Hi/High Resolution Audio EXPLAINED - Is it worth it? (192kHz/96kHz & 24-bit)

Hi/High res audio means music files recorded at 24-bit and higher than 44.1kHz sampling rates (often 48kHz, 96kHz and 192kHz). High definition audio promises "better than CD quality", and is being offered by companies like Apple and Sony. But can you notice the difference? Will your library of songs suddenly become more realistic and lifelike? And is the price that you pay (for devices like the expensive Pono Player) actually worth it?

It is true that the higher resolution has benefits over CD when it comes to dynamic range and extended frequency response. But the files take up far more space on your hard disk for a quality improvement that's all but negligible. In all honesty, the loudness war, recording techniques etc make a far bigger difference in the end result than high-res audio ever will.

====
Hey, I'm HandyAndy and thanks so much for watching my video! If you did enjoy it, then please subscribe to my channel, and make sure to hit that like button!
====

MORE INFO:
[1] Nothing to see here, that's just an example.

[5] Well, we MIGHT. The reason why 96k (and later 192k) became standard in the pro audio world was because it was difficult to build an anti-aliasing low-pass filter that would cut off at 22.05kHz without attentuating the signal any more than was necessary. But for a consumer format, the oversampling on modern DACs has kind of bypassed that requirement.

[6] Yes, they aren't necessary for a playback-only consumer format. But in the pro sound engineering field, I must admit that they are quite useful.

[SoundSpeeds]

I'm a professional sound guy that works in the motion picture industry and I can tell you... your explanation is pretty good but your conclusion is incomplete. You totally miss the purpose of HD (above 48K) audio. The main reason people use it is because they are doing some sort of high fidelity mastering where they could need to layer the sound. What if they need to slow down the sound to match slow motion video? If you slow down 48K, your sound will turn sucky really quickly. You need that higher fidelity sound to properly mix or master.

[les]

i feel like he is staring into my soul

[zahriatan]

My ears tell me different. I definitely can hear a difference, significant difference that adds to my listening pleasure. The proof of the pudding is in the eating and what you hear is what you get.

[heyguyslolGAMING]

I'm considering transitioning to a bat so I can take advantage of frequency's beyond 20-20K and I can virtue signal my new abilities on Twitter while simultaneously canceling anyone who attempts to disprove my new found abilities.

[chriscutress]

Interesting argument, but ... As a professional Audio Engineer I can tell you that 24 bit 96 KHz does make a difference to the trained ear especially for acoustic master recordings. There are vibrational frequencies above our hearing range that do effect the listener when listening through hi-end reproduction equipment. I admit that I can not personally hear a difference between 24-96 and 24-192. But I can hear a difference between 16-44.1 and 24-96 with properly recorded non-over compressed acoustic recording using hi-end microphones and listening equipment. Unfortunately most of the world is listening to mp3 over computer monitors so 16-44.1 is becoming the defacto hi-end format.

[TheChadPad]

3:51 The answer: Because frequencies higher than 20kHz combine to produce lower frequencies within our hearing range that add to the high end frequency timbre of music. This is why Rupert Neve made sure his test equipment was rated to have a linear response of up to 50kHz and a controlled response to 100

[Tunemedia]

Generally, I think you're right, when a song is ready to be broadcast (After being mastered), 16bit/44.1khz will pretty much sound similar to 24bit/96khz. But, at the producing stage, I believe it gives you more freedom to mix, shape, eq and control finer subtleties then lower resolutions. (This subject can lead to a social, philosophical argue, because people are able to enjoy good old lofi music as well..)
It's like, shooting an8k video and play it on a HD tv (which is basically 1k) it doesn't affect the top quality of the tv, but at the producers level, you just have more info to work with..

[reason]

As the owner of thousands of licensed Hi-Res FLAC and DSD titles, I can tell you first hand that most of them are remastered FOR Hi-Res. If I play the CD version of Physical Graffiti from the 1990s and compare that to the recently release Hi-Res version of the album the difference is stark, indeed. So quite a bit of the Hi-Res releases have been updated to take advantage of the capabilities of the medium. The very recently released Hi-Res release of Abbey Road has never sounded better. You should not play these on your phone to compare them. An audiophile hi-if system consisting of an excellent DAC, amplification, and full range speakers, including a subwoofer, is really necessary to be able to appreciate the difference. If you listen to modern pop, don’t bother. It’s about fashion, not music.

[phidrakeresoration]

Partly true. Increasing the sampling rate by 4 increases the top frequency by double. So, if the highest frequency you can reproduce is, say now 44khz and change, it is still only 1extra octave. It's the detail that the extra samples give you which matters. More samples makes the recording more detailed and closer to the analogue information that you are recording. I am in my 40s so my upper frequency threshold is now starting to be impaired by age. That doesn't stop me noticing better stability, detail or pitch. If I loose even down to 10khz that's only the top octave I have lost. Nothing is linear in sound. It's logarithmic. Remember that.

[dvamateur]

There are more considerations than this:
- low pass filters (preventing aliasing) don't have to be as drastic (steep curve) for Hi-Res audio,
- if the track uses fade away type of ending, you won't hear "graininess" of the audio resulting form less bits of resolution at low volume parts
- if you plan on processing audio further, like slowing the tempo down and bringing pitch to original (to practice your favorite part on an instrument), you have more information to work with in Hi-Res audio to get better results after editing
- DSD is also considered Hi-Res. The advantages of DSD are such, that you get far smoother waveform than recovering from PCM encoding. At 1-bit depth, but sampling rates as high as 2.8Mhz, you still get 4x the information that it's on a redbook CD
- SA-CD, which is an optical disc using DSD codec, is unlikely to be pirated, because there aren't any consumer SA-CD burning devices. Also, and artist recording directly to DSD and release SA-CD album, basically makes his or her master recording available to you, because not much editing software is available for DSD.

So, as I described above, there are practical as well as ethical reasons that are in favor of various forms of Hi-Res audio. As long as we don't resort to dynamic compression for commercial reasons (loudness war) or data compression for convenience reasons (56Kbits/sec or even 128Kbits/sec MP3 can sounds rather horrible), and keep recording sources legitimate (no up-conversion from lower bit rates and lower sampling frequencies just to "appear" Hi-Res), I am all for it. I don't like piracy, and I don't like data compression con with 10:1 compression schemes. If I get a high quality recording from a legitimate source, I'll take it on a MiniDisc (ATRAC 290Kbits/sec) or on redbook CD, DAT, 24-bit digital reel-to-reel tape or in Hi-Res audio file format. MP3 or vinyl are not for me for reasons I'd have to write another comment this long, so I won't get into. As far as Hi-Res audio, I don't have problem with it. Pono player is sort of the lowest quality hardware you can get for Hi-Res audio, and most of files in Pono store are not true Hi-Res. So yes, there's some con there. But my experience with SA-CD discs and audio gear, hdtracks.com and something like Sony NW-ZX2 portable player have been really good. If I dig both the sound and the gear, what's not to like?

[samibari]

I guess I like wasting storage on my sd card

[ian_of_glos]

There is one thing I don't understand here. You break down sound into amplitude (which you say is recorded as bit depth) and frequency, which is recorded as the sampling rate, but what about timbre? How is that recorded? I find that it is the timbre that is missing from low definition recordings and, although you can hear the notes and how loud they are it is the quality of the sound that is often poor.

[nobbeuk]

I can hear the difference in high res music, it’s like night & day. Don’t know what else to say but it’s like when I play guitar I can hear old & new strings where others can’t.

[joshuatyler]

It's not necessarily about end-user listening all of the time. With more data within the audio signal, effects can be made to be more robust. It's just like cameras with a high resolution capture. No one is going to look at the 6000x8000 image once the image is edited, but more color data allows the editor to perform more robust editing on the image.

Likewise, someone who downloads a song at a higher than nyquist audio track will be able to apply better effects to the song, mainly focusing on equalization.

Also, because of nyquist, frequencies at the higher end of the sampling range get sampled as low-end, and muddy up the bass tones of the song. Because of this, the higher end of human hearing at 22kHz gets confused as low end when listening to 44.1kHz audio. Increasing the sampling rate to 96kHz frees up the higher human hearing range to be heard clearly.

This was a very short video, and could have been explained better by someone with sound engineering or electrical engineering experience.

[anthonyshisero]

Andy

Your final comment in this video - "unless an album is released with completely different mastering than its CD equivalent, it will sound exactly the same" - is actually all you needed to say.

We are using hi-res formats to record CD quality music, imagining that somehow the change of clothes (upsampling) will upgrade the wearer of the clothes. True, Unless we have access to digital masters recorded from analog at hi-res levels, we will actually never know of hi-res audio is superior.

Furthermore, it may require a lots of time and exposure to tell the difference, much like learning a new language and going to visit another country - the subtle differences in accent and nuances of meaning become apparent over long periods of time. This could explain the resurgence in popularity of LPs.

[sworddice]

gotta love your presentation. loud and clear. english isn't my first language and i usually need to turn on english subtitles in case i mishear some words.

[MichaelW.]

In essentials, this topic would say everything there is to sampling rates and bit depth. But its riddled from being partially factoid and a bit too theoretical, for its own good.

For example: 16 bit depth gives us 96 dbFS of dynamic range. This is not true, even if we ignore, that people believe, you would get 96dB of dynamic range on the music alone. In fact, you don't get anywhere near that dynamic range for the music, nor would you ever need it. Lets take some classical music as an example. How about Mozart Overture 'Die Zauberflöte'? That piece peaks at roughly -3dBFS and goes down to -71dBFS (RMS). So we are talking about a a dynamic range of about 68dB. And it makes perfect sense! Listening to music at 96dB SPL is not healthy for your ears and hearing.

And now for the overall dynamic range possible in 16 bits: While in theory, the noise floor of a 16 bit recording is 96dBFS, pracitally, its not possible to have a noise floor that low. Its more likely to have about -93dBFS (RMS) as a lowest noise floor. That being said - if you get your peak noise below 94dBFS, its likely, that the conversion does not consider what as any signal and as far as your PCM WAV file would be conserned, you'd have effectively NO NOISE at all. But since noise in that recording jumps from -93dBFS (RMS) to ZERO, you have effectively a noise floor of -93dBFS (RMS) and in peak even worse -90dBFS. Thus you have a signal to noise ratio of a mere 25dBFS vs RMS Noise, a mere 22dBFS vs noise peak. But is that a problem? Not really, no. If even if you played back music at up to 90dB (SPL), what can be considered painfully loud on the loudest parts, you would have literally NO NOISE OUTPUT on your system.

So are 24 bits of bit depth useless? This question has two answers: For listening? Yes. Period. For recording though, 24 bits of bit depth are highly usefull, because you can record with a noise floor even lower. Why is this important? Because you can do more postprocessing to make quieter party more audible, without getting the noise floor to an audible level. So 24 bit recordings have their place - but mostly for the people producing music.


So what about the sampling rate? Anything above 20000Hz is useless content, because nobody can hear it, right? Well, when i was 11, i could hear frequencies up to 21500 Hz. But thats beside the point, because i couldn't do it anymore, when i was 16. Since high quality music isn't produced for kids who cannot afford ridiculously expensive gear to listen to it, but rather for people who can afford it, there is almost no point in even having a max frequency of 20000 Hz recorded. 17400 Hz is a tone only teenagers can hear. Anyone 18 years or older cannot hear this tone. So why would anybody want a higher sampling rate?

Not for being able to listen to higher frequency recordings. 48000Hz as a sampling rate has been standartised for digital video broadcasting. It was a sampling rate, that matches pretty much all the frame rates in used around the world at the least overhead possible. No matter if 24Hz, 25 Hz, 30Hz, 48Hz, 50Hz or 60 Hz. It is easiest to process. (And its the same for 16 bits of bit depth. 24 bit signals are more difficult to process. It needs more computational power.) That is the reason 48 khz of sampling rate and 16 bits of bit depth exist. That's why most computerized hardware, no matter if its a smartphone, a TV, a Computer display or whatever the case is standartized at 48000Hz of sampling rate and 16 bit of bit depth. Its easyer to process than even 44100Hz of sampling rate would be. - Today it doesn't matter anymore, as we have easily enough computational power to calculate the playback of even ridiculous sampling rates like 96khz. But back then, when we began to digitalize TV broadcasting, it was very much a thing.

44100Hz was perfect for audio reproduction, but in context with digital video broadcasting, they needed a sampling rate better suited for the job and ended up with 48000Hz. Any even higher sampling rate was an attempt to deal with issues related to aliasing, which is a form of inaccurate audio unique to the digital realm. To adress the myth: No, its not usually audible. You need to to really extreme examples to make it audible. They tried to make dealing with audio aliases easier, but by know, its pretty much proven, that even in recording, using oversampling in the post processing you can deal with even strong examples of aliasing it at both 48000Hz or 44100Hz without a problem. A higher sampling rate only results in higher CPU loads and bandwidths and is harder on the system, at no real world advantage. They do nothing to improve the audible content below 20000Hz, like its stated in this video. But its even worse: Higher sampling rates can result in increased intermodulation, even in the audible range. To put this another way: You pay exponentially more money to realize a pristine recordings AND playback of 96Khz 24 bit signals, than what you would pay for playback of 44100 or 48000Hz @ 16 Bits -- again, at NO advantage. And i challenge anyone who still believes strongly to hear a difference, to do a real blind test. If you don't get at least 80% of your quesses right, you got to admit to yourself, that you're blinded by the numbers.

[fuzzytalz]

As someone who performed as well as listened to music for a significant part of my life, I must interject there is more to experiencing audio than what one hears. There's also the elements that are merely felt, but which contribute greatly to the sense of space in a well-recorded album. There is an immediately noticeable difference in dynamics, for instance, between REM's Bang and Blame on CD compared to a 24-bit/96 kHz FLAC.

[juanmillaruelo]

Andy, the ultimate objective is to having more dynamic range, not less. Your presentation lead to think that since all music has been overcompressed (not true, just pop music for radio use) we already have a surplus of bit depths. 24 bits are making a comeback because you can lower the noise floor and put the dynamic range just where it was before the "loudness wars". Encoding and filtering are not flawless: they create "artifacts" such as spikes and brush. With higher sampling rates this noise can be shifted up into the inaudible region, then filtered away with more subtle filtering (not "brick wall") that lowers this noise.
Furthermore: high sampling rates do not just broaden the frequency range but ALSO improve quality by providing more sampling points that make "interpolation" and "anti aliasing" more accurate to capture and render real world complex music thereby pick up timbre, transients and other subtleties in a way that CD 16/44.1 does not.
You have had the revelation of Nyquist and Shannon, but there is a LOT more science and technology to learn and apply. Your field at present is obviously not digital audio. I encourage you to dig further into the matter if this field interests you. We are reaching a tipping point. The default medium for sound is digital, and we seek the highest possible fidelity. The 16/44.1 is Red Book spec from 40 years ago. Resources were limited and costly back then. CD was barely sufficient, but it got the basic job done. It was a milestone, but a lot has happened since then and there is a lot to learn and do.
Red Book is now the "bare essentials", our basic starting point. Almost "the shellac 78 rpm disk", if I may. And engineers are routinely working today extending accurate music reproduction way beyond those parameters. Dip in further. Contemporary digital sound is fascinating.

[mohanshivamusicfactory]

192khz, , gives resolution, , , , audible frequency is different subject, sampling is different, , sampling means number of points in time line, , record at high sample and down grade to 44.1