Some questions about using Reaper, Mixing, and Mastering

[silky smoove]

You're still ignoring simple math. It's truly just arithmetic. No secret sauce, no... Anything.

Your method has benefits that are theoretically (better signal to noise ratio). My method has practical benefits (leaving headroom to allow for any mix decisions you want to make). The key difference is that the theoretical benefits that you're gaining are in a battle against something that's 60db down from full scale assuming a nominal noise floor. That works out to 64 times as quiet as something that peaks at 0dbfs. Do you really worry about something that low in the mix, or do you worry about clipping (either on the way in or by running out of headroom in the mix)? Even if you have to raise something by 20db through any combination of gain, EQ (which is just multiband amplitude), compression, etc., then you're only getting to 16 times as quiet as a 0dbfs source. Do you know how that translates in a mix? It translates to inaudible. By all means worry about it on your own, but don't pretend that it's a better way to operate for someone who's looking for basic mix instruction (no offense intended to the OP on that point).

To further that point there are some nicer modern interfaces that approach a noise floor that's 80db down! That's a whopping 256 times quieter than 0dbfs. Even if we weren't already there for all practical purposes, the latest and not even greatest batch of interfaces renders this such a non-issue that it's staggering.

 

[Omega Monkey]

Noise is real. Your theoretical need for 10s of dbs of headroom over your average peaks is kind of ridiculous. If you're recording a symphony orchestra playing Rite of Spring, ok. For western rock/pop/jazz/etc... stuff, probably not necessary, I think it's kind of hilarious that your whole rationale is "I might need to get as loud as you track in the first place, so that's why I need to record quietly!", it just doesn't really make sense. I didn't really have any headroom issues on the album I posted above. I think there are MAYBE 100 samples on the whole 40m album that clipped in the end. If that. I don't think we used anything higher than about 5:1 compression. You should be tracking the sounds you want to hear in the mix, so why do you need 20db of headroom per track?

Anyway, noise is not the hugest issue, but it is something to be concerned about, and I don't think I've ever listened to a recording I've done and thought, gee, this is just way too loud, I wish I'd tracked it 15db quieter. I'm not ending up with Californication or Vapor Trails here. No, it's usually the reverse. You're competing with modern mixes and trying to get it to be as loud as reasonably possible without totally overdoing it. And again, you can always carve away, you can't really add back what isn't there, and even when you can, there's a cost involved (usually in the form of noise).

Anyway, you're just parroting some guy you heard on the internet, so this is going nowhere. People will read the thread and make up their own minds.

 

[rufus.K]

This discussion is moving away from the OP's issue, however interesting

 

[silky smoove]

I think it's kind of hilarious that your whole rationale is "I might need to get as loud as you track in the first place, so that's why I need to record quietly!", it just doesn't really make sense.

Your interpretation of it doesn't make any sense is the problem. What you just described would be as if I were adding gain in post with no purpose. The reality is actually quite common. You record a modern rock band for instance. The first thing they want when you send them a rough mix for comments is more bottom end on the kick. You add some for the next go-around. They still want more. They want more bottom on the bass too. By the time you're done it sounds like the kick is from a dance track and you've slotted the bass at about 80hz to the point that it sounds like a hiphop song. It's not where I wanted the track, but it is where the client wants the track. You've now increased the low end content on your mix substantially. If you only tracked with 5-10db of headroom you might not have much room left on your 2buss to do any work on the rest of the tracks short of applying high pass filters everywhere to thin out the bottom end, which works fine until the guitar player starts crying that his tracks all sound too thin. What if there's a keyboard player using some super heavy Jordan Rudess style patches that have a ton of bottom end? You're out of room unless you've given yourself enough room from the start. You could always track hot and then knock everything down in post, but why bother adding a step to the mix process. When I'm doing something for a client I want to get a track where it needs to go as quickly as possible as opposed to applying gain/trim plugins to the top of every channel.

I think there are MAYBE 100 samples on the whole 40m album that clipped in the end.

Give me a slightly elevated but still inaudible noise floor over a single instances of clipping regardless of duration any day of the week.

You should be tracking the sounds you want to hear in the mix, so why do you need 20db of headroom per track?

For the reasons stated above. Additionally, as I've stated repeatedly with most modern preamps where changes in character and tone are not really imparted by the overall sweep of the input gain control you are tracking with the tones you want to hear in the mix whether it's at -20, -10, -5, or -0.1.

his is just way too loud, I wish I'd tracked it 15db quieter

Most, if not all, DAWs normalize the output to 0dbfs when bouncing/rendering a mix. When people talk about loudness in a mix it's a completely different issue. I don't understand the point you're making here.

I'm not ending up with Californication or Vapor Trails here.

As stated above, I don't understand your point. Those two albums had compression issues as a result of mastering choices that many people found distasteful.

No, it's usually the reverse. You're competing with modern mixes and trying to get it to be as loud as reasonably possible without totally overdoing it.

Again, you're not describing something that occurs as a result of tracking at -20, -5, or anything in between. Apples and oranges. Give me something with a whopping 40db of headroom, or something with 10db of headroom and I can make it slammingly, flat as a pancake loud, or too dynamic to listen to on a car stereo while driving down the freeway. You're conflating issues.

And again, you can always carve away, you can't really add back what isn't there, and even when you can, there's a cost involved (usually in the form of noise).

This is true in the analog world, but as soon as you enter the digital world you absolutely can add back what's not there. Quite easily too. In terms of added noise; Once again you're theoretically correct, or correct on paper if you prefer that term. The point I've been making is that in practice the little bit of added noise doesn't amount to anything and doesn't trump the benefits of going in at a lower level, which does have a host of practical benefits that I won't bother explaining for the umpteenth time.

Anyway, you're just parroting some guy you heard on the internet, so this is going nowhere. People will read the thread and make up their own minds.

Actually I've done tons of research on the subject and am speaking from several years of experience doing this type of work for a large cross-section of clients in a professional sense as well as for my own projects. I'm by no means some studio legend with decades of experience, but to suggest that I'm simply parroting things that I've heard without (this next part is me extrapolating from your statement) any experience or understanding of what I'm saying is flat out wrong.

This discussion is moving away from the OP's issue, however interesting

Agreed

 

[Omega Monkey]

Sorry, but if you're adding 10db+ of bass boost on a channel just because a client "asked for it" when it was already properly mixed, you're committing audio malpractice. The job of a mix engineer is to translate the words that a client asks for into the sound they actually want, but also to give them a solid result. If you're name is going on the project, you owe it to yourself and your reputation to make sure the mix stands up in the larger marketplace, while balancing that with client desires, which may sometimes be a tricky line to walk. BUT, again, you should be tracking the sounds you want in the mix as closely as possible in the first place. And if they didn't do that and change their minds later on, you just normalize the damn track down a few db to whatever headroom back that you need. Takes THIRTY SECONDS. You have a solution in search of a problem.

Also I don't believe you're correct about DAW's automatically normalizing mix renders to 0db. I don't remember which ones might have that as option in the rendering menu offhand. But even if you just normalize your overall mix, you're still adding gain one way or another. You may think I'm overstating the importance of avoiding noise, but you're way overstating the need for headroom. Adding 10db is making your track sound twice as loud. If you need to make your -20db track twice as loud, you probably didn't record it loud enough or with the right tone in the first place. Sometimes maybe you're going for a particular effect (like dramatically transforming some mundane instrument or sound into something else entirely), but if you know that ahead of time, you plan for it and track accordingly anyway. But for a standard rhythm guitar track, you're not going to add 8 db at 400Hz or something because that means you were nowhere close while tracking. If you're doing restoration, or a salvage job of what some other engineer wrecked, that's a completely different set of tools and perspectives than recording an album from scratch which is what I'm talking about.

 

[BonesawMcgraw13]

What about when you send all of the drums to a buss where you apply parallel compression?

Is this something a lot of people do? I thought about creating a separate bus for the drums since it's technically one instrument but something like 8 or 9 channels. The First thing I did what I mixed my track was solo all of the drums together and try to get a good mix between them. Then I slowly added in each instrument bringing them to match the drums. It would be nice to be able to use one fader for the drums though once I have them mixed evenly with each other.

 

[silky smoove]

@Omega Monkey I'm intentionally not responding to your gain staging comments as we're just going around in circles and I don't have the energy for it anymore. I'm happy to address your other points though.

Sorry, but if you're adding 10db+ of bass boost on a channel just because a client "asked for it" when it was already properly mixed, you're committing audio malpractice.

If the client is paying you and that's what they want... I had one group keep my name off of the release because I disliked the direction they pushed me to mix it. It happens. It's a business, so you do what the client wants. Well... most of the time, haha.

The job of a mix engineer is to translate the words that a client asks for into the sound they actually want, but also to give them a solid result.

Typically. Sometimes it's better to just do what they want verbatim as it will either show them how wrong they are, or will get the thing off your plate, get you paid and onto the next project.

If you're name is going on the project

See my post above. It was an awkward conversation when they asked why I didn't want to be associated with the project and I've never worked with them again, but it was for the best.

Also I don't believe you're correct about DAW's automatically normalizing mix renders to 0db. I don't remember which ones might have that as option in the rendering menu offhand.

Logic does, Cubase does, Reaper did the few times I tried to use it, Harrison Mixbuss does, Sonar does. I'm sure there's an option to disable that functionality when bouncing down a mix to stereo or surround, but it's the default state a lot of the time. There's really no reason not to amplify the loudest peak in a mix to 0dbfs when going down to stereo or surround. If you don't you're giving away mix volume for no reason.

But for a standard rhythm guitar track, you're not going to add 8 db at 400Hz or something because that means you were nowhere close while tracking. If you're doing restoration, or a salvage job of what some other engineer wrecked, that's a completely different set of tools and perspectives than recording an album from scratch which is what I'm talking about.

Depending on what you're after you might do a drastic EQ move like that. I would imagine it would be more for an effect than for a common rhythm guitar tone, but you never know. I was going to offer that a lot of times I'll get requests to mix something that was tracked elsewhere without my consent, but you got to that in the second part of your post. I've had people send something tracked with a Stratocaster in position 2 (out of phase between the bridge and middle pickups) and ask me to fatten it up because it's too thin. Seriously.

Is this something a lot of people do? I thought about creating a separate bus for the drums since it's technically one instrument but something like 8 or 9 channels

Parallel comping is a very common thing to do, especially on drums. Google it, watch some videos and get an understanding of the process. I tend to do it on all modern sounding mixes. I don't like it much in things like jazz, or very spacey less modern sounding stuff, but it definitely has its place. In rock or metal I would almost always do it. Some people will do another parallel comp with just the bottom end of the kick drum and bass guitar, but I've never been a big fan of that approach. I worked with an engineer at one point that did the kick/bass parallel comp thing and absolutely covered it in reverb. It sounded awful solo, but in the mix, when set relatively low, it gave the thing a very 90s Mariah Carey kind of vibe. I've never implemented it successfully, but when he did it I was very impressed.

 

[Omega Monkey]

Pretty much the first thing I do with a mix once everything is tracked is group the various instruments so I can apply stuff across all the drums, all the guitars, etc... when needed, and also just to organize everything so stuff is easy to find. I don't use parallel compression so much, but I do tend to put somewhat moderate compression on each individual drum (not much on the overheads and probably none on the ambient mic), and then put a lighter compression on the drum bus to kind of glue everything together. But I try not to use any more than necessary. With the album, we kind of overdid a few things in the early stages, but as the whole of each song started to come together we were able to cut all that back. In terms of processing, less usually is more.

 

[BonesawMcgraw13]

In terms of processing, less usually is more.

I'm learning that. I think I went a little overboard in the beginning because of all options I suddenly had. I was pretty modest with the strings though and most of the drums besides the kick.

group the various instruments so I can apply stuff across all the drums, all the guitars, etc... when needed

What would you add across the board for guitars? I feel like both of the guitar players in my band have a different sound.

 

[silky smoove]

What would you add across the board for guitars? I feel like both of the guitar players in my band have a different sound.

For me the first thing I do to a guitar buss is to apply a high pass filter with a shallow slope to get the bottom end out of the way of the bass player and the keyboard player's left hand. This will clear up a lot of mud. There's a sweet spot between clearing up mud and making things sound thin, so use your ears.

The next thing I do is apply a rather steep low pass filter at around 16khz to get rid of any fizz up top. From there I'll lower the LPF as necessary for a genre. Hard rock and metal might be pushed down as low as 10khz. Less heavy rock, more vintagey stuff, country, funk, disco, dance, etc. will probably have it around 14khz. Jazz and classical usually gets left at 16khz.

If I need to add some "glue" to the individual guitars to get them working as a unit I might apply some subtle compression. If I need to be heavy-handed to get a highly processed sound I'll use multiband compression, but more often than not just running the guitars through a single band compressor does it for me. Approach this very lightly. It's one of those things you shouldn't really notice until you bypass it.

If there are a lot of guitar tracks, especially if they were recorded with onboard reverb or in very different sounding rooms, I'll send the guitars to a dedicated reverb buss. This buss isn't used like a conventional reverb buss, but is more to get the each guitar sounding like it was in the same type of space. Some people like having lots of different spaces in their mix (this guitar was done in a concrete basement, this one was done in a wooden drum room, this one was done in a large cathedral, etc.), but I'm more of a fan of a concise approach in terms of spaces. I want it to sound like everyone was in the same type of room, just in different places within it.

 

[BonesawMcgraw13]

For me the first thing I do to a guitar buss is to apply a high pass filter with a shallow slope to get the bottom end out of the way of the bass player and the keyboard player's left hand. This will clear up a lot of mud. There's a sweet spot between clearing up mud and making things sound thin, so use your ears.

The next thing I do is apply a rather steep low pass filter at around 16khz to get rid of any fizz up top. From there I'll lower the LPF as necessary for a genre. Hard rock and metal might be pushed down as low as 10khz.

I actually tried this after watching a video from one of those youtube channels someone had posted. I ended up with an HPF at about 80hz and an LPF at around 10k like you just suggested. The video said between 7.5-10khz was what you were looking for, at least for this genre of music.

The same video mentioned using multiband compression on them as well which touches on your next point. I had already played with using a single band compression on both guitars and felt like I got decent enough results but honestly I don't really know a lot about how to use compression. I never use it on my bass or anything when playing live butI know a lot of people do. I just sort of messed with it until I got a sound I liked out of it. What are the advantages/disadvantages of multiband vs single band aside from the obvious of multiband giving you the ability of a little bit more control over what gets compressed?

 

[silky smoove]

The video said between 7.5-10khz was what you were looking for, at least for this genre of music.

7.5khz is a little excessive for my taste, but I can definitely see someone pushing it down that low with good result. One nice thing of getting it that low is that it will get the guitars out of the way of cymbals. Should make the top end of a mix crispier. Be careful applying an LPF that low on guitar solos as they can sound very dull without some top end sheen.

What are the advantages/disadvantages of multiband vs single band aside from the obvious of multiband giving you the ability of a little bit more control over what gets compressed?

Multiband compression is a killer tool in experienced hands, or potentially a mix destroying tool in anyone's hands. There are a lot of uses for it, too many to go into here, but as a rule I only reach for a multiband compressor if no combination of EQ and single band compression get me what I want. I'll reach for EQ, compression, distortoin, transient shapers, and all manner of other processing elements before I go to multiband compression. It's a great tool, but probably my last line of defense when mixing. It's absolutely critical in mastering, but at the mix stage it's a bit too powerful for most things. Your results may vary, and there are plenty of people that use them freely in a mix, it's just not really my thing if I can avoid it.

 

[BonesawMcgraw13]

Multiband compression is a killer tool in experienced hands, or potentially a mix destroying tool in anyone's hands.

This was my exact reason for asking. The video I watched mentioned it as if it's one of the first things you do when mixing heavy guitars. I find that in many cases things that can give you a lot of possibilities to make something really awesome have the same number of possibilities to make it terrible if you don't know how to use it properly. That's how I felt for a while when I switched from extract brewing to all grain. (beer is probably my next biggest hobby/obsession behind music) I know that's not really related to audio but it's the same premise.

7.5khz is a little excessive for my taste,

That was his suggestion, I think it ended up somewhere between 9.5-10 khz. I just moved it around between those frequencies until I found what I thought sounded best.

It's absolutely critical in mastering

You mentioned before to think of master like the forest with the mix being the trees. This analogy makes a lot of sense. You also said it was adding processing etc to the 2buss or main mix towards the end of mixing. Is it always necessary to add more to the 2buss when done mixing? I'm sure there are times when you're happy with what you have without doing anything else. Is mastering something that's always done or just to sometime to do some finishing touches? I'm sure it's pretty standard procedure when engineering professionally but maybe sometimes it can be as simple as a little extra EQ but be really complicated others if you're looking for a specific overall sound.

 

[silky smoove]

You also said it was adding processing etc to the 2buss or main mix towards the end of mixing. Is it always necessary to add more to the 2buss when done mixing?

That's just how some guys do it. In a pro setting the mix and the master would be completely different events. You'd create a mix, bounce it down to stereo or surround, and then send it off to a mastering engineer.

I'm sure there are times when you're happy with what you have without doing anything else. Is mastering something that's always done or just to sometime to do some finishing touches?

Alan Parsons describes the best situation ever as being the mastering engineering handing you back your mix without having done anything and saying that it needs nothing else. It almost never happens. I would, for all practical purposes, assume that once you're done mixing you'll want to have it mastered to end up with a polished finished product.

I'm sure it's pretty standard procedure when engineering professionally but maybe sometimes it can be as simple as a little extra EQ but be really complicated others if you're looking for a specific overall sound.

Correct

 

[BonesawMcgraw13]

So I'm slowly figuring out that recording has a lot to do with having it sound as much like the finished product from start to finish. A few people in here have said to capture a good performance and then make it better. So this made me think about a couple things related to actually recording the bass/guitars. I had heard that when recording bass specifically that you should record it flat and add any EQ you need later. I'm wondering how true this is. Obviously with the guitars very few guitar amps have a direct out like bass amps. So you're typically going to be putting a mic on the cab. So in this regard, you want the tone coming out of said cab to sound as close to the desired tone as possible. With bass you have the option to direct out much more often which means in theory you could make that signal pre EQ and send a mostly dry signal to the board to sculpt in the box. Some people do this, some people might send a post EQ signal, other may mic the cab or some combination in between.

In the recording I posted in this thread I used a combination of both. When we are rehearsing I go into my ampeg SCR-DI. Then I take XLR out of that direct into the console. The 1/4" out goes into a behringer peq2200 and then into a crest power amp which is ran into 2 eminence 15" speakers. Obviously the direct signal has been processed a little less before reaching the board. Should I be doing less to the signal before it goes into the board? Maybe it's a preference thing and the answer to this is very subjective. I honestly had a lot less trouble getting my bass tone right in the mix than anything else so maybe I should just keep doing what I'm doing. I was just curious how most people approach this.

 

[silky smoove]

Both approaches have their merits. If you don't know what you want as the final product I try to track as simply as possible. Sometimes I'll do everything DI and reamp later to taste. This is usually the case if I'm producing and engineering, but the client isn't sure if their simple acoustic guitar tube should end up as a pop song, something folky, indie rock, etc.

If I know what I want, I track everything as close to finished product on the way in as possible. Most people recording their own projects will prefer this method, but I always recommend taking a "safety" copy of everything (where possible) in case you paint yourself into a corner and can't get out during mixing.

 

[BonesawMcgraw13]

but I always recommend taking a "safety" copy of everything (where possible) in case you paint yourself into a corner and can't get out during mixing.

So maybe create another track that is coming out of the thru jack on my DI as a just in case? This is the easiest thing I can think of doing that won't affect how I sound in the room. I supposed I could also just record another track after the fact using the direct out, disengage the EQ, and put all the EQ controls on my bass at the same o clock. That would be more time consuming though.

 

[silky smoove]

Just throw a DI as far upstream as possible and assign it to an input. Track it along with your "wet" signal and mute it in your DAW. If you need to go a different direction and can't solve it ITB, just reamp the DI track as needed and you're good to go.

 

[BonesawMcgraw13]

So I could just throw another DI in front of my ampeg DI and take one signal to the board that other to my usual routing? I've got a Radial pro D2 that could probably work for this purpose.

 

[BonesawMcgraw13]

After looking at the D2, I'm not sure if that would work. It's a stereo DI so I don't think I can plug into one side of it and have it go out both outputs. I think it is just left and right. I found an old whirlwind DI that has a 1/4" out as well as an XLR. That might work better.