Sound By Numbers: The Rise of Digital Sound
ฝัง
- เผยแพร่เมื่อ 24 ก.ค. 2018
- You can support this channel on Patreon! Link below
These days, listening to music on the analog formats of old has seen a massive resurgence. Whether it’s cassettes or vinyl, it seems analog is in and mp3’s are out. But digital sound is pretty freaking neat! This video discusses how it works and is the launching point for this, the final videos on sound reproduction.
You can support this channel on Patreon! Patrons of the channel are what keep these videos coming. With the support of people like you, this channel has become my full-time work, and I’ve been able to purchase new equipment and spend more time learning video-making skills. If you’d like to join in and support the channel as well, please take a look at my Patreon page. Thanks for your consideration!
/ technologyconnections
Technology Connections on Twitter:
/ techconnectify
The subreddit (warning: I’ve no clue what I’m doing here, but we’ll learn together)
/ technologyconnections
Thank you to the following Patrons!
Hank Eskin, Brady Brown, Giacomo Scaparrotti, Winfield Trail, Luke Biddle, Marc Ethier, Joseph Hill, John Semander, Bill Basch, Robert Toth, Magnus Tranum, blouerat, PK, Richard Lindsey, Lars JJ, Patrick Musson, Matt Taylor, Johan Schoeman, Thomas Jew, Josef Citrine, Marcus Schwartz, Ryan Cozzubbo, Christopher Bassett, Julian Rapoport, Ed Hall, Jason Portwood, Jesse Cardone, Michael Snowden, Drew McClain, Boh00711, Jon-André Haraldstad, The8TrackKid, Daniel J Schless, Sebi Jecklin, Casey Kikendall, Craig Leverenz, Charlie Hankin, Andres Plaza, Revenge, Matthew Baron, André, Gerald Monaco, Tab Patterson, Jason Weathered, h.drew foy, Roger Beal, Gantradies, Chris Hartl, Steve Stone, Brian G. Shacklett, Colton Aubrey Hooke, Paul Mills, Christopher Luna, Kotanu, Brendan Terrett, Techmoan, Mike Mason, Elijah Reeds, Christopher Jett, Justin Talbot, Steven Metcalf, Verysofisticated, Lolucoca, John McLusky, Ashleigh R, Stephen Bell, Logan Beenken, Eric Wood, Stephen Pick, BlancoGrande, Matthew Giraitis, James Ryan, Robert Joscelyne, Francis Fisher, Kedar Deshpande, Lukas Komischke, Dave Anderson, Felix Freiberger, Dane Peterson, Brent LaRowe, Aerospyke, Kieran Cox, Hunter Schwisow, Logan Kriete, Rafał Wiosna, Adam D. Ruppe, Audin Malmin, Eric Hansen, Noah McCann, Jason R Scheuren, Rufo Sanchez, The War Academy, Yaniv, neko, Daniel Ch, Lee Wallbank, Grayson Lang, Neil Bronner, Carl Yazbek, Jeffrey Grajkowski, Christopher Splinter, Eric Merrill, Stefan Berndtsson, Michael Sims, Seth Kneller, Clint Eisinger, Jesse G. Donat, Christopher Leidich, Mike Friedman, Svante, André Waage Sørensen, Ed McCloskey, Philip Daniel, Jasper Thun, Camilo Rodrigues, Dusan Dzelebdzic, Sam Douglas, Sam Redfern, MagicMatteo
Don’t see your name? Don’t worry! Due to the sheer volume of you fine folks, the $5 patron description shoutout has gone on a rotating basis! Rather than eliminate this perk, we’ll just take turns. Right now, you should appear at least every third video. So if you didn’t see yourself the last two times, don’t worry--you’re up next! - วิทยาศาสตร์และเทคโนโลยี
![[TH] VCT Pacific Stage 2 - Playoffs Day 3 // DRX vs GEN | PRX vs TLN](http://i.ytimg.com/vi/CWM1Q-2-o3M/mqdefault.jpg)
OK, time to pin a comment (please read! It's good! It addresses a point you might be about to make in a comment yourself!)
This video is the first of a few (at least). Some things were _intentionally_ left out because I don't want to be making a full-length motion picture here. The most contentious omission (which I now realize I should have identified in the video more clearly) was the fact that the samples are connected in a very specific way because, due to the Nyquist-Shannon sampling theorem, a bandwidth-limited signal can be represented perfectly using a sample rate that is twice its bandwidth. The dots themselves define how they must be connected. That is a complicated thing to grasp and explain quickly, so it was glossed over as "The DAC will smooth out the choppiness of the samples a bit to make the resulting sound a little more natural". Now, this isn't strictly _incorrect_ but it does not explain the mechanism by which the samples are smoothed out--and that's important to understanding how digital sound is encoded and decoded. And it also might imply that there is no detail outside the samples--but in reality there is! And trust me, the next video will address this. That was always the intent.
BUT, I will admit that was sloppy and too simple.
Don't sweat it, you're doing fine. I briefly studied this stuff in college and built a DAC as part of a project (the details of which are now very hazy), and what you're saying rings a bell, if simplified.
Xiph has a great video on the subject of sampling and quantization, for those who are interested in knowing more while they wait for your next video: xiph.org/video/vid2.shtml :D
+Technology Connections Actually, for what it's worth, I think the description in this comment is a great and simple explanation. Certainly better than anything I've come up with. Thank you for posting the clarification. Can't wait to see the rest of the series.
Technology Connections
Headphones and microphones are essentially the same thing...
Ever plugged a headset into a microphone port?
Obviously not because you can still record sound that way (through a head phone system without any integrated microphone to be clear).
Yes, the headsets are definitely doing the recieving. This fact becomes blatantly apparent when you listen to the recorded audio, still proof positive that the two devices are fundamentally the same things with different electrical profiles one far better tuned to record audio the other best tuned to listen to the recorded audio.
Good thing you addressed it here - cause there are so many people that come with their half-knowledge (if they have any) and claim they can hear the supposed stair-steps.
It's very satisfying that you made sure that whenever a sound wave was shown on screen, that it visually matched up with your voiceover. Even the DR-05's dB meters matched up in that close up of it. Also: terrific episode.
Yes that was excellent! :D
One of the most instructive videos I have ever seen! Fun to watch too, amazing!!
Wow! I din't notice that - fantastic!
Indeed, I was noticing this at (or shortly after) 9:28 and came to see if anyone had commented... glad to see I've been beaten to the punch. Also nicely demonstrated at 10:15 and 10:46. Nice work, Alec*!
* completely unrelated side-note to Alec: I somehow was under the misapprehension at some point that your name was Jared. No idea where I got that idea, but I misnamed you in an earlier comment... which if I can find, I'll happily correct... but I'm not sure I know how to find it. Anyway, apologies for that.
@@unfa00 you don’t notice the best edits… it’s funny how our brains work
"64K of CD quality audio lasts this long: [blip]"
I giggled, ngl
12:58
Today I learned 64kb is just enough to exhale once into a saxophone at CD quality
"64 kilobytes of digital audio sounds like this. Doot de..."
THIS VIDEO HAS BEEN REMOVED DUE TO A COPYRIGHT CLAIM BY SEVENTEEN THOUSAND COPYRIGHT HOLDERS.
That's about 0.37 seconds worth of CD audio ! 65536 / (44100 * 4). Two channels times two bytes for each sample.
Royalty free music, baby!
Doot Doot 💀🎺
Kilobyte or Bit for that matter, is already plural you don't add an S to it
@@codname125 that is incorrect, I believe
"Project Green Sally X System Hornet One" Hilarious! Or maybe I'm just easily amused... ;-)
I liked your graph showing the 20-sample, 4-bit digital conversion of a sine wave into numbers. My son asked me a while back how digital audio was recorded, and I made a huge mess of trying to explain exactly what you explained simply in 60 seconds. I'm going to have him watch this video so he'll be able to understand it.
You may be surprised by how poor your understanding is of how digital audio works. I know I thought I had a very strong understanding of it until I watched this video about it from Monty Montgomery. I suggest you check it out if you're at all interested. He even uses an oscilloscope to show how the waveform generated from a digital recreation of an analog signal is identical to the analog signal. If that sounds at all interesting to you (I found it fascinating, and come on, you're watching Technology Connections so you probably would, too), check it out here:
th-cam.com/video/cIQ9IXSUzuM/w-d-xo.html
booooooo
@@andresnider4926 oh boo yourself
@Jeffrey Craig I recently bought an old Atari 800XL computer which has 64K of memory and came out months after the compact disc hit the scene. So it really hit home for me that that short blip of audio would completely fill up my computer's memory but there were already small discs that held over an hour of similar audio.
At first I was like "with modern computers why don't we record the fourier transform of the sound instead of sampling like that" and then after a quick bit of Googling I realized I just described how mp3s work.
Joshua Hillerup great minds think alike, who ever the inventor(s) of MP3 were.
Mosco Monster thanks, although thinking "lets do a Fourier transform" is the easy part.
Because just taking the fourier transform does not change the amount of information, and puts additional load (doing the FT) on the hardware?
Cutting out the bits that nobody perceives anyway, though...
@@brandy1011
I think my professor once told me that the fourier transformation made the effective use of acustic test units (basically sonar) for powerplant pressure tubes possible but not sure ;D.
People have naturally thought about all sorts of creative ways of compressing data. It was a must.
Sir, your videos are just awesome. The quality, editing, dry humor and easy to understand informations are absolutely brilliant. Definitely one of my favourite channels. Keep up the good work!
Agreed!
Alexander Grasl well said
+1!
So the Tascam has two little elves inside it furiously writing down numbers just like cameras have little elves in them furiously drawing and colouring in pictures?
molson12oz, Jeff Session elves are only used in older cameras, as they don't support color photos ;)
Remind me of the Flintstone!
Terry Pratchett invented the 'Iconoscope' , a colour digital camera, rather early!
"The iconoscope didn’t take pictures by letting light fall onto specially treated paper but by the far simpler method of imprisoning a small demon with a good eye for colour and a speedy hand with a paintbrush.
"
Terry Pratchett, The Light Fantastic , 1986
@@NoiseWithRules Yup! Good one sir! Discworld technology was driven by either little demons or little goblins from what I remember from the books. :)
@@NoiseWithRules that's technically not digital, but analog. 😜
Unless demons use binary logic instead of neurochemical methods?
@5:11 - Doo-rectly...doo-rectly?
Gets me every time XD
This channel must be used at technical schools in electronics classes. You explain things like teachers should explain. Congratulations.
My brain overheated from learning. Is that what was supposed to happen in school?
Thanks man, you are a true nerd, i like your explanations. You probably play in class rooms all around the world by now, uh?
Looking forward to the rest of the series. Cheers.
quick copypaste from the tweet for nonfollowers
"Real quick:
I'm about to release a video.
The video has a continuation which is not yet written or filmed.
There are things that are not mentioned in this video which will be covered shortly.
So please don't tell me of the things I have forgotten to mention.
(cue exactly that)"
Sooo...theoretically, if I turned my ears insde-out, I could turn my head into a loudspeaker? Neato!
Your head already has a loudspeaker, but its more of a variable pipe diameter bagpipe with a variable resonance chamber and outlet shape/ size.
Its pretty adapt at mimicing actual speaker sound. Its the mouth
If you had muscles in the hearing bits to move the ear drum. Then yes it's a loudspeaker
would work as good as a hammer used as a nail
please don’t
I've been around ADC/DAC since forever and I've never heard anyone pronounce ADC until you. So bravo good sir, well played.
ADC/DAC were one of my favourite bands back in the day.
@@spodule6000 And Metaloxidic was the other? Just remembering a dumb TV cartoon of the ‘90s my then-6-to-8-year-old son liked, in which the main characters always wore tee shirts with the names of two rock bands, which may have inspired the “Dumb and Dumber” movies.
@@allanrichardson1468 um
i dont know if i should whoosh you or
I have been binging your videos the past few days and fell in love with your charisma, your easy explanations, but mostly the depth you dig into the subjects, sidetracks, revisits and trivia you discuss. Thank you so much!
I loooooove your style and channel. I will share this video several times, to many friends who've brushed away how simple it can be to understand sound sampling.
Great video! There is no one quite as good as you when it comes to summarizing the historical evolution of the development of various technologies. Even if I already know a lot about a particular topic, your historical approach is always enlightening. Thank you for all your awesome work!
Please cover the Nyquist Theorem in your next episode. It is a very important part of digital audio sampling.
Is that the bit about sampling 2.5 times higher than the highest frequency you want to capture?
Still though, there probably should have at least had a quick aside at around 9:20 just to say that ~20kHz is the generally accepted maximum frequency that can be heard by humans (with most individuals being much lower), then just say the reason is the Nyquist-Shannon Theorem but that is beyond the scope of the video.
Actually it is only 2 times, but yes. en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem
It's how sampling at 2x the *bandwidth* of the signal you want to capture is enough to *perfectly* reconstruct it. So if you assume you only need to record 0-20khz, a 40khz sample rate is enough for perfect audio, that's what the theorem says, and it's kinda important.
I also hope he will cover the myth about "stair-stepping". In reality, the distance between the digital audio samples is neither a "stair step" nor a straight line; it is undefined. And a good-quality DAC uses splined curves to "connect the dots" so that a near-perfect replica of the original analog waveform is reproduced. xiph.org has a video demonstrating this using an oscilliscope on the output of a DAC.
I came across your channel recently and really love how you explain the various technologies. I'm an old techie, but I still pick up a few things I didn't know whenever I see one of your videos. Good work! One of my more favorite subscriptions!
Really enjoy your videos, thank you very much. It is obvious that you have a passion for the materials you cover and that light shines through in every video you make. Such a pleasure to watch!
Great video. It would be nice if there would be a video about getting out better analog signal from computer, something like debunking some myths of "audiophiles" or enlighting us. Maybe something that touches sounds cards. This could a series like it was with LaserDisc and that would be super awesome.
The production on this channel just keeps getting better. Keep it up!
Loved the video. The dolly shots and macros look gorgeous, your stuff has come a long way in such a short time! And your explanations balance thoroughness and approach-ability in a fantastic way, it's the kind of thing that inspires me to take the time and prepare classes for my students.
This was one of your best produced videos yet, and lays a great foundation for a series. Keep up the awesome work!
At last a simple explanation of digital recording, very understandable. Nice video production too.
You should be a teacher man. Bravo! I'm always amazed at how interesting you can make seemingly any topic!
Love the detail in this and all of your other videos. Great video. I appreciate the time you take to really grasp the concepts. You’ve definitely opened my eyes on all the electric car misconceptions. It’s also so nice to hear you drawing the line about digital audio being an “absolute”. I love vinyls, but I have had trouble understanding friends on the whole “vinyl has a superior quality/fidelity” idea. Thanks!
I'm so envious of all the stuff you seem to have! Love the video.
the best explanation of how sound is recorded I have ever heard...text-book, my friend!
How can you be THIS funny, without making jokes and without loosing even a bit of seriousness of the topic? Witchcraft!
I imagined Alec himself saying this.
Technically, one of the best channels on TH-cam! :)
Love the new studio, man! Creates the right setting for the video
I really love this channel man! Keep up the good work
“That’s not super helpful” - My favorite line from this video.
One note to add: technically, ADCs don't directly draw straight lines between the points on the graph (really, really old/cheap ones might hold the value instead).
Instead, a line (read graph solution) is drawn that both goes through all the points on the graph, and does not contain any frequency above that of half the sampling rate.
Due to how the formula for deriving this line works, there is only one solution that can fulfil both criteria, and this is our original analogue waveform (pending rounding/dithering).
The ideal reconstruction involves plotting something called a “sinc function” centred at each sample, then adding up all these curves. The trouble with this ideal is that the sinc function has an infinite extent and only slowly fades to zero as you get further from the centre.
But on the hardware side, a low pass filter will smooth even an ADC that just abruptly change from sample to sample since the square waves created contains filterable harmonics at frequencies related to the sample rate.
Such a filter on the output side is called the “reconstruction filter”, while on the input side it is called the “antialiasing filter”. In both cases, their job is to get rid of spurious frequencies above the Nyquist limit. The reconstruction filter is a “usually-good-enough” approximation to the ideal sinc reconstruction function.
One can also use DSP to do the sinc function (often as part of oversampling) which allows a simple reconstruction filter.
If you're curious about a more extreme case of this, look up how a Class D amplifier works. It is also known as a digital amplifier. It basically uses the signal to modulate a high voltage PWM wave and passes that through a reconstruction (low pass) filter to recreate the sound at high volume, but with very low distortion and noise.
awesome video TC, it's great to see your work improve. I have to appreciate the high level of your videos, it's started my interest in VHS and it's fascinating stuff.
i'm so in love with your channel. stuff i already know but still, fun to watch. keep up the great work.
I wonder if you digitally recorded audio on vinyl and recorded of 1 and 0 of an mp3. I wonder how long would the recording last per one side of a 33.3 and at what highest bitrate?
Considering that you could tighten the groove margins considerably since only 2 states would be needed and 2 states isn't practical to pickup with coils so light would be better so you would get a laser to look at the grooves, black is impractical so better to use sliver and oooooh.
Dragon Skunk I was thinking the same thing, though the data would probably be encoded like the sound of an old modem. I found this link, but I wasn't able to follow most of it. www.reddit.com/r/askscience/comments/12iyb9/how_much_data_does_a_vinyl_lp_hold/
Nice!
You probably can use modern modem technology to encode data to the vinyl using all sort of rotated diamond whatevers they use.
Supposing you can use two channels as independent "modem signals", and you use 40Kbps on each, you get as expected 80Kbps of bandwidth, that is probably enough to deliver with OGG a similar quality to 128Kbps MP3.
And given the 52 Minute size of a regular LP, that's around 32MB per disc.
If you want an idea as to how that might work, Retro Man Cave has a good couple of videos detailing how digital data was stored on a record, though that concerned software where the data was modulated into an audio signal. Something you'd very likely have to do for a record just as you did on a casette tape, floppy disk and even a magnetic hard disk.
"black is impractical so better to use silver and oooooh" i feel like this went over a few heads lmao
The point of digital audio is, I read somewhere, to "separate the signal from the system" you just hear the intended signal, not the tape hum, or the record rumble, or the radio static.
Where's the fun in that? ;)
If you want bad and awful and terrible then the nasty fart noises created by the sega genesis are the worst sound ever created by a digital device.
I've never heard it like that, but I can see where they're coming from. Slight problem with the logic is that signal noise is ALWAYS there, it just may be quiet. Even though my room isn't exactly good for a studio setting, my cardioid condenser mic doesn't even pick up the fan of desktop (it's fairly noisy), and while it does pick up the sound of my keyboard (it uses Cherry MX Red switches), and it does pick them up, but it's quiet, even when I'm pounding my fist onto it (it helps that my mic is mounted to a shelf next to my desk, AND is in front of the keyboard).
Indeed, at our FM transmission sites I'm converting the link between signal acquisition and the FM transmitter from analog to digital (AES3) to make the sound more consistant from city to city (eliminate multiple sources of noise, eliminate multiple signal amplifiers (also splitter) that sometimes has too much gain just to be attenuated at the destination, mitigate the effect of ground loops, eliminate discordance between low and high impedance inputs and outputs, and more). That makes it much more likely to transmit exactly the signal that was receiver original.
Yeah, no shit high quality expensive multitrack tape is superior to any consumer tape format, that's the point. It's supposed to have an extremely low noise floor so that by the time you mix down and master to get to the distribution media, the noise floor is still low enough to be acceptable to the listening public. It's the same reason why professional studios record at 24 bit or more, and why people like Neil Young have such a hard-on for selling it to the public, even though it's complete overkill because a CD's SNR is over 90dB which is more than enough to reproduce any kind of music you could possibly imagine, short of maybe 1812 Overture if you wanted to simulate sitting right next to one of the cannons and permanently damage your hearing.
And also, what are you talking about with the hard drive thing? A computer hard disk (not to mention floppy disks, tape drives as well as the defunct Digital Audio Tape (DAT) format) encodes the data magnetically as 1's and 0's, ie. digitally, not as a varying magnetic field like analog tape. It's either on or off, that's it. If you somehow managed to modify a hard disk to record analog audio on it, it would behave kind of like an old style wire recorder and would definitely degrade over time, though not as fast as tape as the heads do not physically rub the disk. Even digital data on a hard disk degrades over time, but that's why we have error correction algorithms.
I love this channel. You are so good at teaching!! The explanation of bit and sample rate was made so easy to understand. Thank you and please never stop making videos!
Best intro to a video dude. You connected with the audience using a familiar object and relating it to us. I am impressed. You are getting gooder!
Edison Player: put a good belt on it...
Your hair keeps getting cooler and cooler
Alexander Grasl inspired me to express my gratitude for your channel. I'm so glad streaming has allowed driven, creative, ambitious and generous people like you to create content we could only fantasize about when tv was the norm.
I love what you do man! Really glad you're putting this information out. I learn so much from you!
I wish you would have made a clear demonstration on why digital is more robust than analog to distortion and damage.
One thing you could have done is write down all the numbers taken from the ADC sample onto a whiteboard or transparency sheet, using a dry-erase marker. Then, you could apply random distortions to these numbers by rubbing them randomly and erasing bits of it. After this treatment, you can show that these numbers are still essentially readable. Even after all that abuse, and even a couple unreadable numbers, you can still recreate the wave with almost perfect accuracy. In contrast, you could make a model of an analog recording with a paperclip or other pliant wire, shaped into a wave. You could apply distortions to this model with a hammer or clamp or something, and it'd be all messed up. Finally, you could compare the regenerated digital wave to the beaten up analog model, showing how the digital recording was more robust.
I think you are driving at a very important point here and I hope Technology Connections will consider it as the subject of a future video. I believe your suggestion can be even refined further. He could start with opening a brand new vinyl record and play the first track on a high end record player. After doing so he takes a pin or some sharp object and scratch the record. The same track will obviously sound different now with that inflicted damage. Moving on to CDs now, he can play a new CD in its player. Take the CD out and grind a few radial slots in it with a thin grinding disc ( there's another very good video demonstration of this, "Errors aren't forever m.m.v. Stan Baggen,Jack van Lint", about 20 minutes into the video). The disc should still be playing perfectly after that provided, of course, he did not turn the disc into some piece of confetti. :D
I’ve always heard ADC pronounced as the individual letters, not as a “word”. DAC of course is always pronounced as a word. Saying “adc” sounds too much like coughing up a hairball, possibly followed with “pfft”.
I realize my brand of humor is sometimes sillier than necessary, but rest assured that was in jest :)
Haha - I love how you used it a few times through the video.
IEEEEEEEEEEEEEEEEEEEeee
I had assumed that the first time was humor when I heard it. I guess I began to doubt when you stuck with it throughout.
Hey man, I pronounce "FCC" as "fcc" just to be contrary. It's all good 😎
This is my favorite video of yours so far, by far. You turn history and science into a compelling story. Great work!!!
You're video are really great. I don't watch everything you upload but I learn something cool from the ones that I do. Thanks, man!
Another two bits of miscellany that the audiophiles drive themselves absolutely crazy over. - "Quantization error" and "Dithering".
And "Jitter" and "ehrmagerd digital stair steps!" on and on and on. The rabbit hole only goes deeper from there.
Stairsteps happen on only the cheapest or oldest DACS
Nope, they dont happen ever. There is no stairsteps.
Ugh, don't get me started on stairsteps, you take one graph and quantify the x value as a function of time and all of a sudden people think that's actually what's going on. No dang it! Even a naive dac will simply push voltage and the loudspeakers will travel the analog distance required between sample points, never mind the dacs that read ahead, buffer and build lines of best fit then super-sample the output back to the speaker so that the exact representation of the sound is sent down the pipeline. It's simply just that lollipop graphs are harder for our stupid meat brains to wrap around.
Quantization errors, ugh, you push a technology past it's limits and it breaks, how is that a weakness of the technology and not you being a dumb-ass, even if it records a bad number check-summing will bring the signal back to where it's supposed to be, oh, your recording isn't 1/44 THOUSANDTH accurate to the original? Farts, snacks, how do you even know?!
And dithering, dear god dithering. How the hell can people complain about sending harmonics to tune the noise floor to actually have it LESS noticeable, even though it's damn near beyond the range of human hearing.
Quantization? Dithering? Stairsteps?
Anyone even remotely confused or interested in any of that stuff needs to watch Monty Montgomery's video. Here, I'll post the link for you. Learn something: th-cam.com/video/cIQ9IXSUzuM/w-d-xo.html
In FM transmission with with digital AES3 signal (which is not new), most of that are not an issue. Most of the aliasing is frequencies beyond our hearing capabilities at the standard 48 kHz (and therefore can only affect marginally the energy density of audible frequencies (that would be the only argument for higher sample rates)) , and can be largely eliminated by applying a low-pass pre-filter (but modern broadcast equipment has been dealing with those issues for a long time and they deal with them easily). Jitter is an issue, especially with AES3 reclocking, since the standard is somewhat old, they used an old serial transfer protocol (instead of packetized) with a baud rate directly related to the bit depth and sample rate, so a slight frequency difference between the oscillators (like, not a perfect 48kHz) of two element in the chain (like a switcher) can make audible ticks in the signal sometimes (a few times a week). The solution is that if you have multiple elements in your AES3 link, to make sure the clock of an element synchronized to the source. My personal favorite is to use a IP steam instead! If the link clock is significantly higher than the signal clock (like putting stereo PCM in a 100 Mbps ethernet link) then jitter there is no issue (assuming the link is reserved).
This video makes me want to etch 320Kbps MP3s on a wax cylinder.
LOL ! Now there's a good project for you. Not all that impossible, really.
Awesome explanation! Looking forward to the next episode in the series that got me interested in this channel in the first place
the visuals you really make it much easier to understand, thanks for your hardwork!
Please address DSD in some part of this series.
somdusazerate I second that.🍻
Based on how simplified his explanation for encoding pcm was I doubt that he would go into an even more unintuitive system like DSD. I don't think it would fit for the vibe of this series because giveing an analysis on a topic that in depth would make the simplification somewhat misleading. I think he barely got away with the explanation in this video.
+bob caldwell
You've gloriously hit on one of my pet peeves and let me enlighten you on why it is such. This is based on me making an assumption about your intent here, so grains of salt etc.
"I think he barely got away with the explanation in this video"
This assumes that I am finished explaining this. There is clearly going to be another video soon addressing this topic of digital sound. If you are upset by my use of "connect the dots" (which I'll grant is somewhat misleading), just know that I intentionally did not bring in the Nyquist-Shannon sampling theorem here because doing so seems like a clunky side-note that would rapidly derail the path of the video. I *did* say that the DAC smooths out the choppiness of the samples, which was an incomplete (but not inaccurate!) way of explaining that the digital samples *do not* create a stair step waveform. I simply didn't explain how this is done, and why a simple reconstruction function can recreate the original bandwidth-limited waveform because that's getting far more technical and pedantic than I'd like to for an introductory video.
In every presentation on every topic, choices are made on what to prune out for length and flow. My choice here regretfully does not make clear that I was oversimplifying specifically this thing right here and that I would elucidate later, but knowing that a second (and probably third!) video is coming would hopefully persuade someone to refrain from shouting into the comment section that the creator "forgot something" or is "wrong".
We will get there in time.
Sorry for my jumping to conclusion then. You put vastly more effort into the video than I did my comment so any judgment on your part is fair. Don't get me wrong, I loved the video and don't care as much as some people in the comment section seem to about technicalities. Your explanation was far better then what I've seem other people (or large marketing departments) peddle around the internet as fact. I didn't even think about how DSD relates to SACD and I'm sure that would be for a future video. I can't wait for the next video!
Came here to say this!
How on earth did you encode a 1khz sine wave onto a Vinyl record!?
It could have been a disk with various tones and combinations sold for calibrating and servicing audio equipment.
I love how methodical your videos are, keep it up!!!
Always entertaining, always educational, always the top of my 'to view' list when uploaded!
My cats both sat up and perked their ears when the wax cylinder music started playing. You know how cats sometimes do that at seemingly no stimulus? Well, my hypothesis is that ghosts sound like wax cylinder recordings of violin music.
“That’s not helpful.”
Lol now I wonder how many Commodore 64s would be needed to play the equivalent of one CD? 🤪
A CD contains 640MB, a C64... 64KB, so easily 10'000
@@LeopoldoGhielmetti I'd like to see someone try that! ...hopefully we have at least 10k C64s lying around...
Nice video man, audio and video sync was nice to see. Keep up the good work
I appreciate the 'time' it took to sync-up the VU meters with your voice during this video, well done! Also the _'DAC' - 'ADC'_ line broke me up! Your subtle humor and inflection in your videos always floor me, hilarious comedic timing sir!
13:00 Jimmy Fontanez - Floaters?
Yes, you can name that tune in 64 kilobytes!
I was thinking more Going Home by Dire Straits
None of these things are the real reason why digital won over analog. Analog can not recover errors, digital can. In a long daisy chain of components, errors simply add in the analog world, whereas in digital world each component is able to recover the exact signal, thus never accumulating them.
Over time, audio equipment became more and more complex, and analog devices struggled more and more to counteract the inevitable errors each subcomponent introduced. Only with digital processing is it possible to daisy chain things without destroying the signal on the way.
*Did you notice that this series isn't over yet?*
Technology Connections I did not! I apologise if you are planning to address this in a later installation.
Apology accepted! And please accept my apologies for the snark; it's a bit of a pet peeve of mine when people assume a thought has been overlooked, when in fact it simply hasn't yet been mentioned. Though I can't help stop myself in refuting that I vaguely hinted at your comment's substance in the video when I discussed how digital sound is "absolute" rather than dependent on the physical properties of everything involved. I just haven't yet elucidated there.
rumborak Have you ever tried to play a scratched CD? Digital can not recover all errors. Also digital errors are often far more noticeable than analog errors.
Well, any medium, sufficiently damaged, is unable to play back. A badly scratched vinyl record is unplayable as well.
I hate that I only recently came across your channel a year or so agp. Accurate and informative. 2 things I like, as well as being somewhat quirky... Thumbs up. Fantastic content.
What an enjoyable channel. Your videos are wonderfully entertaining and informative! Thank you!
0:15 Whoa there, I don't understand all these complicated words!
Heary bits - see, your ears can understand binary.
Michael Stevens no, it's the bits that hear as the bits are "heary"
Vinyl Worshipers will hate you XD
(they cannot simply accept the rational reasoning
on the principles of quality, dynamic range and the physical world)
On the other hand, analog distortion is "sweet" to the ears and listening to a good recording on good vinyl played on a good stereo is a very pleasurable thing to do. Maybe it's less to do with not accepting "rational reasoning" etc. and more about "I like the sound and the listening experience of vinyl". I like listening to albums and I like having to turn them over.
Not everything has to be tribal or political.
It's fun to mentally divide groups into "rational" and "irrational" -but it's not conducive to understanding the world as a whole and it doesn't help your short game or increase chances with your intended partner of choice. -just a thought.
Who divides what? Have you oversimplified my statement and tried to put me in the box of tribalism? I have written about fanatics, vinyl fundamentalists, a small segment of the mankind which very frequently comes up with superstitious reasoning. A segment of people and not a black and white situation.
And my sweet genius here is my lifestyle tip for you: don't play wise after 3 lines of text and try to tell how others perceive existence (it was absolutely irrelevant after all in this context), and it will probably improves your social life (lesser chance to seem to be a moron). Have a nice day!
THAT was a good chuckle. Thank you. You made a sweeping generalization and I inferred from your statement that you, like so many of us today like to see things in digital black/white terms and not analog "somewhere between here and there, close enough for jazz" terms. It's a common enough affliction, and if it doesn't bother you, who am I to point it out? -i apologize for rustling your jimmies.
A single statement can tell us a lot about a person, but I seem to have touched a nerve with my inference. Rather than attempting to correct my (possibly incorrect, at this rate we may never know) understanding, you went straight for the personal insult. Well played, internet stranger.
Ok wiseman, who is also a stranger and be happy, and generalize me as well.
I am a vinyl worshipper, but I totally agree with you. Vinyl is totally flawed, but I still like the sound it produces, the harmonic distortion makes it sound warmer, but is nothing a properly created filter can’t replicate digitally.
As usual, a great video... I really like your channel! Keep up the good work!
This series is going to be dank. Keep up the good work, proud of you
I always described the two methods thus: Analogue sound is imperfect playback (because of dust, wear, etc.) of a (theoretically) perfect recording, whereas digital sound is perfect playback of an imperfect (because it is, even at super-high resolution, an approximation) recording.
Except all recordings start out as analogue, in that they record analogue sound waves, so the idea that digital is somehow more perfect than analogue (the original sound vibrations), is like saying that a digital scan of the mona lisa is more perfect than the original.
Not true, any sound recorded digitally at a sample rate above the nyquist frequency is a perfect representation as long as it's played back correctly and is 100% above the noise floor.
C. H. Exactly right. A misconception about sampling is that it discards information, but it doesn’t. Search for xiphmonty’s video on digital audio to see why.
Please pin this down! Also, please, explain that on your next video. Pretty please!
C. H. ...and as long as the source has no information above the nyquist frequency! Getting these frequencies out of the source signal without somehow changing the lower frequencies is the tricky part.
I have just bucked up on your videos... and I have to say... your info is first class... and it seen that you have got you're 'groove' with regard to presentation... top work fella...
Fantastic video. You really start to appreciate the technology, even more than 100 year old tech, when you break it down. I kinda feel bad that lots of people take these things for granted.
Fantastic and clear explanation. Thank you. Really enjoyed it, and will be sharing it with quite a few people.
Your videos have been increasing their quality by every metric in a astonishingly rapid manner since I was a subscriber pretty much at day 1. This video however, represents a light-year leap. Your videos have always been fascinating by delving more deeply into esoteric technologies than anyone mainstream, and present them in a wonderfully comprehensible manner. But now you have crossed the line into geninley entertaining content! The 64KB audio clip had me doubled over! I'll definitely become a patron member. I want you to make as many videos like this as possible, and cover every topic until all viewers obtain a godlike understanding of the universe.
I really enjoy your explanations. I'm already a geek and had a grasp on how this process works, but I still enjoy hearing from you. Thanks
All of your videos are superb. Thank you.
Brilliant explanation of PCM! I love your channel, everything very well laid out.
What a nice video! Every detail is so cool.
Well done! I've been using digital audio for over 30 years and it never even occurred to look into what the various bitrates meant other than "bigger is more samples and better quality" I assumed the reason would be over my head, but this makes so much sense :) Thank you!
Some of your best content to date, earned a like.
I have tried to understand this for so long and this video explained it amazingly easy.
Man, your videos are amazing !
Congratulations for your channel. Your videos are really accurate and explicative. I await your next video explaining band-limited waves quantization and Nyquist theorem.
Possibly the best explanation of a complex issue that I have ever seen. The graph paper charts were key in illustrating the two dimensions that are captured. Well done.
Excellent video I somehow missed before. Also, love the analog and digital backdrop in the shelves.
Please accept my thanks for such thorough, clear and well-explained videos, for you take into consideration all means of understanding: visual, auditory and logical. Forever am I subscribed. Warmest regards from Peru.
This channel keeps getting better and better.
Great, clear description, as usual. Thanks.
Great walkthrough, dude! Can't wait for the entire history of the compact disc! 😃
Fantastic video. Looking forward to the inevitable video that bridges your older analog tv video to digital TV. (Might have been part of the pixels series)
You have a true gift for analogies and visual aids to inform and educate.... and of coarse entertain.
Very educational video, you taught me how digital audio works! Way simpler than it seems like it could be. Thanks so much, I'm now proudly a Patreon of yours.
From another source, VOIP is sending 333 MP3s per second, each one lasting only 300 ms. Each one is just a simple recording of this, and the delay is mostly in how long it takes to FTP the MP3 to the other end.
This is great, dude. Quality work.
Superb stuff. Thanks for yet another informative and entertaining video.
I just recently watched the vids leading up to this one, so this is great timing for me!
I Really enjoy your channel and your Great content! Thank you!!
Love your videos, their quality, every piece of knowledge inside them. And those details, VU meters on the recorder you were filming corresponding to your voice, jesus!
Seeing a new upload from you makes my day better
Wow, I never knew that's how this stuff works, pretty in depth and all. Thanks for the video!
I have been watching your videos for a long time now. You seem to grow on me every time I watch your video. Also I like your sense of humor. No doubt you are very smart and clever guy but you also admit your mistakes and fight for what is correct and what is wrong. You are the kind of guy other people should be listening to. Well done.
Love how you always insert funny bits into your videos :)