lyons
Full Member
Posts: 28
|
Post by lyons on Aug 20, 2024 15:38:22 GMT -6
Hi everyone,
So out of curiosity and to check as a reference, I loaded up a record taken from YouTube (very successful artist) and had a look on the eq graph. The thing which stood out was that there was almost nothing above 16k. I’ve also seen a mix engineer online talk about cutting at 15k to make a mix more radio ready. What did he mean by this?
And why would a hit record having no information above 15-16k? It was almost a brickwall cut despite having a healthy high end up to that point of the cutoff. Does YouTube diminish everything above a certain point? Even if so, why would cutting at 15k produce a more ‘radio ready’ mix?
Really interested to understand this as I haven’t found much online by way of explanation.
Thanks
|
|
|
Post by bossanova on Aug 20, 2024 15:52:53 GMT -6
Hi everyone, So out of curiosity and to check as a reference, I loaded up a record taken from YouTube (very successful artist) and had a look on the eq graph. The thing which stood out was that there was almost nothing above 16k. I’ve also seen a mix engineer online talk about cutting at 15k to make a mix more radio ready. What did he mean by this? And why would a hit record having no information above 15-16k? It was almost a brickwall cut despite having a healthy high end up to that point of the cutoff. Does YouTube diminish everything above a certain point? Even if so, why would cutting at 15k produce a more ‘radio ready’ mix? Really interested to understand this as I haven’t found much online by way of explanation. Thanks Youtube is often lossy encoded at a rate that cuts off frequencies above 15k to save space in the audio file.
|
|
lyons
Full Member
Posts: 28
|
Post by lyons on Aug 20, 2024 16:00:06 GMT -6
Thank you! I guess that explains it.
The only thing remaining is why a cut at 15k would produce a more radio ready mix. I’m not even sure I even know what radio ready means.
|
|
|
Post by tonycamphd on Aug 20, 2024 16:25:24 GMT -6
WWIW, I can’t hear a sine wave above 15khz with my crusty old ears, now take a band from 15kz to 20khz of white noise and I can hear it
|
|
|
Post by drumsound on Aug 20, 2024 16:41:04 GMT -6
They may have been considering radio transmission when cutting above 15K. Was it an old school mixer who mixed a bunch of radio hits?
|
|
|
Post by Tbone81 on Aug 20, 2024 16:55:12 GMT -6
Some mixers aggressively cut high end from most instruments. There’s not a lot of useful audio info that high up, on most instruments, and it leaves room for the few instruments that really need it…like OH, vocals etc. Also lost people over 40 can’t really hear that high anyway. That’s not to say you can’t over do it. If you brick wall filtered everything above 15k a lot of mixes would fall apart.
|
|
|
Post by EmRR on Aug 20, 2024 19:32:11 GMT -6
FM radio cut off above 15k so it’s a cultural reference at very least. Cutting all that makes for massive phase distortion, FWIW.
|
|
|
Post by drbill on Aug 21, 2024 9:13:45 GMT -6
Be very careful. I had to master a record from a group / mix engineer with little real world experience that held one of the three letter mixers in high regard. Said mixer had a very popular interview where he said that to make things sound more "analog", he did low pass (HF cuts) around 12k. The mixes for this record sounded horrible and muddy and closed in. At some stage of the production chain, the mix engineer had reprinted all the tracks with nothing above the 12-13k range - so there was no "going back". The original files were gone. That was a VERY difficult record to master, and I had to pull out every trick in the book to even get it sounding normal again. I mentioned to them my analog decks went to upwards of 30k, and they thought I was crazy. But then, I don't have a 3 later acronym for a name.....so........
If you MUST take this approach, I'd suggest rolling off the top till you start to hear it - then go back up 1-2k. Do not do all instruments the same cut. Leave room for some aspects to breathe, and please, don't permanently print them that way.
|
|
|
Post by niklas1073 on Aug 21, 2024 9:30:05 GMT -6
Yeah, mp3 will often have this “feature”, I think it was even on mp3 bounces from pro tools i had (might have been another software conversion too) just a while ago was cut at 16k. Then it showed a low peak at around 20k. Very strange but i think this is common on mp3
|
|
|
Post by christopher on Aug 21, 2024 9:54:02 GMT -6
If this is a production thing, and might be part of beating the LUFS algorithm, I’ve noticed a lot of tracks have way more quiet part/loud parts, silence, and short moment of full loudness, then back to quiet rolled off parts. Seems to help the avg loudness, maybe it’s intentional? Here’s the way I manage above 8k.. The highs above are like spotlights. So if you want the singer to be in a spotlight, everything else should be rolled off. (Or at least low volume/wet verbs if you want to keep ultra highs) Very common for pop. If it’s a singer and her acoustic, probably will let the acoustic have some ultra highs as well. Dance? You want the kick and snare to flash the dance lights so they are gonna have top end as well. The harder you roll other stuff off the more the spotlight effect will work. You don’t want the singer up there to be outshined by a cowbell For rock or country chorus you might like all the lights, so everything might work better with ultra highs. Hope this helps
|
|
|
Post by bossanova on Aug 21, 2024 10:01:24 GMT -6
Yeah, mp3 will often have this “feature”, I think it was even on mp3 bounces from pro tools i had (might have been another software conversion too) just a while ago was cut at 16k. Then it showed a low peak at around 20k. Very strange but i think this is common on mp3 Some of the mp3 quality levels have a thing built in where there's a hard cut at a given frequency but there's leeway to momentarily extend it for certain peaks. V-0 cuts off at 20K but can extend up to 22k if I recall correctly. I think 320 is the only one from the original spec that went all the way up to 22k. (I also remember that V-0 was encoded in joint stereo vs 320 retaining the true stereo channels. All these things were a big deal way back when.)
|
|
|
Post by EmRR on Aug 21, 2024 10:10:42 GMT -6
For the most part I find that a desire to cut treble like this means the mix isn't balanced in the first place. Might need to do it if you're seeking an 'unnatural' balance between some elements.
|
|
|
Post by tonycamphd on Aug 21, 2024 10:26:22 GMT -6
Never on a master(i'm not a mastering guy) I use maxed out linear phase hp/lp filters on every track except kick and bass(natural phase pre ringing), they are on every single channel of my templates first plugin in series to zero out D compensation, I high pass appropriately but low pass is trickier to me because i can't hear all that high, i use the display in ProQ to see whats up there, you can easily see the weird stuff that has to go, and then i conservatively cleave off the upper stuff leaving room for the fact that i cannot hear most of that, i do believe multiple tracks with no low pass filtering create excess hf hash that i can certainly hear especially after running it through all the bells and whistles of further plugs and processing By doing this FIRST a good portion of the successive plugins and gear i use after creates upward harmonics that fills the space i made with desired hi freq air vs hash, clean and airy up there, thats my story, i'm sticking to it, it works for me
edit; i should add that my understanding WITHOUT filtering your speakers are working to represent all the unmusical stuff above and below filtering points causing inefficiencies hash and mud, that said my mixes after filtering and before anything else feel dull AF, but if i was desert islanding it, id' want hp/lp filters over almost anything
|
|
|
Post by Dan on Aug 21, 2024 10:31:08 GMT -6
I’ve never low passed anything except for nasty electric guitars, synths, and fx sends. I’ve never lowered a high shelf except for overheads and things recorded to tape where it was cranked beforehand. I’ve used the tdr ultrasonic filter on stuff recorded with detective converters at 44.1 or 48 kHz that didn't even have half band filters, usually a brick wall filter with inadequate attenuation. People need to stop using mics, gear, and plugs with a messed up top end but stupid is as stupid does. The bad Chinese condensers, aliasing, crap Japanese opamps, and phased top ends continue.
Dan
|
|
|
Post by Dan on Aug 21, 2024 10:45:05 GMT -6
Never on a master(i'm not a mastering guy) I use maxed out linear phase hp/lp filters on every track except kick and bass(natural phase pre ringing), they are on every single channel of my templates first plugin in series to zero out D compensation, I high pass appropriately but low pass is trickier to me because i can't hear all that high, i use the display in ProQ to see whats up there, you can easily see the weird stuff that has to go, and then i conservatively cleave off the upper stuff leaving room for the fact that i cannot hear most of that, i do believe multiple tracks with no low pass filtering create excess hf hash that i can certainly hear especially after running it through all the bells and whistles of further plugs and processing By doing this FIRST a good portion of the successive plugins and gear i use after creates upward harmonics that fills the space i made with desired hi freq air vs hash, clean and airy up there, thats my story, i'm sticking to it, it works for me edit; i should add that my understanding WITHOUT filtering your speakers are working to represent all the unmusical stuff above and below filtering points causing inefficiencies hash and mud, that said my mixes after filtering and before anything else feel dull AF, but if i was desert islanding it, id' want hp/lo filters over almost anything Solution is to ditch the analog and digital processors that make the nasty high end. “Emulations” and limiters are the worse culprits. Typical early digital limiters and the ones that imitate them no matter how sophisticated they look on paper are basically downwards bit crushers with smoothed attack (some of them have lookahead without the windowed maximum smoothing so sound even worse) and sound awful. Easiest solution is to get rid of them. Others have too long lookaheads so randomly ramp down past events based on the future and others are distortion devices or clippers. Some that aren’t and work pretty well are the limiter 6 ge (best peak limiter. Zippy to clean to soft depending on how you set it), stealth limiter (best maximizer that is like a bus comp into a limiter but it does change the sound and mix a bit), ursa boost rev b (insane upwards distorted to pumpy limiter, cleaner but pumpier than the stealth limiter, less top end lost than tdr in multiband but it is pumpier and you can make it zip bass more on faster lookahead/attacks. I tend to think it is not as good as the other two but it is a cool effect and is the loudness king now). The multiband stuff will all defeat your mix doing anything real off. So will the too long lookahead stuff. They will undo your eq and fader moves.
|
|
|
Post by frans on Aug 21, 2024 10:54:24 GMT -6
Back in the days the german broadcast modules had filtering to have less signal above 15k, somthing with the radio transmission gear i rather leave out than to explain it wrong. But apart from that, why filter anything if your signal is sounding okay?
|
|