This is the greatest internetting I have seen in ages. Thank you. Also, I completely agree, as my experience with recording both bands and unicorn serenaders has yielded exactly the results you describe!MoreSpaceEcho wrote:forget RMS. don't even think about it. it means nothing.
the shortcut is just using your ears. they work great for this.
anything with super loud bass (or sustained low mids) is gonna read a high RMS. means nothing as far as loudness. anything that's super bright/lots of high mids is gonna sound loud, regardless of what the meters say.
as noted, a solo guitar/voice tune is gonna sound loud compared to a full band arrangement. you could have the biggest, meanest, most-tattoo'dest, beardiest sludge metal band of all time, and you put their record on next to a recording of an 8 year old girl playing ukelele and singing a love song to her pet unicorn, and the kid will make the metal band look like chumps. if her sister plays a flute solo in the middle, it's really all over.
do what lyman did, compare your stuff with records you want to stand up against. compare BY EAR, ignore what the meters say. listen to them loud and listen quietly.
perceived loudness has way more to do with arrangement, eq, and the sounds themselves than it does with limiting or compression or anything. a great mix of a great tracking job is going to sound louder than a shitty mix of a shitty tracking job, even if you squash the shitty one to death and it reads louder on the meters.
tl;dr meters are useless, ears are good.
Aiight, let's talk RMS level during psuedo-mastering
-
- alignin' 24-trk
- Posts: 71
- Joined: Thu Jul 03, 2008 8:53 am
- Location: Upstate, NY
"Sounds to me like you're just a little bit cranky"
-
- zen recordist
- Posts: 6686
- Joined: Wed May 07, 2003 11:15 am
"As noted, a solo guitar/voice tune is gonna sound loud compared to a full band arrangement. you could have the biggest, meanest, most-tattoo'dest, beardiest sludge metal band of all time, and you put their record on next to a recording of an 8 year old girl playing ukelele and singing a love song to her pet unicorn, and the kid will make the metal band look like chumps. if her sister plays a flute solo in the middle, it's really all over. "
That's hilarious, and I've noticed this in stuff I'm working on. I have a gut feeling about why, but what's the succinct explanation? I know this but couldn't explain it to the sludge metal guys.
Is it "no low end"? So all the energy is available to scorch your ears with higher frequencies? Is it human hearing or perception?
That's hilarious, and I've noticed this in stuff I'm working on. I have a gut feeling about why, but what's the succinct explanation? I know this but couldn't explain it to the sludge metal guys.
Is it "no low end"? So all the energy is available to scorch your ears with higher frequencies? Is it human hearing or perception?
-
- zen recordist
- Posts: 6686
- Joined: Wed May 07, 2003 11:15 am
alas i am not nearly that powerful.
i dunno if i can explain it very well, but look at the speakers (and the space between them) as "a box".
now, when you're mixing, and listening to your stuff in isolation, you can have any size box you want. if you want to have the bass super booming, and the snare so loud you wince every time it hits, you can.
and it might sound super awesome like that, in your room and on your monitors.
but if you take that mix out of the room, and play it anywhere else next to any other record, it's going to sound like a tiny, yet muddy, little thing. it's gonna be super quiet compared to most everything else, not just modern loudness war casualties, i mean any reasonable rock record.
if you don't care how your record sounds next to anything else, It's My Art and Fuck You, that's totally cool. most people though, me included, want their stuff to at least be in the neighborhood of the ballpark of other stuff in their collection.
so, a big part of mastering is making everything fit into the same size box.
a classic loudness war mastering technique is rolling off the subs/low end, and boosting the high mids. rolling off the subs buys you a little more headroom, boosting the high mids gets you more apparent volume (remember our ears are most sensitive around 3k).
of course, too much of this gets you a weak-ass, thin sounding record that hurts to listen to. but hey it sure is loud.
anyway, getting back to our make-believe box....what's going to be easier to fit in there.....the little girl with her ukelele or our metal band of viking warriors with their huge downtuned guitars, monstrous drums and bong-rattling bass?
we have to cram our vikings in there while the little girl can stand up straight and still have room to run around.
i think it's really two things mainly:
1. less dense sounds louder than more dense. counter-intuitive, but lots of things in acoustics are.
2. our ears are just more sensitive to high frequencies. think about seeing a rock band live....do you ever have a problem hearing the tambourine? nope. in a recording scenario, has anyone ever had to turn the triangle track UP? nope.
it's all that stuff.Is it "no low end"? So all the energy is available to scorch your ears with higher frequencies? Is it human hearing or perception?
i dunno if i can explain it very well, but look at the speakers (and the space between them) as "a box".
now, when you're mixing, and listening to your stuff in isolation, you can have any size box you want. if you want to have the bass super booming, and the snare so loud you wince every time it hits, you can.
and it might sound super awesome like that, in your room and on your monitors.
but if you take that mix out of the room, and play it anywhere else next to any other record, it's going to sound like a tiny, yet muddy, little thing. it's gonna be super quiet compared to most everything else, not just modern loudness war casualties, i mean any reasonable rock record.
if you don't care how your record sounds next to anything else, It's My Art and Fuck You, that's totally cool. most people though, me included, want their stuff to at least be in the neighborhood of the ballpark of other stuff in their collection.
so, a big part of mastering is making everything fit into the same size box.
a classic loudness war mastering technique is rolling off the subs/low end, and boosting the high mids. rolling off the subs buys you a little more headroom, boosting the high mids gets you more apparent volume (remember our ears are most sensitive around 3k).
of course, too much of this gets you a weak-ass, thin sounding record that hurts to listen to. but hey it sure is loud.
anyway, getting back to our make-believe box....what's going to be easier to fit in there.....the little girl with her ukelele or our metal band of viking warriors with their huge downtuned guitars, monstrous drums and bong-rattling bass?
we have to cram our vikings in there while the little girl can stand up straight and still have room to run around.
i think it's really two things mainly:
1. less dense sounds louder than more dense. counter-intuitive, but lots of things in acoustics are.
2. our ears are just more sensitive to high frequencies. think about seeing a rock band live....do you ever have a problem hearing the tambourine? nope. in a recording scenario, has anyone ever had to turn the triangle track UP? nope.
I try not to master my own stuff if I can at all help it, but when I do (for demo purposes, usually) I try to use my ears and not watch the meters for RMS. Peak, sure, but that's already covered from when I've completed my mixing anyway. If you go by meters for RMS you can get a lot of really weird anomalies in perceived volume from low end frequencies and the difference between very clean, high headroom sources and very distorted/compressed sources. Nothing new, it's been well covered in this thread already, just adding to the chorus. Use your ears might be the standard response, but definitely for good reason.
-
- tinnitus
- Posts: 1094
- Joined: Fri Oct 20, 2006 7:52 am
- Location: Washington, DC
In case you still are looking for anything like a number, after I use my ears and get it to sound like I like (again, just for my own music and my bandmates), it's usually between -12 to -6 dBFS RMS, which is fairly hot, it seems. I've also figured out something that's probably old news to almost everyone reading this: marketing copy seems to say that more advanced processing can help your mixes sound louder, but my experience so far has been the opposite, and that less processing of any kind makes my mixes sound as loud as possible, as well as just better overall.
-
- zen recordist
- Posts: 6686
- Joined: Wed May 07, 2003 11:15 am
^^^^ yes. cleaner, less processed/distorted/saturated mixes are much easier to make loud.
as bob ohlsson has said a million times elsewhere: "distortion accumulates." all the typical loudenizing stuff we do in mastering is basically distortion in one form or another, so cleaner mixes hold up better to the abuse than more distorted ones do.
of course, if you're going for a blown-out sound, then blow it the fuck out and don't worry about it, but if you're not trying to be the next dave fridmann, cleaner is better.
as bob ohlsson has said a million times elsewhere: "distortion accumulates." all the typical loudenizing stuff we do in mastering is basically distortion in one form or another, so cleaner mixes hold up better to the abuse than more distorted ones do.
of course, if you're going for a blown-out sound, then blow it the fuck out and don't worry about it, but if you're not trying to be the next dave fridmann, cleaner is better.
- Snarl 12/8
- cryogenically thawing
- Posts: 3511
- Joined: Sat Dec 20, 2008 5:01 pm
- Location: Right Cheer
- Contact:
-
- zen recordist
- Posts: 6686
- Joined: Wed May 07, 2003 11:15 am
Holy fuck! Ya think? I'm not mixing metal, but I do horrible distorted pure noise all the time and never really come close to -6 overall.GooberNumber9 wrote:...usually between -12 to -6 dBFS RMS, which is fairly hot, it seems...
I usually get it sounding good and full in Reaper and then bounce and bring it up in SoundForge. The Normalize process has a little button says "Scan", and it tells you the level of the loudest peak as well as the overall RMS level of the file. The difference between the two is the crest factor which is what we're really talking about here. I shoot to have that somewhere between -14 and -12, though some things want to be a bit wider (-16 or so), and some a little shallower. I guess I'm actually a little conservative in this respect, but my mixes sit pretty decently in shuffle mode.
What I actually do is look at the waveform (zoomed out to see the whole song at once) with this info in mind, and try to eyeball how much of it is going to get smashed when if I try to bring it to that range. It usually works out that there's two maybe three "rows" of peaks. There will be one or two big quick peaks that stick up sometimes several db above anything else. These are truly aberrant - sometimes just weird glitches, but more often just the point where everything accidentally pushes the same direction at t he same time - all they do is waste headroom, and can usually be smashed down without hurting anything at all. Then there will sometimes be a second row of peaks that poke up above the bulk of the wave by another db or two. Super hot snare hits, or a spot where just a couple instruments are pushing together a little two far. Depending on how common these are, they can usually be squished down to buy another db or two, but you sometimes have to be careful, maybe a little less aggressive than with those big aberrations. Then there's the "meat' of the project where all the normal dynamics things happen. You can even sometimes shave off a db or two at the top of this, but this is where you really need to be careful to avoid massive distortion, pumping, breathing, and all the other issues that over-compression brings.
So, if I knock just the aberrations down to that second row, what kind of crest factor does that give me? If it's good enough, I'll usually just tell SF to do RMS normalize with "Compress to avoid overs" turned on. (This forces the RMS level to whatever you ask, slams the peaks up to 0dbfs, but tries to be a little more polite to the peaks that would go over. Still can cause distortion, but for just one or two quick peaks I can usually get away with it. I usually follow this with a Peak Normalize down to -0.6 to leave room for "intersamples", cheap DACs, and/or mp3 conversion) If I have to get too far into the second row, or worse into the meat of things, I go back to the mix.
Of course, I then listen to make sure that whatever I did didn't fuck things up too badly.
And yes, I have studied a number of different files by scanning them the same way in order to get a sense of perspective on where folks are putting things, and what different crest factors kind of sound like in different contexts.
-
- zen recordist
- Posts: 6686
- Joined: Wed May 07, 2003 11:15 am
if we must talk numbers, most records that everyone thinks sound good are probably sitting between say -14 and -7.
most modern super-smashed stuff is at or above -6 pretty much the whole time. if it's been meticulously tracked and mixed, it can still sound sort of ok. most things don't sound good that loud IMO.
some stuff can go well over -6 and sound fine...i have a couple neurosis songs amongst my references on my computer, and those get up into the -4, -3 range on the really crazy parts. they sound totally great, and they're not "loud" compared to lots of other things. again, lots of low end, sustained low mids are what's pushing the RMS up.
some 80s stuff is down in the -16 to -11 range. you either think "wow, the fact that this isn't limited really draws my ears into the mixes, how refreshing!" or "man these drums sure are sticking out", depending.
breeders "pod" is probably at like -20 but still kicks the ass of pretty much everything, you just have to turn it up.
most modern super-smashed stuff is at or above -6 pretty much the whole time. if it's been meticulously tracked and mixed, it can still sound sort of ok. most things don't sound good that loud IMO.
some stuff can go well over -6 and sound fine...i have a couple neurosis songs amongst my references on my computer, and those get up into the -4, -3 range on the really crazy parts. they sound totally great, and they're not "loud" compared to lots of other things. again, lots of low end, sustained low mids are what's pushing the RMS up.
some 80s stuff is down in the -16 to -11 range. you either think "wow, the fact that this isn't limited really draws my ears into the mixes, how refreshing!" or "man these drums sure are sticking out", depending.
breeders "pod" is probably at like -20 but still kicks the ass of pretty much everything, you just have to turn it up.
I never, nowadaze, mix into a compressor, but do use a limiter on the two-mix.
I mean, I use a compressor a lot in tracking, and frequently on vocals and sometimes other stuff at mixdown.
But I don't mix into a compressor (my mixing is ITB), so I'm essentially getting the last of my smash in volume from the software limiter, (hopefully transparently) and I'm liking the sound consistently in that -13 to -15 area, which is obviously a lot lower than the numbers above.
What brought this issue to the fore, FWIW, is a good friend of mine is Svengali'ing some rappers, and his critique of my mixes "lacking professional gloss" seems mebbe concerned with RMS, (as well as my limited use of additive EQ and obvious time-based stuff.)
I mean, I use a compressor a lot in tracking, and frequently on vocals and sometimes other stuff at mixdown.
But I don't mix into a compressor (my mixing is ITB), so I'm essentially getting the last of my smash in volume from the software limiter, (hopefully transparently) and I'm liking the sound consistently in that -13 to -15 area, which is obviously a lot lower than the numbers above.
What brought this issue to the fore, FWIW, is a good friend of mine is Svengali'ing some rappers, and his critique of my mixes "lacking professional gloss" seems mebbe concerned with RMS, (as well as my limited use of additive EQ and obvious time-based stuff.)
-
- zen recordist
- Posts: 6686
- Joined: Wed May 07, 2003 11:15 am
it's probably more eq-related. my stuff lacks professional gloss too. i tried, but it sounds better to me just being all crunchy midrange.
but it could also be that he's just used to hearing stuff way more smashed than your stuff is. that pressed against the glass sound of heavy clipping/limiting constitutes "gloss" for a lot of people i think. i don't really agree but i can see what they mean.
but it could also be that he's just used to hearing stuff way more smashed than your stuff is. that pressed against the glass sound of heavy clipping/limiting constitutes "gloss" for a lot of people i think. i don't really agree but i can see what they mean.
Who is online
Users browsing this forum: No registered users and 45 guests