+Roy Unit I can explain what's going on, I work as an acoustical engineering consultant. 24-bit audio simply has a higher range of possible amplitude or level values than 16-bit, independent of frequency. This translates to a range of audible sounds of 144dB from the quietest sound to the loudest sound for 24-bit (which equates to roughly the maximum range of human hearing), 96dB range for 16-bit (CD quality), and roughly 50dB for a vinyl record. So with a 24-bit recording you could capture all the detail of a mouse sneezing at 1m followed by a jet taking off 50m away, while a 16-bit recording and a noise floor set at 0dB SPL (sound pressure level at the location of the microphone), would capture the mouse sneeze, but the jet would be all distorted and clipped (falsely being reproduced at only 96dB). Setting the noise floor high enough to accurately record the jet, say 140dB, means the quietest sound you could record would be 44dB SPL or approximately the ambient level of a quiet home. In terms of playing back audio recorded in the above example, pretend you had an amazing home stereo that could accurately playback the full 144dB range of 24-bit audio, in a perfectly sealed anechoic chamber. You calibrate the speaker output so that the levels you hear match the levels of the point of recording. You listen to the 24-bit recording, hear the mouse sneeze, then hear the jet take-off. Assuming you can still hear anything, you play back the first 16-bit recording - the mouse sneezes, then the jet takes off, but this time its only 96dB instead of 140dB, and sounds distorted due to clipping. You turn on the second 16-bit recording, and you hear the soft sound of noise similar to the level of your living room, no mouse sneeze audible, then the jet takes of at 140dB, deafening you again. With a vinyl record, the resulting playback is - with a noise floor of 0dB, the mouse is audible and at the right level, but the jet is only 50dB, and with a noise floor set to capture the jet at 140dB, the resulting background noise is at a whopping 90dB. What you are seeing in the audacity spectrogram is simply that noise floor of the 16-bit file is visible as dark purple, because the range of levels (color axis of the spectrogram) goes from -20dB gain to -120dB gain - a total range of 100dB, while 16-bit only contains information across 96dB of levels, or -20dB gain to -116dB, so the darkest possible color on that 16-bit spectrogram corresponds to -116dB on the color axis legend. As far as listening to the 16-bit vs 24-bit versions, in this case you are only hearing a maximum range of 50dB anyways, because it was recorded on vinyl. For music not piped directly into your brain via USB, you will never notice any difference between 16-bit and any higher bit depth. It's still best to record and mix in the highest bit-rate and sample rates possible, but the final output can be in 16-bit for any practical listening setup.

@chunkylover5367

6 жыл бұрын

I know this is an old comment but thank you for the clarification. It makes perfect sense.

@heartofjustice6041

6 жыл бұрын

this is indeed excellent i copied and pasted this for future reference

@ziggydeath9397

6 жыл бұрын

Finally, an audio engineer speaks about the realities of 16bit vs 24bit! It is an amplitude-only measurement, not a quality measurement. It is for us audio engineers in the studio, and does nothing whatsoever after the final mix down happens. We can balance track levels with more resolution with higher bit depths. That's all. When you are using an already mastered track like the Paul McCartney track, you are starting with a false assumption about what 16bit and 24bit means in digital audio engineering.

@kevintomb

6 жыл бұрын

Partially true, except nothing is ever recorded much below about -50 to -60 db. Add in dither and CD goes far beyond 96db which is still more than needed for any recorded material ever made. You forget the noise level of a recording limits the dynamic range, not the noise level of 16 bit or 24 bit. CD is never the limiting factor, recordings are.

@yueying7838

6 жыл бұрын

SovereignMan85 that's good if I'm listening to mice tapdance behind a jet take off

Personally I have found a good master will sound excellent even as an MP3 file. A terrible master still sounds terrible at 24/96.

@trophywolfe

3 жыл бұрын

But a good 24/96 master will sound all the more better than if it were still mastered good for mp3

@pracheerdeka6737

3 жыл бұрын

96 khz clocking captures some blank samples in audio but less timing error.

@pracheerdeka6737

3 жыл бұрын

24 bit is good of you recording less COMPRESSION audio on digital.

@gayusschwulius8490

3 жыл бұрын

This. Good mastering is far more important than high resolution. For most recordings, most people will not hear a difference between a 96KHz 24bit recording and a CD-quality recording unless they are in a completely quiet environment with 500 $ headphones and a 500 $ DAC - something which is really the exception, not the norm. They will, however, even on a crappy phone or iPod with 20 $ earphones, always hear the difference between a good and a bad master. People who think that they like hires recordings usually only like the way those recordings are remastered compared to the awful loudness-wars CD masterings. But that has nothing to do with the resolution of the format. Really, the main benefit I see in hires-recording lies in the recording itself. You get a wider dynamic range to play with in mastering. But as a delivery format for the end user? I think the use cases are very, very limited there. Maybe 1% of all music listeners can actually benefit from that. Not that I'm opposed to it - it's always great when something is released in close to original quality by the artist/label. But the hype really isn't justified.

@gameblogua5274

2 жыл бұрын

)))

The 24 bits are important only when you process the sound because you have more data to work with, just like in photo editing

Dude. This is so cool that you actually discovered this completely on your own just by curiosity. You researched it independently and then confirmed your observations. Really cool.

Nice video. A couple of thoughts for consideration: 1- as producers we spend countless hours trying to track perfect sessions only to then (if mixing in the box) use tape delay, saturation and countless other processes to emulate a vintage or analog flavor. We take our perfection and add "noise"...... Lol 2- What bit depth would vinyl be equivalent to? Reel to reel tape? 3- The final test is always, how does it sound? If you can't hear a difference on a bose wave radio, on a mono speaker or a pair of headphones, my opinion is that it is a non issue. 4th and final opinion: Mastering engineers like to receive files in 24bit so, that's what I give them. Other than that, it makes no difference to me. I think the people who buy records (or iTunes dowloads) but them because they like the music, get a good feeling from it and listen on far less advanced equipment than we use a producers. In most cases they don't even know the words to the song. They certainly arent hearing the things we are trained to hear. So I say, do what sounds good and feels good. That's most important. Just my two penny's worth. I did enjoy the video. Thanks for helping me to confirm my thoughts on the subject.

Thank you Roy, I heard the difference and I'm being grateful for your sampling work

I noticed that the 16 bit sounded more 'wider' in frequences and better to listen to. Interesting

@sleightofmind2016

Жыл бұрын

Same here....

Great video! This is one of the most interesting videos that I have seen on KZread! Thank you very much.

Thank you for your effort in A-B testing bitrates across variables. I valued how well you explained your instrumentation and methodology. I also learned a great deal from the commentators. With appreciation, I Ro

Great video, alot of work behind this.Thanks!

Thank you for the trumendous work you put into this research! I don't know how is this with others, but I could hear the difference in your blind test, and guessed right on the versions. 16 bit audio sounds "more rough", less detailed, more simplyfied to me.

@GodsMistake

5 жыл бұрын

I can easily tell the difference but there's no convincing the people who can't, they just won't 'listen'.

To mask quantisation errors dithering is applied. Dithering is essentially adding noise, and the higher the bit depth the lower the amplitude of the dithering noise. So most likely, that noise you see is just dithering. If an audio clip is played on 24 bit doesn't mean the dynamic range of it is going to be 144 dBs, it means it can reach that dynamic range if the clip's loudest point is on 0 dB and the quietest point is at -144 dB, but if a song has a dynamic range of 13 dB, you won't find any difference in the dynamic range when comparing a 16 bit version to a 24 bit version of it. So in digital audio, a highest dynamic range doesn't mean you can go louder, it means you can go quieter (except for the floating point modes). Oh and, by the way, the RMS level is not your signal-to-noise ratio, RMS stands for root-mean-square and, to quickly sum up, is just an average level, as opposed to the Peak level, which tells you the loudest level reached. To understand this a bit better I recommend you to take a look at "Principles of Digital Audio" by Ken Pohlmann :) I hope I was of any help to you!

@dazr6604

6 жыл бұрын

I was screaming dithering all the way through this! I bet noise shaping is applied during the 16bit FLAC export but no options in the UI like you'd get with SoundForge etc.

@jasondoe2596

6 жыл бұрын

Daz M, haha, same here! It was immediately obvious to me that the maker of the video discovered what dithering looks like. Audacity probably applies it during downsampling without even asking. And I'm not even a sound engineer; just an amateur photographer. But the basics of signal processing are universal :)

@peterveer7798

5 жыл бұрын

What for speakers is on 10:16 ?

@chadfranklin47

4 жыл бұрын

It is definitely dithering. Audacity doesn't ask at export, but it is in the preferences.

@alexander1989x

2 ай бұрын

Exactly this. Dithering is something you do in low-bit depth audio to reduce quantization errors. Esentially is a tradeoff between high fidelity reproduction of analog-to-digital-to-analog audio and noise you can't perceive.

thank you my fellow audiophile, this couldn't be done any better. i think i covered all of my concerns. now i can continue to save gb's and gb's of space!

Awesome video. Thank you!

You may need 24-bit during edition to get 16-bit dynamics in the final result, like you may need RAW images because you get an over/underexposed picture, but a perfectly exposed (or corrected) image can be 8-bit JPG.

This is was a great watch, informative and insightful.

What you see on the spectrogram is either the dither noise used by default before truncation process or the quantization noise if the dither wasn't used. So if you recorded the data directly from analog into 16 bits and 24 bits resolution the spectrograms would look identical.

Awesome video. I was just thinking about making an A/B test between 24 bit and 16 bit. Now I don’t have to.

Thanks for your work.

when you convert from 24 to 16 you are adding dither to minimise quantisation error, this is why you have the high noise on the 16 bit, but that's not due to the 16 bit. if you convert without dither you would have some quantisation error but no high hiss noise added... noise shaping is used to alleviate this, 24 bit vs 16 bit is only dynamic range - 16 bit is well 96dB which is way over what a tape or vinyl used to have. 24 bit is 144dB but this is limited by your converter capacity (I don't think there are 144dB s/n converters, correct me if I am wrong)

@neowavemusic

5 жыл бұрын

yes 100% conversion with dithering, in theory they should not differ on the spectrogram

First off, i dug your post. Well executed. DL'd your sample zip. I easily noticed the separation, depth and clarity difference between both your flac samples. #1 all the way. #2 sounds "acceptable" and decent, however you can totally hear how it's not as spread out like #1 is. Of coarse just all depends on how big of a deal that is as well as what type of music one is listening too..plus obviously IF indeed it is "true" 24-bit or not. Cheers! :-)

Great channel! very entertaining! I was wondering if you have ever considered doing an episode on NEO GEO ports and all of the various iterations on all of the different consoles, as well as comparing and contrasting them against one another?

Nice evaluation btw. The question is what source material would best reveal the difference if we could hear it. I would try complex orchestral music. Reason is all those different instruments have their own wave shapes (deviations from pure sinus) that make them recognizable. Now add those all up in an orchestra and listen if you can distinguish the viola from the violin, the clarinet from the saxophone, and oboe from English horn in a tutti fragment, plus can hear where they are in the soundstage.

Thank you, very informative and helpful.

Thanks so much Roy! Super work! after watching this I think that what matters is if the music was recorded at 16 or 24 bit. I am checking this because I record music and I have 2 Tascam DAT recorders that I use for converting analog sound from my console with oscillators filters synths and all analog sound. So I have 1 converter that converts up to 24bit and the other one to 16bit. So I will do a test with both of them recording just ground floor to see if i can tell the difference and I will let you know. Thanks for sharing!

Thanks for uploading. Def learned from it.

As soon as I was told which one was the higher bit rate I could tell, despite there not being a difference. Magic.

@MacXpert74

6 жыл бұрын

Placebo is a magical thing :D

@daftmoonz9395

6 жыл бұрын

It's like going to a restaurant for the rich being served a meal on a gold plated crystal plate and you say umm that's better food than the other place I go to. The Chef comes out and tells you that other place cooked it. It adds to the experience having nice looking cables and music files in the same format as the studio recording. Some things benefit the sound some things are based on it and used to exploit for cash.

@craZivn

5 жыл бұрын

I could swear there was a difference, and I called it correctly immediately upon hearing the second sample start. But, to be fair, I had a 50 percent chance of getting it right anyway so there's that.

Great job on this video :) interesting and very well explained. when i get a chance I'm going to run some tests with my audio interfaces as well. keep up the great vids Cheers!

@RoyUnit

8 жыл бұрын

+illBeatz Thanks a lot! I'd be interested in seeing your results. Glad you liked the video.

So, my small amount of experience was that if you turn up the volume, the quiet parts stayed relatively quiet, and the loud parts got a lot louder. It's like an HDR photo, it isn't any clearer, it's just that the bright parts are brighter and the darks are darker.

@Strepite

6 ай бұрын

Nope… that’s just your brain telling you because you want to believe… Nyquist Theory, no way you can hear the difference because your ears are not designed for that

Thank you for your effort, very helpful!

Brilliant.... to a layman that just loves music and was trying to understand if I needed to sample my music at 16bit FLAC or 24bit FLAC. You answered my questions in full...... Cheers mate and happy listening :)

i bought album online because i couldn't wait for it to get in stores and the first thing i noticed when i eventually bought the cd too was the difference from the cd version and the online version, in my quest to see why the online one sounded better i found the difference was the online one was 24 bit. so now i'm here to confirm if i am/was imagining things. because i actually expected the online to be equal or worse than the cd version hence the reason why i must always have cd. but now that i have noticed the difference i'm prob. not going to buy cd's anymore. i can totally hear the difference with i guess high end headphones.

Very good video and so educational

Good Job, thanks for this video, cheers

Isn't that noise at -110~-115db? That basicly means that it's below your fancy stereo's noise ratio right? So at a point where you'd amplified enough to expect to hear that, you'd instead hear the noise generated by your equipment instead. Am I right?

That was amazing thank you.

Very Informative, learned so much ^^

May I know how the 32 bit float master was obtained? I would tend to agree with Mycoolharmony. There are several aspects to the A-D process itself that can make the depth comparisons later null.

Thanks for a very interesting video. I've watched several such videos lately, and have come to the following conclusion. My personal opinion is that the combination of my middle aged ears, and my consumer grade stereo equipment makes the entire issue moot. There is no way I can hear any audio quality difference between vinyl, CD, or a native 24 bit audio file. Therefore the CD is still my format of choice. Vinyl is fun, and interesting, but takes up too much storage space, and is a pain to convert to digital for modern use such as for digital playback in the car. Digital downloads are convenient, but require that I maintain them digitally, and we all know computer hard drives can fail. This leaves me with having to rely on the vendor I purchased the music from as a backup. CDs are great because I can easily digitize them, and the CD itself is its own backup.

Ty for this great video! :)

When I got the feeling of this is a more natural sound experience specially when recordings of real instruments I usually got the 24 bit right choice...With synth is getting more difficult

The noise near the top end is most likely Noise Shaped dither added by your audio editor. The program tries to put the noise where it is least likely to be audible, and scoop it out of the midrange. Monty Montgomery describes the process later. The noise level inherent to the vinyl medium completely overwhelms dither noise at almost all frequencies, and any extra resolution in the midrange won't be observed. Some programs default to highly shaped dither, but one can usually chose a milder curve in the settings. It is difficult to tell from the images because the blue detail has been washed away by KZread chroma subsampling. I would zoom in another 10 db and closer in time, and use the warm color scheme that SoX or Izotope has, which is easier to look at and transmit over video.

On 16 bit releases, the dither is usually noise shaped, i.e. it is applied to the high frequencies only. That more than likely explains the 16 vs '24 bit difference at the 4 min mark here.

I did the test 30 times, I opened both of the FLAC files in VLC media player and chose shuffle. I did it completely blind and tallied how many times I guessed right. I ended up getting 21/30 right, which is pretty curious. I felt more confident of my choices in the beginning and got the first 5 correct before making the first mistake. As time went on, I found it harder to tell, possibly just due to fatigue. Unless I just got lucky, I suspect there is an audible difference.

Roy, this is an excellent video. I show this to all my friends insisting recording at the highest sampling rate and bit depth is the way to go. I challenge them all them time to tell me which songs are 44/16 and 44/24. They mever get it right. The ears jist aren't that good enough to hear the noise at those high frequencies. It ain't gonna happen

You were very helpful.

The noise you're seeing added to the high-end of the 16-bit files is "dithering". This is an over-simplified explanation, but it's basically high-frequency hiss added to the track to mask any small errors the software makes when resampling down to 16-bit.

Gran análisis!!

I think it matters more what bitrate the songs were originally exported. It seems pretty counterintuitive that a song originally exported in 16 bits for example, could be increased in quality because converting it to 24 bits would be kinda of just like making a recording of a recording, you can’t make it sound ‘better’ or at least it won’t be 24 bit until you re master it at 24 bits.

Thank you SO MUCH...

The noise you were referring to was what is called dithering. In a highly simplified explanation dithering is the intentional introduction of low level shaped noise to cover up any quantization errors that sound like distortion. The reason you see it at those frequencies is that for most adults they are almost inaudible.

@axilleas

3 жыл бұрын

@ReaktorLeak correct, I was oversimplifying intentionally. Digital audio is hard for those of us who work with it, let alone for people who don’t have to.

i can differentiate easily the example sound you gave in the video using my slightly worn out headphone, for me 16 bit feels more firm, and the sound produced by instruments feels more contrast to each other than 24 bit. while 24 bit feel more smoother and well mixed to each other. personally i would prefer 16 bit.

Excellent in-depth! Btw, anyone noticed the song 'Tug of War' in the list had a duration of 420?

There are two parts to an audio file. There is sample rate, which is how many samples can be taken during a given time, and there is bit rate, which is the range of samples that can be taken within a given amplitude of the audio signal. For instance, if I have a frequency range of +/- 100Khz, and I have a bit rate of 4, that means that any time I take a sample, it can only be one of 15 different values at steps of 13.333Khz. If the sound I am sampling doesn't fit perfectly at any one of those points, it is considered either to be at the next higher step or the next lower step. At lower frequencies, we notice less difference than at higher frequencies. This is because the sample rate is a larger ratio of frequency than at higher frequencies. When you hear high frequencies, the averaging that happens works as a muffler, as there are more samples that come out in a shorter period of time that are off from their original frequency than when you are hearing lower frequency sounds. Now, this is not a problem for many people for a couple reasons. First is that every person has a different range of frequencies they can hear, so they often don't hear the highest frequencies in the first place. Second is that most people have no experience trying to distinguish one sound from another.

well i was only looking for a 24-bit version of for the damaged coda and i ended up here. good chat.

The difference between 16-bit and 24-bit is the dynamic range. So when comparing the two, don't use an audio source that has a smaller dynamic range than either of them (i.e., a vinyl record)!

What setup did you use to transfer your vinyl to digital? I've tried multiple pieces of software, USB audio cables and even tried using a TASCAM field recorder - I always end up with a lot of hum and the quality is low.

Hi, i have a question: i want to download all my CD's on to my computer and in the past I've used windows media player. Problem is that it doesn't keep the artwork properly and ive lost a few tracks. Also it doesnt give me many options i might need ( like recording levels for old cd's that were recoded very softly... or burn quality like your talking about ( 16 or 24 bit/ upscaling). Can you ( or anybody else) help me in recommending good software? I want to be able to keep it on a separate hard drive so i can use it on any computer. Any price point is good.... thanks in advance and very interesting vid.

@titmusspaultpaul5

6 жыл бұрын

And btw, i was listening through my Bose heaphones and i could pick the 24 bit straight away. It was more punchy and vibrant.

ThankU for this easy explanation. I bought Player, DAC, Over-Ear Head Set Unit & Audio / Songs that tagged HiRes ~ listening for about 2 months, yet I still can hear the different and enjoy the HiRes Investment :-) But al the device look's cool anyway..., yet I need put the 'break' for a while

Never mix bits. Start with 24 bit end with 24 bit or start with 16 bit end with 16 bit if not you start to here loudness in certain parts of your drops. Sloppy channeling can cause this to, but long story short what ever you export in digital or analog 16 or 24 bit master it in the same format. make sure your audio box have the same connection in each program you use.

You will probably only see the difference with subtleties, like lower and more subtle sounds, that's where it could be indistinguishable (part of the noise so completely lost) if the depth is low enough or it you can hear it as clearly as you should if the depth is high enough

how can you know a lp has enough resolution for you to be able to hear a difference? or your speaker system is good enough?

nice video, nice discussion, nice links! Re: 24 vs 16 filesize. Not best to say 'nearly twice as much', when you readily know you mean exactly 50 percent more, which you went on to express that anyway. Not a criticism, since you clarified it well, but of course your discussion deals directly in quantities, so a correction almost worth mentioning. I have a relevant question.. does the noise floor 'add up' when mixing (combining) tracks in a DAW. This might make is more important in that case to record at 24 bit.

great vid, u put a lot of effor into it, vape naysh

@Scottydj01

6 жыл бұрын

vape naysh ya'll haha

i can hear the difference EVERY time. the track selection is sublime :) the noise on top of the trumpet (or whatever that blowing instrument is :D) is so evident... i listened on lousy microlab speakers connected to my laptop powered from a USB jack using whatever player this windows 10 has. so no need to go through the hassle to turn this up on my hifi setup

@barbararabalalaala3123

2 жыл бұрын

the trickwith ABX is to take it slow, isolate sections, take a break, if you listen too much in a short period you will get used to it and focus on the musicality of the track and not the interpretation of the sound machines. by the way i am 38 years old and kicking, you know that MEDICAL S#!T that you lose your high frequencies part of hearing first...

Outstanding presentation, ROY UNIT! The money available in my checking account THANKS YOU!

One of the more honest attempts at answering the question even if you could not tell the difference...

A question to anyone who can help. my entire music collection is verified 320k mp3s. i want to have my portable player's library at 192k mp3s to have space for a lot and still have decent quality. will a 192k mp3 which is down encoded from a higher bitrate mp3, be of the same quality as one down encoded from a lossless format or one directly encoded to 192k from cd (in perception at least, i assume it would have less actual data). i'm mainly concerned about the artefact of "drowning of percussion" in low quality mp3s. flac is just overkill for me, so can i down convert my 320s with a peace of mind or just use them as is?

@markjacobs1086

5 жыл бұрын

Converting MP3 > MP3 is a bad idea, you effectively apply the same compression twice in the exact same spots in your audio, now if you can't tell which is which in an ABX test go for it, but generally quality will be degraded way too much then if you would transcode from 320kbps MP3 to 192kbps AAC.

I can't believe that you even have the Audacity to do this. 😉👌

@seth1455

3 жыл бұрын

Groan

Ever heard of the term "random noise dithering" or just "dithering"? It is usually added by default on 16bit encodings in audacity, but can be turned off in the settings And there you have an explanation for the added noise

Bit depth properly dithered purely effect the noise floor.

BitDepth is interesting for dynamic range. Any idea what the signal to noise ratio is of your analog listening system? For digital recording it gives extra playroom (stay away from the 0 dBFS) for unexpected peaks. The dynamic range of 24bit (144 dB) is theoretically. The component is our devices (even if they are switched off) make more noise then that... en.wikipedia.org/wiki/Johnson%E2%80%93Nyquist_noise p.s. What you 'see' in the high freq is dither of 16 bit. ;)

So much knowledge! Now let me use it in my audio experience and be a better person :D

So this is a very technical and scientific way of telling me to just stick with 16 bit. mixdown. Got it!

@dongerramarco9617

4 жыл бұрын

agree

@TD402dd

4 жыл бұрын

I disagree because the DAC builders are setting the stage for higher bit. If you buy one it will already be 24 bit, moving to 32 bit, and eventually 64 bit. You will have a hard time finding one that only works in the 16 bit range.

can flip one wave form and subtract them and then try to listen the difference if you can hear it

I was tempted to test my ears and see if there is actual difference 24 bit vs 16 bit music files and ended up listening Mozart in 24BIT/352.8kHz and in CD quality 16BIT/44kHz. There sure is. 16bit sounds ok but 24 bit is amazing. To start with when you change from one song to the other, it feels like the large concert hall suddenly becomes a large room and instruments lose their sound range. I am not an expert in sound, just had 4 years of piano and solfege as a teenager and i was singing for 12 years in school choruses, but the instrument differences where audible even at the 2 files you submitted in the description. Not that much of a deal some can say. But after listening to these files my life in audio will never be the same. I used Focal Sphear S for my listening session. Still kind of new (8 hour pink noise and about 30 hours of house music for the bass kick) Thank you for your video.

@QuarkPrince

6 жыл бұрын

Çerastes, i am sure i hear a difference. To me the 24 bit file was much better. Especially the highs, were much clearer.

@QuarkPrince

6 жыл бұрын

Çerastes www.2l.no/hires/index.html These were the audio files that i tested, not the ones on the youtube video. And i of course downloaded and listened to the two audio files in the description. KZread of course has no audio quality reliability. Have a look for yourself.

@echodelta9

6 жыл бұрын

My first thought on reading about this stuff in 1978 was not enough bits. I counted on my fingers bits till I was happy... 24! Nailed it. But I swallowed the Nyquist thing whole. The truth. The top third of the audio spectrum in a CD (2000-20,000Hz) is pixelated! 96 K samples per second or higher is more important than bit depth. It takes an exponential amount of bits to go up in frequency range and depth as well.

What is the maximum sample rate, bit depth, and bpc youtube allowes? Thanks in advance.

Could tell before being told which was which was which - certainly a difference in clarity and range

I gave a listen on the ballroom song for a couple if times. I noticed the difference on the trumpets. The 16 bit one is darker. It could be the high harmonics of it is less audible in relation to it's fundamental frequency.

2:50 "might" KZread encodes audio as 256 kbps mpeg.

@IAmNeomic

4 жыл бұрын

128 kbps actually.

@IAmNeomic

4 жыл бұрын

@nullvideo It depends on which playback device you use whether you get AAC or OPUS though. I'm not sure which gets which, but there is variation for whatever reason.

@cardioandfriends

4 жыл бұрын

@@IAmNeomic 192

@IAmNeomic

4 жыл бұрын

@@cardioandfriends There used to be 192, but once they added the OPUS codec, everything was brought to 128 for whatever reason. Possibly bandwidth, since that was about the same time they added the HEVC video codec.

@lukaslangrock7878

3 жыл бұрын

@@IAmNeomic but OPUS gets transparent at around 96 so it will sound very good at 128 (don't actually know if that the bitrate youtube uses). I have seen some videos where someone claimed to have uploaded 192kHz 32bit music to KZread and all the audiophiles were talking about how good lossless sounds on their fancy thousand dollar hifi but they were actually listening to low bitrate OPUS and didn't even notice hehe. As long as long as you don't think about the bitrate or the codecs KZread uses, the audio here will sound very good to you. Or of course, when you are trying to compare two lossless audio files, that is not gonna work. And a small correction, KZread uses VP9 and recently AV1 in addition to AVC, not HEVC. AVC is for older, low powered devices or just inactive videos and VP9 for pretty much anything else, then the very popular videos get AV1 as it saves an additional ~30% in bandwith but requires a hell lot of computing power for youtube to encode. the classic AVC+AAC combo is only used for livestreams as far as I know. You can actually see the current audio and video codecs on the video you are watching by right-clicking and selecting "Stats for nerds".

It is important to understand that Quantization error (QE) is nothing like random error. In fact it is incorrect to call QE noise. The waveform is converted back to analog by the ear which has mass and is thus an averaging device. QE does not distort the analog rendition in the way that random error would. So as long as the sampling rate is high enough the QE does not result in any additional perceived noise. Who can distinguish between 16x192 and 24x192?

One more outstanding video that brakes 16 vs 24 bit myth. Thanks again

The "noise" on 16 bits audio is propositional! When you down-rates the audio resolution from 24 to 16, the converter software will put some noise to reduce the distortion of the digital quantization, and so make the sound more natural to our hears. This noise is called dither.

@willb3698

6 жыл бұрын

Artur Bernardo Mallmann - Thank GOD someone finally starts talking about the key issue - especially on the down sampling of youtube. I was getting really fed up with the arrogance of people saying placebo etc. And the evident lack of knowledge that goes with it.

All a bit rate does is reduce quantized noise. So yes it improves sound quality but it's not that much of a difference between 16bit and 24bit

@MacXpert74

6 жыл бұрын

In reality you'd never hear the difference. There simply aren't any recordings around in any music genre that would use the full 96 db range of standard CD 16-bit sound. Having more bits available does nothing for the sound if the recording already has a higher noise floor than the 96 db. And this will ALWAYS be the case, even with the most critical classic recordings or whatever. With pop / rap / dance music etc. the dynamic range is particularly very limited and would probably sound fine with 12-bit or less ;)

What type of device do I need to create 8 bit audio or even lower if possible?

thanx for the trouble you have gone too.Think I kno now.thnx

I could here very clearly. At least the first time I listened to it. I immediately knew. But idk how long I would be able to tell or why I was able to tell, I’m on an iPhone 10S’s built in speakers as well, which I know is making the sound suffer. But I was able to tell the first 3 seconds

Hi...when you mentioned record...do you mean you recorded the audio file in real time ?. Did you input the source from audio line in of the computer?...please explain...thanks.

How much oversampling was there in the DAC? If there was none the antialiasing filter could explain the difference in noise.

after listening to the 2 audios I right away could tell the different. to my ears the 24 bit depth sounded fuller and richer than the 16 bit.

Take a 1khz sinusoidal wave, then set the amplitude to half way between two sampling levels, digitise at say 4khz sampling rate, then restore back to analogue (using a bandpass filter between 20hz and 20khz. At what point can you can stop hearing the difference between the original analogue wave and the sampled one as you increase the bit depth? The sampling levels are spaced in a log and not a linear scale. So more noise may be introduced at lower than at high levels. It would interesting to see again, difference this.

you need to record at the bit depth you want to test. converting down from 32 bit is still a 32 bit recording but with more noise from the grainier quantization.

The only time I have ever noticed a difference is when I was forced to downsample. I found that when I recorded and mastered in 16bit recording for making CDs, I had a lower level of noise than if I recorded in a higher bit depth and then down-sampled to 16 bits. There was something about the down-sampling that increased the noise level. This seems to be paradoxical to me and I have no explanation for it. But that is what I found. Needless to say, as long as I was recording for CDs I would always record in 16bit so that no down-sampling was required.

I’m young so I think I can hear the difference because I can still hear high frequencies. The 16 bit sounds harsher, like the jumps between notes are distinct while 24 sounds like a smooth transition without high tones all sounding the same.

Using a vinyl record as the source immediately invalidates your test. Get a native 24 bits track.

@youkounkoun2

6 жыл бұрын

of course...

@yueying7838

6 жыл бұрын

Tiberiu Nicolae lol but vinyls are the pure energy of sound. I can hear the skin falling of the guitarists fingers

@pilotavery

4 жыл бұрын

@@yueying7838 lol it's recorded from a digital mic on most anyway, and additionally, frequency response is limited. Vinyl is not a better sound or higher quality sound than digital, it just sounds nostalgic. Kind of how your 6cyl Civic has as much power and drives nicer and is safer and more comfortable than an old Pontiac firebird, but the firebird is the more fun classic car to drive. Same thing. Not better in any way other than... Just being awesome.

@bruhdamartinasty3636

4 жыл бұрын

Not necessarily. If he had put a microphone in front of the speakers of a stereo system playing vinyl, then that would've been fine. That's the same thing as putting a microphone in front of a singer or an acoustic guitar. Then he could've changed the bit rate in his recording system and made the comparison.

Perhaps downloading a 16bit tone and a 24 bit version of it could be easier? Or use audacity to generate these tones. Great info!

Unnecessary comparison as the source is poor. Vinyl has an average dynamic range of around 60-70dB, essentially offers about 11 bits worth of resolution. Sampling it at 16 or 24 bits will make no difference. That is why all the old digitization of LP's is nothing more than a marketing gimmick. Even analog studio master tapes using Dolby-A noise reduction can only have a dynamic range of around 80 dB. What is the use of capturing a low grade source at high sample rate?

@morskoyzmey

6 жыл бұрын

Noise on vynil records is not equal to noise of quantization errors.

@kevintomb

6 жыл бұрын

60-70 is not the average, but about the best it will usually be.

@nazcaplain

6 жыл бұрын

So obviously in this example 24 bit is overkill dynamic range for encoding a LP. However a higher Khz rate makes sense for older analog mastered LP's.

@KRAFTWERK2K6

6 жыл бұрын

@ Tejas: well you see… there's just one problem with your comparision. Vinyl is ANALOG. So it doesn't freaking matter if it COULD be 11 bits worth of resolution with it's dynamic range. The Audio is analog and behaves completely different than digital informations that need to pass the bottleneck called Encoding and Decoding. And since Analog is never 100% the same like digital bits, you gotta make sure the resolution of the recording is as fine as possible to be an somewhat exact digital representation of the analog audio. However yes, i would not use such a vinyl as a reference unless it is an MFSL release. Also I'm not even sure if modern re-releases of Vinyl albums, that you can get nowdays, are even 100% analog but use 48khz 16 BIT digital masters instead, that was used to create the vinyl groove for the mass production stampers.

@MarkTillotson

6 жыл бұрын

True vinyl noise is very different to quantization noise, but at 16-bit the main issue is that the vinyl noise is generally much louder, and vastly vastly louder at very low audio frequencies (rumble). Surface noise is less easy to ignore than Jonhson noise too. Thumbs down for vinyl, and we haven't even mentioned wow yet.

Its like trying yo pour a can of soda in a a empty gallon container. How much soda do you have? Not a gallon, still a can

How can you hear any difference when both are coming from an mp3 audio track on a youtube video?

good job!