I was thinking about this last night and wondered why we define sounds as being 'sonically' analogue or digital. I ran a 16 track studio for decades and mastered to half inch two track. On paper, from the source at my microphones, through to the finalising process, everything rolled off well below 20kHz. Now, without those limitations we can place that "air" factor into our recordings. Then we go looking for tape sim saturation plugins to get the glued sound so sought after. Perhaps a smooth slope roll off at 16kHz would even out the digital-ness of DAW produced audio.

 

Im not a mastering engineer, but have plans to really learn about it, though 20khz cut seems way too high imo.
13-16khz seems to be the sweet spot. Certainly depends on the material, but i cant imagine heavy metal or edm going so far up it would destroy your ears unless you eq all the channels and busses like you are supposed to then usually nothing hits that high anyway.
Lindell te-100 has a hi cut set at 12khz and can be used as a mastering eq.
You dont have to boost the 15,16khz range to get sweet silk highs, there are ways to boost just 7,8,10 khz range if you need more presence or whatever.
Though its better to resort to saturation plugins in this case, ivgi or in cubase you can choose magneto on the channel strip and boost some highs with it and it sounds better, non fatiguing.

 

Analogue audio and digital audio are quite different. Well, what I'm referring to is that in analogue there is no nyquist frequencies, there's no limit on the frequencies. I would completely agree if we were talking about analogue hardware, but we're talking about digital plugins. There's an interesting article here you should read up:
https://vladgsound.wordpress.com/2014/12/21/tdr-ultrasonic-filter-alpha-version/

In essence, when mastering at 88.2 or 96kHz and it sounds better, especially when you use outboard hardware and ADDA, it is a good practice to cut high frequencies before any digital dynamic processing like limiting. It is not essential and it doesn't make a huge difference, but it's a good practice and it doesn't hurt to do it. It could depend on the quality/genre of the music, too. Acoustic, ambiental music, jazz and classics should preferably be mastered by people who know how to do it properly. ;)

Cheers!

 

Last edited: May 4, 2017

Generally 12dB or 18dB/octave [2-3 pole] HPF from 20 to even 50Hz, depending on how much excessive bass there is, and 6dB or 12dB LPF [1-2 pole] at between 15-18kHz, also depending on how much highs there is. You have to remember that this is not a brickwall filter. You don't lose those frequencies, you just attenuate them with gentle filters like that.

And the very good thing about HPF is that it acts as a DC suppressor so you can get more dynamics from your track. DC is very bad, it consumes the headroom without being of any benefit to the sound and it's quite important to keep it at bay at all times.

Cheers!

 

no they won't, if you do it properly.

Cutting what you can't hear will increase headroom for what you can hear.

 

was the file wav or mp3???

 

I cut around 17/18k.
IF I CAN'T HEAR THEM, WHY OTHERS SHOULD!

edit: i actually can't hear above 16k, starting from 15k

 

Last edited: May 9, 2017

FLAC is FREE = mp3 is dead. (unless rare occasions
you don't want to cut below 20hz or above 20kz (but use a piece of gear that has the slope you want if it must be changed
If anyone is cutting like that, they missed steps BEFORE they bounced.
you don't want to go in and start hacking things, you will not get a final result as the best mixes and references in the genre.

 

No, they use AAC on Youtube, Spotify & Apple Music.

 

Never! I only use 368Khz 64bit dongle floating point WAVs.
I just grab 128kbs youtube mp3s and upsample & interpolate them with Harrison Mixbus using a low brain pass filter.

Then I perform a losslessmess compression to NCW (I'm sorry guys, FLAC is for losers ) and I use Kontakt as music player. One nki for album, one nkm for discographies. Art scans in nkr, lyrics in dummy scripts and drugs in whisky.

Hell yeah, I'm the man...

 

Last edited: May 4, 2017

Octuple-rate DSD
A further extension to the development of DSD is DSD512, with a sample rate of 22.5792 MHz (512 times that of CD), or alternatively 24.576 MHz (512 times 48 kHz). Hardware such as the Amanero Combo384 DSD output adapter, and exaU2I USB to I²S interface, and software such as JRiver Media Player, foobar2000 with SACD plugin and HQPlayer are all able to handle DSD files of this advanced sampling rate fully natively.

 

Finally a format that can record my neighbor screams without aliasing...

PS: Which reminds me, now seriously, as long as these high frequency components/noise/harmonics aren't aliasing artifacts it should be fine. You can filter based on your needs.

 

Last edited: May 5, 2017

I'm not a 192kHz/24bit freak, but there is a big difference in sound quality between an MP3 and FLAC file at 44.1/16. It's just that the music nowadays sounds so crappy that you can't even notice the low quality, because it's already pretty low quality and full of distortion. That distortion kinda masks the MP3 ringing artefacts, so it all nicely sounds like a terrible mess.

 

What makes most MP3s unlistenable IMHO is the butchering of M/S image, and huge differences between encodings. In other words, there's no guarantee that one 320kbps will sound as good as the other one - bitrate is "meaningless" when trying to determine quality. I have seen moderately acceptable encodings at 192kbps and unusable ones at 320kbps.

Clipping is another factor. All modern lossy codecs support floating point, so hitting positive dBFS values is of no real concern. Not for MP3.
Even MP3 compatibility is more and more challenged by MP4/AAC and OGG-Vorbis, there aren't many devices that won't play AAC.
OGG-Opus is still not widely adopted outside Youtube and perhaps Rockbox, but in time it'll surely receive its recognition.
Personally I keep my whole music collection on rockboxed Clip+ in 128kbps Opus and I have yet to spot any subjective difference compared to lossless.

 

Extremely interesting this new Opus codec. I'm afraid as long as its free it won't be fully supported. Mp3 and Aac are standards but not free, just like aac's video counterpart x264.

 

VP9 is open-source and it's widely adopted. Ogg-Vorbis too, it's not that popular but good 90% of all audio devices can handle it.
The time for Opus will surely come.

 

It says it all.

 

Please guys see this video!

 

To the OP: You're talking about 44.1khz material? 44100hz means it can't reproduce frequencies above 22050hz, and stuff above that frequency _must_ be cut to prevent sampling errors (i.e., aliasing). An analog cut can't be 'too steep' because it would create resonances.
Easy way of explaining: To make sure there is absolutely _no_ frequency-content above 22050hz, you have to start cutting 'gently' from somewhere before that, let's say 21000hz.

What I mean with all this is that unless your track is completely recorded and mixed in the analog domain (and then delivered for output mastering to a digital studio), you're ending track will not contain any (real) content around 21khz and up anyway. 20khz and 21khz you can be sure no one is going to notice, so somewhere around there people decided it would be safe to cut 'gently' anyway to be safe.

You absolutely don't have to cut everything, there won't be any 'real' content there anyway. A good mastering engineer will make sure there are no loud frequencies around anywhere that can damage speakers or hearing or so, but around 20khz and up it's pretty safe, unless you're actually peaking at that frequency :P.

The only reason why some people might be cutting at around 16khz / 18khz is that they think like this "MP3 or audio file formats these days are going to cut around there anyway, so I want to preview how that sounds in case I want to compensate somewhere else in the frequency spectrum". Not that true any more with modern AAC / Voribs / blablabla file formats.

And let's ignore the fact that there are very little people on the world who can actually hear any details around 18khz and up .

 

this is inaccurate, play a high quality 192khz 24 bit file with highs extending well beyond 22khz
compared to a file that actually has zero above 18khz
(one ex. is black sabbath SACD dsd converted to flac avail at "the bay"
bring it in your daw, convert to your fav shit codec with brick wall filter at 18kz 16 bit.
A to B homie.