeed, but storage is cheap these days.
I upgraded my computer with two 1 tb drives couple of years back, as I couldn't fit my digitized library on it anymore. Currently at 456 gb of WAV + 387 gb of ISO files containing virtual copies of my music CD's.
Great presentation as usual. "Tagging" is an additional considerations if you have a music collection stored on (for instance) a home server - sourced from ripped CDs, and purchased digital downloads. For that use case you want file formats which support storing tags (eg metadata such as album name, artist name, genre, cover art picture). But in addition there need to be music players, music servers and tagging editors which support those formats. Most of my experience is with MP3 and FLAC files which have comprehensive cover.
Some applications claim that they support tags in WAV files, tho I have never encountered tagged WAV files.
I'm 26 and with the headphones I can reasonably afford I couldn't hear the difference between FLAC and 320kbps MP3 to save my life.
I'm 20 and I can barely hear the difference between 256 kbps AAC and 16-bit 44.1 kHz FLAC. I'm totally content with lossy, even though it's worse. I grew up on WMA 128 kbps, so anything that or higher, I can enjoy
Can you hear a difference between a 96Kbps Opus file and a 320Kbps file?
@@EmergedFromReddit I don't encode files in Opus so I wouldn't know. Opus and AAC on TH-cam music is a far cry from FLAC, ALAC, or 256 kpbs AAC / 320 kbps MP3
Just what I need! I've just picked up an iPod classic, no idea which file format should I use.
I copied a lot of files, ripped from my CDs, to it in MP3 format. It sounded fine.
Like the other guy said, MP3 is fine. Stick with 320k. It does support ALAC (at least mine did) if you want lossless, but it will affect battery life. Unless you have some high end headphones you probably wouldn't be able to notice the difference between lossy and lossless anyway.
If you flash the Rockbox firmware, it'll play 320kbps FLAC no problem. That's what I've done to mine, along with the iFlash SD card adapter, and not only does it hold more tracks than the original hard drive, but it's more battery efficient too.
This was a much needed video.
Excellent work!
I love learning. Thank you, Christopher!
This is the first time that I've not heard ".WAV" pronounced 'wave'.
wav is wav, we don't make the rules
I am sitting in the morning at the diner on the corner...
Suzanne Vega?
@@michaelhall9211 yes. Not randomly mentioned either, that song was used to test the digital compression schemes while developing mp3. She's sometimes called "the mother of the mp3"
Just this week I went and watched one year old videos explaining video formats and codex....then this comes out!!
An interesting coincidence!!
Spooky! :)
@@ExplainingComputers
Another TH-camr was saying something about video codex es, and showed a list of videos on TH-cam. I recognized your thumbnail style in the list I went looking for that particular video. Turns out I had already watched it, (and give it a thumbs up), but I watched it again....
AV1 and AV512 we're not mentioned in the video, but I still got the idea....though I wonder where they fit in.
Brilliantly explained Chris, thank you very much.
Do to the fact that we now have terrabyte hard drives, all lossy formats should be gone including mp3. Same goes for lossy picture formats like jpg.
Network bandwidth is also an issue, not just storage space. :)
@@ExplainingComputers
I know, I'm just having wishful thinking. 👍😃
Audiophiles incoming in 3...2...1...
They would argue which format is the best until the end of time. However, there was an experiment which proved that even the best music student with perfect pitch could not even differentiate between mp3s and lossless audio. Great video as always!
As always, very informative video. Thank you Chris.
Informative video as always, Chris!
Uhm, you made WavPack (1988) out to be older than WAV (1991). Surely that's not correct.
WavPack is from 1998, so there must have been a bit of a mix-up there. Great and comprehensive video as ever overall though!
My bad -- typo. :(
Thanks for going into such detail. I already knew some of this, but even that, not very well.