/mu/ - Music (Temp full images)

320 mp3's sound pretty much the same as flacs for me honestly. Unless you paid for the high quality and got the 320 then there is nothing to complain about anyway.

So many people don't check the spectrograms of the stuff they download on Slsk either. That place is riddled with tons and tons of horrible transcodes that people have blindingly kept and shared on. Spectrogram your shit people.

>>52390735
Yeah it seems to me like there's no real difference between 320 and flac unless you have a really nice expensive system and a room with good acoustics (or nice headphones I suppose). There's definitely a difference you can hear between 160 and flac or 320 though. I'm not complaining, I know I can't expect much out of free music. Just trying to help people who might not realize that what there getting isn't what they think it is.
Here's the opening of the same song, first from a .flac I torrented and then from a 160kbps mp3 I transcoded from that .flac. Identical except for the noise above the flac's cut off at the same frequency as the mp3. 23.1 mb vs 4.14 mb for the exact same audio output, so more than five times the file size with not difference in quality because it's a bad transcode. People need to be aware of this shit, that's a lot of wasted storage.

>>52391062
they're* woops

>>52390994
I've heard other stuff with bad sample rate conversions, really low quality mp3 encoders, poorly applied noise reduction filters, etc.

You can't really trust people to not fuck things up completely when ripping music.

What's a good spectrogram program? And I'm just looking for low cutoffs like in your image?

>>52391137
spek

>>52391114
this
i dont dl anything that isnt a v0 sourced from what.cd anymore

>>52391214
I want into what.cd so bad but I've never used irc so I dunno how to set up the interview ;_; Do they actually remove bad transcodes?

>>52391197
Okay I downloaded that and I don't know what I'm looking for. This is the spectrogram of the song I opened.

>>52390329
>>52390329

don't mix up kbps and kHz- "kilobytes per second" and "kilohertz"

kbps is how much information is in the signal, kHz is the frequency scale.

your original post mixes these up a lot, and it's very hard to understand. I'm an audio engineer and I almost lost the entire point of your post.

>>52390329
>hard cut at 16 to 18 kbps
should be "hard cut at 16-18 kHz"

and

>>52390329
>20kbps in the higher nitrate

20 kHz in the higher bitrate

beyond that, you're right. you can hear this stuff all over youtube if you look in a spectra. sometimes people upsample into flac even after things have been coded into mp3s.

beyond the cutoffs for lower rates, mp3s mask drop bits that happen after loud transients in the spectrum, so for instance if a big cymbal hit happens that stretches from the lowest part of the spectrum to the high, it will make the subsequent upper partials of other instruments temporarily drop out- that's why low bit rates sound all "swishy", it's the upper part of the spectrum being modulated unnaturally. with some training you can kinda hear the difference between 256 and 320 and flac.

>>52391942
Oh shit I didn't even notice that. I know the difference, just typed it up in a hurry bc I'm trying to clean out my flacs that aren't actually in flac quality and replace them with mp3 til I can find real flac as quickly as possible. My bad. It's pretty late where I am as well.

As for the dash instead of to between 18 and 20 I don't really understand why it matters but ok

>>52391321
What you want to look for from what I've read is a "shelf" at a certain frequency, where there's no information above that frequency. A real full quality flac file, assuming that it's in the music, will have sound from 20-22,000 Hz. This isn't necessarily true for the entire song/file because some parts of the music may not have frequencies that high, but they will probably show up somewhere in the song/file. I hope that made sense.

Again, from what little research I've done, a flac file that's been transcoded from a 320kbps mp3 will have this shelf or cutoff at around 20khz. As the quality in kbps of the original lossy file, which is usually an mp3, goes down, the shelf gets lower, down to somewhere between 16kHz and 18kHz for ~160kbps source mp3.
If the guy from a couple posts up is reading this, I know bitrate probably isn't necessarily a direct measure of "quality" but I'll use it as one for the sake of brevity.

Sorry if any of that is confusing or not as direct as it could be, I'm kinda tired.

I know I'm blowing my own thread up here but I'm now wondering if the master will be intentionally cut short at or above 20kHz for some records, because my copy of AM appears to get up into 21kHz. Is that probably just artifacts?

>>52392704
>master will be intentionally cut short at or above 20kHz for some records

I imagine they do that in some cases because if you have excess sound high in the spectrum, it can create foldback distortion. the nyquist frequency is half of the sampling rate, so they probably tend to give themselves a buffer since people almost never hear above 20k