Audio File 01: Bit Depth & The 24 Bit Audio Myth

24 bit audio for music files is pointless. Here's why!
● MORE INFO:

"Digital Show & Tell" is distributed under a Creative Commons Attribution-ShareAlike (BY-SA) license.

The following images used in this video are distributed under a Creative Commons Attribution-ShareAlike (BY-SA) license.

"Radio-frequency-anechoic-chamber-HDR-0a"

"Engineer at audio console at Danish Broadcasting Corporation"

"Ipod 5th Generation white rotated"

The music used in this video is "Laserdisc" from Chris Zabriskie
"Laserdisc" is distributed under a Creative Commons Attribution (CC-BY) license.

This video itself is distributed under a Creative Commons Attribution-ShareAlike (BY-SA) license.

If you use this video, please attribute back to this Youtube page.

////////////////////////////////////////////////////////////////////////////////////////////
Click the like button and subscribe if you this video helpful!
////////////////////////////////////////////////////////////////////////////////////////////

Notice of Fair Use:

The graphics and visual elements are used for educational and instructive purposes.

They are of sufficient resolution for identification but lower resolution than the original graphics. Copies made from them will be of inferior quality, unsuitable as artwork on pirate versions or other uses that would compete with the commercial purpose of the original artwork.

[tjgopi]

24 bit audio is meant for production purposes as you are likely to have more than a 100 tracks running simultaneously and the noise will add up to real noise. also the dynamic range does matter in these environments. this really does not apply to the end listener. there what really matters is the quality of the DA converter. that decides the quality of sound you hear.

[RamonSmits]

24 bits is not useful for the end-user that is for sure but for recording 24b makes sense due to the much larger noise floor and precision so that these recorded sounds still have enough precision after applying EQ, high/low pass filters, compression, etc. All such 'effects' are basically conversions that result in quality loss.

In the end all such tracks are mixed and a 44.1kHz/16b is really enough. By the way, compression tools like ogg and mp3 usually DO result in better encoded files when the source was in a better format then 44.1/16.

Enough studies have proven that recording in 44.1kHz/24b is enough and also the minimum for applying effects unless you are going to stretch audio as then a higher frequency sample could benefit when played at a lower frequency.

By the way... within the recording community there are enough people that claim that they hear the difference between a unprocessed 96kHz/24b and 44.1kHz/16b recording which is complete bullocks. Most manufacturers of audio interfaces also use such qualifications :-). Well, maybe an ear trained 15 year old can indeed hear a difference in an perfect acoustic reflection free environment but otherwise... 

[colinm]

Man, I was all pleased today, as I finally got my head round Nyquist–Shannon, DACs, PCM, file formats, and after so many years of it just being a simple, but common, misuderstanding of DACs. But wow, Ive got so long to go to get to level of some of the guys posting below. You must know your stuff when you can hear things that dont generate sound, and can also modify tone by changing the mater volume setting.

Sadly, when I once thought I was hearing bit-depth, and it was affecting tonal quality, I was most disappointed when further research eventually tracked it to an obscure encoder anomaly from the week before, rather than having developed the ability to hear sound that does not exist

Ive also been able to realise that "warmth", "feel", emotive reasoning for vinyl being fundametally better than cd, can now much more likely be attributed to the fundamentaly different frequency response of vinyl and CD, vinyl being log and the other linear, the differences in dymanic range, and now realising that so much vinyl is "better" because the CD will have had a slapdash master. If that was fixed, warmth would be getting calle distortion

[lachlanlikesathing]

dunno why it should be all that serious. Btw, I know how the internet uses that as a generalised slur but I'd like to keep the comment section a safe place for people. I've had to remove your comment.

[Headphoneaddict]

Applause for great initiative! I sure learned a thing or ten! :D

[noturnleftunstoned]

16 / 48 flac with a good starter high-end DAP like Fiio x1 with a decent amp and AT50 headphones is incredibly good, honest and vibrant sound all for around $200.

[Thejgomez]

Bravo sir! Would be also good if you can demystify vinyl, so I don't have hear people saying over and over again how accurate that format is, when in fact even MP3 320kbps does a better job!

[leogruijters]

I am 62 years old and always work(ed) with electronic and computers and in addition I am an audiophile for most of my life.
When I read something like this, I always think back of the time when I was (much) younger.
I worked as a repair technician in an audio/TV shop.
We also sold amplifiers, but with valves(tubes), not transistors. They were bulky, inefficient and typically produced about 10 Watts max. at 1% THD (Total Harmonic Distortion). Then Philips announced the first all transistor (stereo) amplifier. It would produce about 2 x 15W at 0.2% THD.
Wow, that sounded great...
When we finally received the first unit, we could not wait to test it.
Then, total disappointment, it sounded like total crap...
Later, others found out why. They discovered nasty things like crossover distortion which sounds terrible, even in very low percentages, while the valve amplifier produces mainly pure harmonic distortion which 'colors' the sound but, for some, even sounds pleasing...
The lesson is: don't rely on a limited number of parameters to judge something as complex as sound...
My advice: If you are totally happy with MP3 quality, just enjoy and save a lot of money.
If you want to have the best possible sound quality, use your money wisely on things that really matter and not on a lot of 'hot air', I am sad to say, is abundantly present in the audiophile world...

[markfischer]

Technically correct, there is no difference possible in perception between 16 and 24 bits.

10 years ago I bought several Toshiba DVD players with 24 bit 192 khz audio DACs. Price per player $30. Performance, flawless. Extra 8 bits, superfluous.

[geoffgmartin]

I'm sorry - but there are a number of fundamental flaws in this lecture that need correction. For example, you are confused between the difference between quantisation error and dither (the noise floors that are generated by your 4-bit and 8-bit reductions of the piano recording are dither, not quantisation error). You are also confused between dither and noise-shaped dither (which you do not mention, but "noise-shaped dither" is what you mean when you say "dither" in the video).

In addition, your main conclusion is flawed, since you are assuming that a signal to noise ratio is enough of a description of an audio signal's quality to determine a required dynamic range. This would only be true if both the signal and the noise were steady-state and had identical spectra. Neither of these conditions is true in the case of music with quantisation noise or dither - whether noise-shaped or not.

Finally, your lecture, like the xiph.org posting, makes a number of assumptions about the playback conditions (both in terms of hardware and software) that are not necessarily true in all cases. As a result, once again, the basic statement that audio recordings distributed with a bit depth of 24 bits is a waste of space for all persons is incorrect.

Of course, there are many examples of equipment and playback conditions where a 16-bit correctly-dithered audio signal is adequate. These may, in fact, be the majority of cases. However, it is easy to think of a number of examples of a consumer playback situation where the difference between a 24-bit dithered signal and a 16-bit dithered signal will be audible.

Cheers
-geoff

[jeremyclayton-travis]

Thank you that was a very informative piece of information. I sent most of my life from the late 1960's through to the 1980's as a music enthusiast and manager of a Hi-Fi shop in London. It's good to put theoretical parameters into perspective. 
I recently brought a FLAC player and was bitterly disappointed but perhaps that's because I am getting on in years and my hearing is not what it was. 
I currently use an iPhone 5s 64 gig and have loaded it was 16 bit Wav files which fulfils my needs adequately. 
When the original Sony Walkman was released as a major Sony dealer I was invited to the launch at the Hilton in Park lane London where I met the head of Sony. 
The original walkman was a cassette based system and despite the tape his and limited dynamic range I enjoyed it for listening to cassette tapes I had recorded on high quality cassette decks to play back on the walkman ~ which I still own today with most of my other Hi-Fi equipment.

[marcomark]

Sir you are Brilliant, Funny with comments and illustrations and an outstanding teacher!!!! A true asset to the YouTube community. Thank you for your work and teaching me in a very comprehensive manor, EXACTLY what I needed to know. May you have a day as wonderful as you!

[JoryRFerrell]

Basically in a nutshell:
Bit Depth is like the color depth of an image. You can have a image, 256px X 256px low, in detail, but with a high range of colors (65, 000 different colors instead of just 256), and people are mistaking the bit depth for the resolution of the image.

[RealHomeRecording]

Great video!

If you want better quality audio, buy better speakers. After you get nice speakers, upgrade your digital to analog converter and amplifier (assuming your speakers are passive). There really isn't much beyond that.

[doublefivetwo]

Ok YT, there are 4K TVs out there but I'm still on the fence. When I got my Onkyo my friends loved to remind me that (at the time), few movies had a 7.1 codec. Now we can buy 4K TVs. How many Blu Ray are 4K. Those of us that love audio more than video have always had to wait. The market responds quickly to visual demands, but is slow to higher audio quality (for 3 reasons). If it were my world all music would be RECORDED to DSD. Then shit out to 24b minimum. The only way RECORDING standards will ever rise to the level of video is through consumer demand. We're the few that are passionate about audio. After all, one must cooperate with audio. Phase, position, and listener/observer tastes are so varied. Not so with video. All one need do is look. Gabba Gabba Hey!

[shaolin]

Audiophiles=human being that listens through his eyes

Yes audiophiles are special human beings that are able to hear things no other human being can hear as long as they know what they are listening to. When a Blind test is done, suddenly those HUGE, easy to tell differences they heard before, disappear....gotta  love snake oil. ;)

[PeterKese]

CAREFUL: "recordings these days only have 20 db of dynamic range" -- while that is indeed TRUE, it certainly doesn't mean that there are only 20 db of music in there (otherwise try listening the music encoded in 24 db or 4 bits -- which is what you just proved that doesn't sound well).
20 db dynamic range just means that there's that much 'volume level' difference between loudest and quietest parts of the recording. But there is certainly more than 20 db of musical information in there, otherwise you wouldn't hear anything in the more quiet parts.
Otherwise I totally agree that 16 bits is enough. Thanks for the video.

[alparsandor]

Actually, as more and more people convinced that 24bit or 32bit is something they should buy will increase the sold volume of such equipment making price lower for us who are relying such technologies in music production. :)

[syn]

The audiophile pooh-pooh industry will never focus on things that actually make a difference. Audiosnob woo-woo is far more profitable and is just as ridiculous as homeopathy and chiropractic. When you work with music reproduction objectively, the insane boutique prices become inapplicable because you're working with tried-and-tested cheap components.

It's why I shake my head and sigh when people turn their nose up at the NJM4556 opamp and spend thirty times the price on something that isn't even designed for audio. When the 4556 was not only specifically designed for audio, but specifically designed for _headphones_. And yet I really can't tell the difference between a stock O2 and any multi-kilobuck amp.

Gee, I wonder why.

[Nik__._.]

Excellent video. This is all correct. I can approve these facts. Very well done. Keep it up. (In Addition: As you said, 24bit is used for mastering or recording audio in the studio as well as a higher samplingrate "4:4:4, 4:2:2" for video cameras is the same - for recording video  - as this is a problem on the video sector sometimes, too.)