Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

24/192 Music Downloads and why they make no sense (2012)

people.xiph.org

Post reply on HN
Print view View original

301–310 of 328 posts

Re: 24/192 Music Downloads and why they make no sense (2012)

#301

Post by KozmoNau7 »

Mediterraneo10 wrote: post #20
vortico wrote: post #7

Earlier quoted context omitted.

To thank the article even more, you could resample to 16/48kHz.

Even if the lower sample rate of 48kHz would be entirely reasonable and 96kHz is overkill, 24 bits still makes an audible difference for the material I listen to (modernist classical music and ECM jazz), which is why you can find 24/48 from some labels. For pop music, which of course is distinguished by little dynamic range, then 16 bits would be fine just like on the CD format.

To put things in perspective, 16-bit PCM audio has a noise floor around -96dBFS, ie. the difference between the loudest possible sound the format can contain and the noise floor is 96dB. That's what the bit depth determines; the level of the noise floor in relation to the loudest reproducible sound. It does not add any more detail, it's not like the resolution of an image file, the added bit depth does not allow for finer-grained details, audio doesn't work like that.

96dB is a lot more than you probably think, it's like the difference between an anechoic chamber (nominally ~0dB) and someone jackhammering concrete right next to you (~90-100dB). Add to this that even a quiet room has a noise floor around 20-30dB, and to even hear the noise floor in CD quality audio, a full-scale peak would hit 130dB!

Try generating a sound at 0dBFS, the attenuate it in steps of 10dB and make note of when you can't really hear it anymore. At -50dB the sound is already extremely low and barely audible, and there would still be 46dB of attenuation available.

In addition to this, noise-shaped dither can push the noise floor towards frequencies where the human ear is less sensitive, giving a perceived noise floor of around -120dBFS. In other words, 24-bit audio for distribution and listening is absolutely pointless and has absolutely no audible difference when compared to 16-bit audio.

Re: 24/192 Music Downloads and why they make no sense (2012)

#302

Post by AsyncAwait »

nirvdrum wrote: post #241
AsyncAwait wrote: post #187

Earlier quoted context omitted.

As far as I am aware, iTunes is not even able to play back FLACs, so when I am on a Mac, I use Clementine, ( https://www.clementine-player.org ), or cmus, ( https://cmus.github.io ). Converting etc. I do exclusively on my Linux desktop, so can't help you there.

Heh. I'm actually on a Linux desktop but figured most people would reply with an iTunes-based solution. Cmus looks interesting I'd love to hear what your workflow is for converting, naming, tagging, getting artwork, etc.

Heh, nice :-) Yeah, cmus is incredibly convenient for rapid playlist management once you learn the shortcuts, (there's an excellent quick tutorial $ man cmus-tutorial).

I mostly use 7digital & HDTracks to acquire FLACs these days, but when I rip from CDs, I use https://github.com/whipper-team/whipper to do the job.

FLACs from 7d/HDTracks are already named & tagged properly so I only deal with it occasionally and when I do, https://picard.musicbrainz.org works well for acquiring tags & artwork.

When I need to rename/tag manually, https://kid3.sourceforge.io has been working nicely.

Also I haven't used it myself, but there's a lot of positive chatter around https://github.com/beetbox/beets for tagging etc. I just prefer not to have my files touched in such an automated way :-)

I rarely actually convert from FLACs these days, since I have set up Airsonic, (https://github.com/airsonic/airsonic), on my home server. I now have access to the lossless files directly, from anywhere.

When I do convert, I usually just use https://github.com/kassoulet/soundconverter - nothing fancy, but does the job. I do not maintain my whole library in both, lossless & lossy formats since I have set up Airsonic, but when I do want to save data & do not have access to WiFi, I just let Airsonic use lame to transcode to MP3s on the fly, (rare). If you cannot do that, don't have regular access to data on the go etc. I'd honestly just use https://ecasound.seul.org/ecasound/Documentation/examples.ht... and put it in a script that checks if a .flac file in a folder or subfolder has a corresponding .mp3/.ogg file and convert if not, then just use find to filter out the format I don't want to copy over. :-)

Re: 24/192 Music Downloads and why they make no sense (2012)

#303

Post by dusted »

Mastering should be done for studio monitors. No, studio monitors do not sound the same, but they are somewhat neutral, they sound somewhat in the same ballpark, which is the point of them to begin with, have a flat frequency response (which does not imply that they "sound flat" just that music mastered for active-subwoofers sound flat).

This way, those who wish to hear how the music was intended to sound, will have a somewhat decent chance of coming near to what it sounds like, and people who want other flavours can still simply buy equipment which colors it in the direction they desire.

Re: 24/192 Music Downloads and why they make no sense (2012)

#304

Post by ZoomZoomZoom »

jrace wrote: post #154
widerporst wrote: post #118

Earlier quoted context omitted.

Interesting. Are you sure you were using the right settings and a recent version of LAME for this? The only artifact I can reliably hear in 320 kbps MP3s is pre-echo, for instance with castanets, and only in a few very specific situations. Apart from this, V2 and above sounds completely indistinguishable from the original to me.

Which is great ...for you... but what if in the future you could hear the differences? And now your entire collection is in MP3 V2 - now what? No reason to not rip everything in uncompressed FLAC these days. But if your happy with your audio now, great!

There was never a reason to rip in anything less than lossless codecs! Lossy codecs are only for consumption in technically restricted conditions, lossless are for storing and listening.

Re: 24/192 Music Downloads and why they make no sense (2012)

#305

Post by ZoomZoomZoom »

tohnjitor wrote: post #91

High resolution audio is important to me as a sound designer because of the ability to severely slow down a piece of audio without any aliasing or stuttering. At 96KHz and higher with certain samples I can slow down by 80% and it will still sound good.

Do you mean slowing down while lowering pitch (without resampling)? If so, you're correct, as you bring harmonics from out of limit of human hearing back and the result sounds natural.

But if you mean just changing the speed of the sound, than you need to change the algorithm you're using. There should be no difference in quality due to sources having different sample rates.

Re: 24/192 Music Downloads and why they make no sense (2012)

#307

Post by Majromax »

tomc1985 wrote: post #283
Applejinx wrote: post #257

Earlier quoted context omitted.

No, not at all. Vinyl has a wildly inconsistent noise level where rumble predominates, and people conflate this with bits of resolution. Vinyl's behavior is not easily pinned down relative to 'bits of resolution', because the noise floor is skewed so intensely towards low frequencies. To say nothing of how generally available vinyl records (especially old ones) have wildly different rms/peak measurements than general…

2-bit digital audio? Like only four total values of dynamic range total? 4-bit meaning only 16 total possible amplitudes? Is that even physically possible? ;) I agree that the quality of the record -- AND its playback equipment -- among other physical factors will dramatically effect the numbers. My "10-14" quote only applies for ideal conditions: a newly-minted, unplayed disc on a high-quality preamp which together…

> 2-bit digital audio? Like only four total values of dynamic range total? 4-bit meaning only 16 total possible amplitudes? Is that even physically possible? ;)

Surprisingly, yes. With noise shaping (https://en.wikipedia.org/wiki/Noise_shaping), very coarsely-quantized digital audio can produce high signal to noise ratios in the audible frequencies, via quantization techniques that push the error towards ultrasonic frequencies.

This doesn't violate information-theoretic limits because noise shaping requires very high sampling rates. The 1-bit Sony DSD format (https://en.wikipedia.org/wiki/Direct_Stream_Digital) used a 2.8MHz sample rate.

In the case of vinyl, the effective sample rate is physically limited by the (linear) record speed divided by the vinyl grain size, and to a rough approximation the bit depth would be log of the maximum groove amplitude divided by the grain size. However, the analog cutting mechanism would greatly limit the opportunity for dithering and noise shaping -- for example a needle cannot cut a wave shorter than the tip size.

Re: 24/192 Music Downloads and why they make no sense (2012)

#308

Post by justin66 »

gwbas1c wrote: post #288
justin66 wrote: post #183

Earlier quoted context omitted.

> When I casually researched the upper limit of human hearing, I came across something that mentioned that some people can detect lowpass filtering up to 27khz. Link?

https://en.wikipedia.org/wiki/Hearing_range#Humans

Thanks. It's wikipedia but there's actually some good stuff there. The important link from that page seems to be:

https://asa.scitation.org/doi/full/10.1121/1.2761883

A decade has passed and it would be interesting to know how many people have reproduced the results detailed in the abstract. I gave it a quick read and at first glance it looks like an impressive experiment:

Hearing thresholds for pure tones between 16 and 30kHz were measured by an adaptive method. The maximum presentation level at the entrance of the outer ear was about 110dB SPL. To prevent the listeners from detecting subharmonic distortions in the lower frequencies, pink noise was presented as a masker. Even at 28kHz, threshold values were obtained from 3 out of 32 ears. No thresholds were obtained for 30kHz tone. Between 20 and 28kHz, the threshold tended to increase rather gradually, whereas it increased abruptly between 16 and 20kHz.

Re: 24/192 Music Downloads and why they make no sense (2012)

#309

Post by grawprog »

KozmoNau7 wrote: post #294
grawprog wrote: post #172

Earlier quoted context omitted.

You also lose some of the fullness on the extreme low end, it's noticeable even with a fairly low end subwoofer. Something also ends up missing in the midranges. I was working on a track once where all I had was 320 mp3 version of the vocals. At some point I replaced it with a flac copy of the same vocal recording, from the same original wav source and the difference was noticeable right away without changing any of…

The low end thing just isn't true at all, and I don't know where that myth comes from. I have a room-corrected setup with two properly adjusted subs, and MP3 does just fine on deep bass content. Regarding solo vocals, the history of the MP3 format says: "The song "Tom's Diner" by Suzanne Vega was the first song used by Karlheinz Brandenburg to develop the MP3. Brandenburg adopted the song for testing purposes, listen…

>The low end thing just isn't true at all, and I don't know where that myth comes from.

Well for me it came from running multiple copies of the same bass heavy tracks encoded in different formats through spectrum analysers. But I guess those lie?

>Regarding solo vocals, the history of the MP3 format says: "The song "Tom's Diner" by Suzanne Vega was the first song used by Karlheinz Brandenburg to develop the MP3. Brandenburg adopted the song for testing purposes, listening to it again and again each time refining the scheme, making sure it did not adversely affect the subtlety of Vega's voice".

Human voices come in a wide range of tones and frequencies. Optimizing something for one voice doesn't mean all voices will benefit from the same optimizations. The specific track I was referring to had a lot of variation in high and low notes. You can tell me all you want what I did and didn't hear.

Re: 24/192 Music Downloads and why they make no sense (2012)

#310

Post by magduf »

sjwright wrote: post #273
magduf wrote: post #112

Earlier quoted context omitted.

Actually, no, it might sound better cut to vinyl. Remember, vinyl doesn't have the frequency range or dynamic range that digital audio does, and it has to be mastered using the RIAA Curve because of the properties of the medium. One factor here is that the stereo separation on vinyl can't be too large, or else the needle will literally jump out of the groove! In short, you can't just take CD music (no matter how well…

> it might sound better cut to vinyl There's nothing extra in a 192kHz signal that would help with the vinyl mastering process. You could make a technical argument for the benefits of a 24-bit source, but in practice even those benefits would be utterly swamped by the SNR of vinyl.

You're completely misunderstanding. No, there's nothing extra in the 192kHz signal that would help, but that's not the point. Just remastering it for vinyl, which has certain limitations, might make it sound better than the digital format. As I said, you cannot just cut CD audio to vinyl; the medium won't allow it. You'll probably have the needle bouncing out of the track. So remastering for vinyl might actually make it sound better than the abomination that is the sound of modern CD-audio (because of the Loudness War and extreme compression), not because the fidelity is better (in fact, quite the opposite), but because vinyl's limitations will prevent them from making the audio sound as horrible as it does on CD.
Post reply on HN