Question- Does the volume of individual frequencies cause clipping at the master bus or does a song clip because of the summed volume across the frequency spectrum?
Context- I'm trying to make my recordings as loud as possible without abusing master compression. I know that a good arrangement can make a song appear louder. So I'm trying to make sure that the music is properly balanced in the frequency spectrum in attempts to make the music appear "louder."
In other words:
If a song is near clipping because it has lots of volume in the bass range, I know I can't add more bass-heavy instruments (from an understanding of good composition), but can I still add treble information without causing the master volume to increase? In other words is the clipping volume point a sum of all of the volume levels across the frequency spectrum or is it a scenario where one frequency range can clip while another range still has room to grow.
P.S. this is my first post on this lovely forum, even though I've been reading it for years. Please be gentle :)
It's the waveform shapes (amplitudes) not the frequency content of a sound that can exceed a limit and be clipped. Adding a sinewave of amplitude -5dBFS to another (uncorrelated) one at -5dBFS will will give a sum waveform that clips in places. This is independent of the frequency of the waveforms, assuming they are not from the same source.
First off, thank you for your response. Secondly, although I'm inclined to believe you, a quick experiment I just did seems to point in the opposite direction:
I started with one track that has a white noise generator running on it. I checked the EQ to make sure that it was completely flat. I then made a big boost around 1500hz. I then raised the overall volume so that the track was about to peak on the master bus. I then identified that boosts to any of the other frequency ranges would make only minor increases to master bus. However, a boost at 1500hz caused an immediately noticeable and dramatic increase to the master bus.
So the only conclusion that I can make is: the volume of individual frequencies really does matter. And if you're trying to add "more sound" to a mix, using a eq spectrograph analyzer across the master bus might be a good place to look for areas that can still be boosted.
Sorry if I misused any terms. I haven't thought about music by these terms since my "Science of Music" class back in college, haha. But it sure would be nice to know why I can add more bass to a track that is treble heavy. I would have thought that it was the summed volume of a piece of music that causes it to clip, not the peak volume of any of the frequencies.
Ugh... I'm going to bed. I hope this makes more sense in the morning, haha.
Your logic is sound where the Ways and Means may not be.
While it's important to record at proper levels for maximum depth and resolution, this really has not much to do with your mixing summation. High-frequency energy appears in a smaller spectral frame than does low frequencies which takes up most of the energy of reproduction. Low frequencies put the greatest demands on the playback system. Whereas the higher frequencies put the least demand on the playback system. This is how spectral energy in sound goes. And when doing a mix, it's usually a summation of these low frequencies that presents the most obvious overloads. High frequency transients can actually survive clipping due to their extremely short durations without really upsetting the perceived sound much. In fact it can actually be used as a tool in certain types of musical genres usually not legato. Drum peaks come to mind here. But not other stuff where it is wholly inappropriate.
The summing bus in software has a finite peak to limit unlike analog tape had. This does not make you are mix sound louder. Only good mixing technique and good proper postprocessing a.k.a. Mastering makes the difference. Sometimes folks will actually makes through a stereo bus limiter. Not something I usually do unless it's requested. But I do it after the fact in the Mastering when I am the one who is Mastering. But I first put a mix together that already plays well the way I like it to. I will then frequently do a number of different Mastering jobs to each one of the cuts either for improving consistency or creating a definite difference between cuts. So the Mastering process is much different from the mixing process. It all has to play well together. Sometimes I'll actually clip drum transients on purpose and then I will back down those clips to a proper level so as not to clip the out put digital to analog converter. While still retaining those transient clips on drum tracks which can improve the aggressive odd order distortion components. But that's on heavy rock 'n roll stuff and not on smooth sounding inspirational cuts.
So when mixing in software, you absolutely do not want to clip. Which means, you cannot attack the mixing bus with much over modulation. That just flattens everything out and makes it sound completely dull. No life, no dynamics, nothing good to listen to. It becomes an overbloated blob. And that's not loud sounding on playback. It's just a blob. A blob of sound that won't sound loud played back on-air of a radio station per se. That Loudness really comes down to how you mix it to sound to begin with, without overloading your stereo bus. Then the Mastering can ratchet up the apparent loudness level through good dynamics and EQ processing in a parallel configuration. And that's how you get loud sounding. It's all in the mix and the technique. And not in how you overblow your mixing levels.
If it was that easy to do, everyone could do it but they can't. That only comes through years of experience and technique. It's not a drive-through process that comes from a single plug-in button that you press. Though there are those that just simply use a Waves multi-band limiter as their final process. I do something similar with the multi-band limiter in the VST plug-in version of IK Multimedia's, T-Racks multi-band limiter and nothing else. And that alone makes a huge difference in and by itself when tweaked the way I want it to be. Which is never a standard preset. I don't use any presets. They are all over the top. Good as a starting point. Bad all by themselves. But I don't need those jumping off points since I know what I actually want and know how to get it through years of experimentation.
Spectral dynamic range processing is usually the key to good Mastering. Today, I think it's used more for the similarity of sound one used to get from the spectral processing used to every radio station just prior to the transmitter? These on-air processors not only included multi-band spectral compression and limiting, they also included downward expansion in those same bands, so as to control noise from compression flying all over the place. Yuck. But the software recording processors don't include any downward expansion capabilities. So I have to fight back with different types of downward expansion on individual tracks. That way, I can accomplish my multi-band processing without too much objectionable noise rushing up from tracks. And sometimes, I'll even add some downward expansion after the stereo bus limiting? Just to keep noise levels and reverb tails in check. And where I used to do this with KEPEX-1's I now do today in software, by drawing my own downward expanders from a compressor GUI.
I should also note that downward expansion and gating is generally a very precise threshold adjustment. Compression and limiting can be a lot more forgiving as to where you set your threshold at. Not downward expansion and not gating however. I only use gating on drums, never vocals. I only use downward expansion on vocals, never gating. I use downward expansion on keyboard outputs, guitar outputs and cabinets. Never gating. And it's amazing what this can do for you and your mixes. It adds a whole new element to your sound and mixes. It can negate that acoustics, bleed between other instruments. It can really tighten up a drum set to the point where your friends won't believe what you've done. And that's what makes for a fine recording and mix and loud Playback sound. And where overload is not necessarily a part of this discussion.
I have embedded a full mix of a full song on each one of my five pages of my website at Crowmobile.com . There is where you will hear recordings where all of the drums have been gated and vocals downwardly expanded with plenty of compression and limiting on just about everything throughout, individually done. There is no mastering here. This is the way it played on-air live. Except for a single cut which was a remix from an eight track recording and that was the really funky number which I think is on page 2. You'll also need Microsoft Internet Explorer 6 or newer or, Google Chrome unless you are running a third-party player, it won't play from Firefox, Safari, Opera or smart phones.
So there ya go.
Mx. Remy Ann David
Sabin333, post: 398963 wrote: So the only conclusion that I can make is: the volume of individual frequencies really does matter. And if you're trying to add "more sound" to a mix, using a eq spectrograph analyzer across the master bus might be a good place to look for areas that can still be boosted.
Use eq to make it sound subjectively better, not objectively louder. If you want it louder use a mastering limiter or other dynamics processing.
Thank you for the responses. Bouldersound- I suppose I don't necessarily mean louder. Perhaps I mean, more full? I just have this theory that people tend to like music that is evenly spread out across the full hearing spectrum. Like if a mix has equal amounts of bass, mid, and high end, it seems more interesting to the listener.
So mixing to me seems like a constant struggle of representing all of the frequency ranges, but also being aware that current music listeners like their music LOUD. So if you've got to be competitive, how do you emphasis all of the frequencies. Again, I'm rambling, but I'm realizing the only answer to these things is simply experience.
I've never sent anything off to a pro mastering engineer. I think after I hear my stuff through that process, I'll have a better understanding of how eq at the summation phase can be transformed through master bus compression etc.
Hi Sabin333, nice to see you posting!
Its Interesting to see you on the Hybrid Forum. I'm curious why you chose this forum and not the DAW Pro Audio forum? It sounds like you are ITB?
That being said, I actually asked myself this question more than once over the years. You are unsure at this point and its understandable. I actually get what you are thinking ( I think :) )
If I were to elaborate what I do, everyone here would desert this topic and Remy and I would go on for 2 days lol. Right Remy!!! so I'll simply echo what everyone has said, add a seed and leave lol.
Depending on what type of system we have and our style of music and studio, everyone comes at it unique to their goal in mind. So what works for one, may not work, apply or even make sense to another. All of us have different systems and styles of music too. A compressor does this, a limiter does this, eq etc.
I no longer mix the traditional way. My system is hybrid so I mix a session ultimately on the Master computer right before the upload. I do two versions. One brickwalled and one without, leaving that for a ME.
So I am sort of playing mastering engineer. I use a star monitoring system. In other words, I am able to listen to the entire process from pre or post (going in DAW1, going out DAW 2) right before the internet upload. My system is hybrid mastering quality but I don't consider myself a ME by any stretch. And my goal isn't all about sound quality either. Its actually about getting the most quality of level.
I have a limiter on the master bus but also use other compressors in stages throughout the creation of music. I can come at it any way you send it to me.
I use two computers and have an analog summing system connected between the two computer systems. No SRC either.
Most music I like, has more energy in the low freq and less on the top. High freq travel faster so it could also be why we push it this way and also why monitoring, room acoustic treatment and speaker and sitting placement is so damn important. If your room is bassy, you are going to mix thin sounding music and learn towards music that also sounds better in your room. This leads to no good..
Looking at music, The spectral graph is always peaking at the bass while the the top end, 20k, is barely moving a 1/4 way up. I see 20k moving but I actually use a LPF on my mastering EQ to roll that off. Low end converters have an especially difficult time there. So, if you are using cheaper converters, get rid of that. Most converters everyone is using I would roll that off. Plus, the internet doesn't sound that great up there either so whats the point. But there are tricks to this as well and this is where hearing and the right tools comes in.
I also use a LPF to roll off the subs. Low end takes a lot of energy. If I was mixing for DubStep, I would have more down there, less for rich acoustic music. At least a different accent there. So, not everyone is mentioning the kind of music they are referring too either. And when taking advise from people, we should all QUALIFY what they are doing before you apply or make purchases on something as gospel. But, gear is gear and it does what it is designed to do, if you know what you are doing.
If I am looking at the spectral graph and it was looking flat, I would be hating the sound indeed. Use your ears. If everyone was listening to music on ear buds , there would be no sense in pushing music with subs. But, if its mixed right, it should sound good on all systems. Some people are mastering music on very tiny speakers today, but the target is for the Ear Bud Crowd.
Loudness Wars are inspiring new ways to do things. I'm still trying to figure this out. I'm ordering tiny speaker tomorrow.
This is how I see it today... Tomorrow may be different. I may end up ITB someday again.
As Bolder said so well.
Use eq to make it sound subjectively better, not objectively louder. If you want it louder use a mastering limiter or other dynamics processing.
It is a red herring to think that one needs to fill the spectral field. If all the freqs are constantly filled up then the ear becomes tired and disenchanted and shuts down. Think about blending paints. You start with primary colors to make secondary and then tertiary colors. If you put every hue and shade into the mix you get something that looks like cat poo. At some point one just has too much going on. What is the correct balance of color? That will depend on the medium. Is your "canvas" wood or glass or paper or actual canvas? Are you painting with water color or oil? Color and sound do not have a one for one crossover but since many people are visual this may help to describe the conceptualization.
Good point, explanation John,
I actually turn off lights, deactivate most meters a lot of the time. I find it all distracting. I never use spectral meters for mixing. trust the ears .
TheJackAttack, post: 399041 wrote: It is a red herring to think that one needs to fill the spectral field. If all the freqs are constantly filled up then the ear becomes tired and disenchanted and shuts down. Think about blending paints. You start with primary colors to make secondary and then tertiary colors. If you put every hue and shade into the mix you get something that looks like cat poo. At some point one just has too much going on.
This is brilliant advice. I didnt' think of it this way. But now that you mention it, this makes perfect sense. If you fill out the spectral field you get closer to a flat signal, which is white noise. So I guess it's the absence of frequencies, and the creative usage of these "colors" that makes the music interesting. I can't believe I didn't realize this sooner.
Thanks guys! :)
I sort of look at the real time spectral display as billowing white fluffy clouds on a beautiful day. Whereas the stuff that is usually Pro-Tooled to death and is super hopped up, the spectral display usually looks like an overcast day where you are socked in. And you can plainly see that the sound has nowhere to go. One is done for stylistic and competitive reasons the other one is done for artistic reasons. And neither the two shall meet.
I just want a piña colada on a nice hot sunny beach somewhere.
Mx. Remy Ann David
All great info and advice from my colleagues here.... the only thing I would add, as a place to start, might be for you to gain a little more knowledge on "average" frequencies of certain instruments.
As remy and the others mentioned, low end freqs carry the bulk of the energy, which, can be good, but... they can also add frequencies to a mix that aren't doing anything but adding useless - and sometimes harmful - energy to your tracks and mix. Reducing certain frequencies in the low end spectrum can be very helpful because it can suddenly add more definition and clarity to those tracks you have, and, can allow for more amplitude as well because you're not including frequencies that do hold energy but don't do anything productive for your mix.
For example, and this is just a rough example so everyone put your guns down LOL... the average female singing voice doesn't really carry much useful frequency energy below 200 Hz or so ... this is an "average"... it could vary a bit... but to have frequencies of say, 80 hz playing back with that track aren't really going to do much more than to include energy that doesn't do anything to help that track, and in fact, it could actually hurt, because you maybe including things on that track that do "live" in that frequency range; furnace noise, traffic, it all depends on what environment you recorded the track in, of course...
Anyway, knowing the average ranges and bandwidths of the tracks you are mixing can help out a lot when it comes time to start using reductive EQ. Not much point in trying to reproduce 70 hz if you're recording a flute, is all I'm saying.
IMHO, of course.
DonnyThompson, post: 399078 wrote: they can also add frequencies to a mix that aren't doing anything but adding useless - and sometimes harmful - energy to your tracks and mix.
Yes, thank you DonnyThompson. I've been thinking about this a lot recently. I've always remove 50hz-80hz depending on the instrument. But recently I've also heard that there can be something called Psychoacoustic Masking at other frequencies. I might start another topic just to ask about that. Because its clear to me now that using subtractive EQ at other frequencies besides the sub 80hz, can make the rest of the mix clearer. Great point.
For instance, psychoacoustic masking can pertain to a track that might be a little too hissy. That noise won't be heard when there is another sound source louder over top of that noise. Then you don't hear the noise even though it is actually still there. Psychoacoustic masking can also be obtained by utilizing downward expansion and/or noise gates. These devices can actually quickly lower volume levels that can help to eliminate an otherwise not so flattering room acoustic signature. And that which will also tighten up the sound of a mix by changing the acoustic structure of the recording through the downward expansion, gating and dynamic range compression and limiting. Which can also help to mask other lower level audio abnormalities.
Another example of psychoacoustic masking would be the addition of artificial reverberation. Reverberation can frequently mask lower levels of audio distortion making those elements virtually inaudible of any audio aberrations. So sometimes a bit of reverb can mask some of the distortion elements that may have been poorly recorded. Though this only works to a certain point where you cannot mask something that becomes too horrendously obvious.
So psycho and acoustics can have varied meanings and strikingly different results when you put on a plastic Halloween mask and then all bets are off. The rubber masks work even better LOL. And it sounds really amazing when you put one of those over your microphones. The psychoacoustic masking then makes it sound like you are listening with a rubber mask on LMAO. But that only works well on Halloween.
I'm going to the store to look for a pretty new psychoacoustic mask.
What a girl has to do to sound good...
Mx. Remy Ann David
Another example of psychoacoustic masking would be the addition of artificial reverberation. Reverberation can frequently mask lower levels of audio distortion making those elements virtually inaudible of any audio aberrations. So sometimes a bit of reverb can mask some of the distortion elements that may have been poorly recorded.
Just be careful of the "Barry Manilow" reverb... you know, the kind that you can still hear an hour later. LOL
And no, I don't hate Barry Manilow. I just hear that stuff now and wonder if the reverb isn't maybe still happening in real time from the original FX aux bus from over 37 years ago... LOL
Hey... I was an assistant engineer for one of the Barry Manilow, five different versions of Copacabana at Media Sound, NYC back in ' 79. They cut it in the big studio A room which was the original sanctuary of the 58th Street, Baptist Church. And while most of that was EMT plate, it's certainly reflected an element of that large room acoustic. But definitely mostly EMT. And the ladies love it. Recorded and mixed on the Neve and Altec Lansing 604 Big Reds with the Doug Sax crossovers. Recorded and mixed by Michael DeLugg. Some folks may remember the GONG Show "with Milton DeLugg and his band with a thug" which was Michael's dad. And then he ended up being David Letterman's music mixer when David moved into what used to be the Ed Sullivan theater. And always be reminded of those Barry Manilow recordings for the rest of his life LOL. I wonder if he's ever heard the end of that... reverb LOL? And why haven't those multi-track masters been remixed and re-released like so many other things today? But would Barry ever sound the same again? I don't think so? Think about how ridiculous his songs would sound without that EMT LOL? I think he would quickly become forgettable? I mean if the Moody Blues could use all of that EMT plate, why couldn't the plate be used as much on Barry Manilow LOL? It's only fair with how much those things cost to get your moneys worth out of them LOL. I mean that EMT plate cost nearly $5000 in mid-1970 dollars. And which would have been more than three or four times the cost today. And we're a Neumann 87 was still only something like 600-$800 new back then. And which were used as drum overheads about 10 feet overtop the drum set to get all of that big sanctuary room sound. So that may have actually been enhanced with quite a bit of the acoustic signature from that room? Take out all the pews and stick a drum set in the middle of the room and not much EMT plate may have been used? They were gating everything with KEPEX-1's so maybe it wasn't the plate at all LOL? I wasn't in the mix session. Only tweaked the 24 track machine and the mix down machine. And I really don't know which one of the five Copacabana versions I worked upon, that day? And Barry was a sweet guy in more ways than one LOL. So well mannered and so charming. No wonder all of the old ladies loved that nice gay kid?
A mother would be proud (you have to say that with a New York Jewish accent like you just came back from a shopping spree at Bergdorf Goodmans)
Mx. Remy Ann David
But would Barry ever sound the same again? I don't think so? Think about how ridiculous his songs would sound without that EMT LOL?
LOL... yeah, I'm pretty sure the real time reverb trail for 1976's "Could It Be Magic" finally just ended last year. The verb on that track made the drum reverb at the end of Simon and Garfunkel's The Boxer sound like it was recorded in an anechoic booth. LOL
By the way and seriously, how big of a room did it require to hold that EMT plate? Those things were massive... and heavy.
They weren't as heavy as they were large. Approximately the size of a queen size box spring mattress. So you needed at least 6' x 10' room that had been literally " soundproofed ". And shock mounted from the floor. Some studios actually hung them from the ceiling rafters by springs, in storage areas elsewhere in the building. I liked mine to be more readily available for service purposes. The one I had in my last home studio, had to go down vertically into the basement OMG. It took four guys. It only weighs about 250 pounds. But the plate has to be loosened on its tensioners before shipping. It then has to be re-tension when you put into service and without moving it around. So you really had to think twice about where you put them.
I also loved the stories about those technicians have found that awful sticky coating all over the plate. So they cleaned it off. And then the plate rusted LOL. Plenty of great sounding rusted plates out there however. And plenty of underwhelming sounding plates in near perfect condition. No two sound the same. My last plate was the Studio Technologies, stainless steel plate. Talk about bright. It was about 50% brighter than the EMT 140 ST. So if you liked that sizzle, there was sizzle for days. I had the remote for my EMT but I didn't for my Studio Technologies. So I would have to go into the room and tweak the decay time by hand.
The density of those plate reverbs just cannot be reproduced by any algorithm or DSP chip. Not yet and probably not in our lifetime? I was always rather intrigued by EMT's 240 gold foil reverb. I get the impression they didn't last terribly long? I think I've seen all of one totally broken one for parts only once on eBay? I actually thought about getting one of those for my truck? I just gave up my AKG BX-20 E along with my Neve sale to Fabrice Dupont. That spring would have died on the first big bump. It remained in my home studio after the plates. Those were just so fragile though. So it was really disappointing to ship set off to AKG for a broken support spring, only to receive it back with a new broken support spring in the other channel. AKG told me that even though it had those lovely recessed chrome plated handles that it was never intended for any kind of portable applications. And it wasn't designed to be moved around a lot either.
So it's no wonder that we have embraced the cool sounding but density lacking silicon versions that you can carry under your arm.
I dream about EMT plate reverb
Mx. Remy Ann David
Remy, do you have a picture anywhere including the guts of one of these?
If I do have any pictures, they are certainly stuffed away somewhere? There really wasn't much to look at on these beasts. Rather underwhelming considering the original cost.
On part of that rectangular metal frame, there was an additional, effectively, a large piece of sealing tile. Through additional supports and hinges, this basically large piece of acoustic tile could move towards or away from the metal plate. It would get within 1/8 inch from the metal plate for the shortest decay time, but never touched the plate ever. Moving it as far as the mechanism would allow, away from the plate, approximately 6 inches, it would give you a full 5 second decay. One thing you could not do, remote control or not, was to be changing the reverb decay time.
I only played with the transistorized EMT 140 ST and never played with the earlier tube powered original 140. I would still like to build my own if I ever have the opportunity to have a place to put it? And yes, I would go stainless steel this time again.
The question is, What came first? The metal bed frame or the sheet metal?
Mx. Remy Ann David