Go Back   Cockos Incorporated Forums > REAPER Forums > Recording Technologies and Techniques

Reply
 
Thread Tools Display Modes
Old 08-14-2019, 06:43 PM   #41
3buddhas
Human being with feelings
 
Join Date: Oct 2018
Posts: 54
Default

Many thanks, ashcat_lt, for taking time to write out these tutorials. I'm learning from them.
I like this idea of previewing the master fx chain at the individual project stage, make sure everybody's gonna be on best behavior come mastering. But turn off those fx to render the individual projects. Then use the same chain in the mastering project.
The tips on mastering for DR and RMS/LUFS are welcome. There's a lot in those subjects to study, and I'm just starting to float those boats. If you'd care to disclose some of your master fx and reasonings why you use 'em, I'd be an avid reader.
Cheers
3buddhas is offline   Reply With Quote
Old 08-14-2019, 09:15 PM   #42
ashcat_lt
Human being with feelings
 
Join Date: Dec 2012
Posts: 7,271
Default

I need you to take all this with a big grain of salt. I do sometimes get paid to record and mix for other people, but none of the things I have done are actually selling. I don't honestly believe this is really because of how they sound. My clients always seem satisfied if not just blown away, but...


So my "master chain" is usually just super simple:


ReaEQ - preemphais - shelving down the low end a little bit, sometimes pushing the harsh upper mids and maybe the way high "air", all meant to shape and focus the action of the following


ReaComp - lookahead leveler - RMS 500, Precomp 250, A/R 0, ratio usually 1.1:1 (now that it can go lower I might start messing with that), knee maybe 9-12db, threshold set so that we almost never really get there. This very subtly controls the big dynamic swings so that like the loud chorus is not quiet so much louder than quiet verse but also like the one beat where everything just happens to be pretty loud at the same time isn't quite so much louder than the other beats. Things kind of push other things around, and it tends to be what I think of when I hear the word "glue". It is a good, transparent way to decrease DR (and thereby increase perceived volume once normalized) without destroying the faster transients.


ReaEQ - deemphasis/final tone shaping - First basically undoing what the first one did. Together, they work very much like putting an EQ in the sidechain of the compressor, except that here they don't have to be - and usually aren't - exactly opposite. Since the low end isn't getting squashed quite so hard, it often doesn't need to be turned all the way back up, and I can even set different corner frequencies on either side if that works better. It's also just the last EQ in the chain, and I use it to kind of define the overall sonic character of the record. It's way too late to surgically fix individual frequency bands here. I consider it much more like the frequency curve of the master tape back in the analog days.


ReaComp - rail - all time constants at 0, ratio inf:1, knee like 6db, threshold set so that it almost never actually reaches it, auto makeup on, wet at -0.6db. It's a nice curvy soft limiter/saturator that just makes sure nothing ever goes out that's louder than -0.6dbFS. This is a number I chose long ago as a hedge against "intersample overs" that also happens to help with mp3 conversions all the the other things that can make DACs clip even though there are no actual samples going "over" 0.


I don't worry about True Peak and I don't actually care about LUFS. The actual peak sample sample level and the RMS have always told me everything that I couldn't figure out just from listening.


Edit - In certain rare instances where I feel it's necessary, I will put Otium BassLane at the front of this chain to force the real low end to mono. There's a lot of reasons for this, but mostly just to share the load between the two speakers and sort of open headroom for higher frequencies on each channel.
ashcat_lt is offline   Reply With Quote
Old 08-15-2019, 06:01 AM   #43
3buddhas
Human being with feelings
 
Join Date: Oct 2018
Posts: 54
Default

Thanks, ashcat_lt, for taking pains to type out another fine tutorial. If I find any big grains of salt around my house I'll be sure to take them, but for now almost all this is stuff I've not really examined before, so it's helpful. And these little primers you make will sit in the forum archives and I'm guessing will help others as well to get on the road to Reaper / recording self-reliance.

I like that your mastering fx chain is so Reaper-centric. I was worried you might come back with a hundred tools I didn't/couldn't own. I've set myself a kind of personal challenge to not spend money on this album, apart from the upcoming expense of cd duplication. So I appreciate the frugal sensibility. To date I've bought exactly four pieces of recording-related software, by far the most expensive being Reaper, with which I'm very very satisfied.

This afternoon, after the dentist, with my face all numb, I mean to dive into the tutorial and try out the chain on my projects. I like this idea of look-ahead leveling. This is something could be applied to shaping a whole life maybe. You look out at your future and see the potential hotspots or pinch points, the potential clips, and apply a bit of pacifying attention there.

I don't worry about TruePeak either because I don't know what it is! The total freedom of ignorance.

The EQ tips are understandably -- being that EQ is an art form -- poetic. I'll see if I can divine what's going on there. Every YouTube mastering tutorial I've watched has, when it came to EQ, pretty well mystified me. It's like you do a little harm with EQ then try to undo the harm with EQ. Hm.
Cheers
3buddhas is offline   Reply With Quote
Old 08-15-2019, 09:06 AM   #44
ashcat_lt
Human being with feelings
 
Join Date: Dec 2012
Posts: 7,271
Default

Idk about harm. In this case, the first EQ is meant to focus the attention of the compressors on the areas of the frequency spectrum that I want it to focus on. For example, by shelving down the low end, it lets low frequency transients have a little more impact and reduces things like pumping and breathing. But then that does in fact change the frequency balance, and you kind of have to put it back. Usually none of this is really doing much when I’ve got it on a full mix. Maybe a couple dB cuts or boosted, and the compressor barely kicking in. It doesn’t take much, though it’s so clean and transparent that you can push it pretty hard without really realizing.

I did not invent this pre/de emphasis technique. I’ve been using it one way or another for almost ever, but Dan Worrall made great vid about it recently:
https://www.youtube.com/watch?v=H-Gs-o39C5o

But I do things a lot differently from most other folks. If you use a multi-band compressor (ReaXComp), it pretty much just does all that or at least kind of defeats the purpose.

It’s only recently (since they fixed the knee, make sure Weird Knee is NOT checked) that I could use ReaComp as the saturator at the end. I was using a JS plug I’d written, but I’m pretty happy to dump that and use the native solution. It’s a little less convenient to get asymmetrical saturation, but for this usage it’s not really necessary and maybe not really advisable. I’ve bought a couple plugins, but I really try to keep it free and Reaper native when I can. I definitely don’t need any othe emulations. I don’t need my EQ to also distort and I don’t need my compressor to have its own EQ curve. What if I don’t like the decisions the coders made? I just build my own with chains like this.

The name of this preset is “rail” because analog gear cannot put out any voltage greater than its power supply. We often call that power supply the voltage rail or just rail. Reaper itself has no such thing. It will very gladly (try to) put out voltages billions of times greater than your DAC could ever create. When it does that, what happens is basically out of your hands and could be different on different playback devices. There will be distortion, but you won’t ever know for sure what the distortion will sound like for your individual listeners. So I’m taking control of that distortion. Many of us who used to mix analog used to rely on this sort of rounding off and limiting of the peaks. I actually used to choose boards based on how they sounded when they were slammed. I do use this on a lot of tracks/busses throughout a mix. Again, handling it at the lower level so there’s less to do at the master. On the mix bus this won’t usually do much. If you start to hear it, you’ve gone too far.

In fact, this whole chain is a lot like a channel strip for me. I drop it in all over and tweak the values to suit the situation. The lookahead leveler with a bit more ratio and a more aggressive threshold is great for taming vocals, kick drums, bass guitars, or pretty much anything that has a bit too wide of a dynamic range. In those cases there will often be a more traditional compressor afterwards, and the leveler helps that next one work more consistently.

But of course most of this relies on the fact that you are recording, mixing, and mastering, and can fairly easily go back and fix things at the more granular level so there’s nothing to do at the master. When you’re working with other people’s stereo mixes, it can sometimes take a whole lot more. It starts to lean more toward forensic restoration which is not something I particularly enjoy doing.
ashcat_lt is offline   Reply With Quote
Old 08-15-2019, 12:14 PM   #45
3buddhas
Human being with feelings
 
Join Date: Oct 2018
Posts: 54
Default

Outstanding. Thanks. Two things as I dive into this. Kindly,
1) Where is this rail preset? I've looked in ReaComp and ReaXComp. Didn't find it in the versions I have. Nor do I see Weird Knee anywhere.
2) I'm sitting in the dentist chair getting drilled, it occurs to me, "but wait, ashcat_lt, where is all the fancy mastering metering I've read about? Where's the tool that's gonna make sure your RMS/LUFS is maximized across the whole song, automate all that compression? Diagnosticize and prognisticate all your crest values? Are your tools called just ears? If so, I wonder if mine are seasoned enough.
Cheers
3buddhas is offline   Reply With Quote
Old 08-15-2019, 01:15 PM   #46
ashcat_lt
Human being with feelings
 
Join Date: Dec 2012
Posts: 7,271
Default

Quote:
Originally Posted by 3buddhas View Post
1) Where is this rail preset? I've looked in ReaComp and ReaXComp. Didn't find it in the versions I have. Nor do I see Weird Knee anywhere.
Oh sorry! That's my own preset. I described the settings above.
Quote:
2) "but wait, ashcat_lt, where is all the fancy mastering metering I've read about?
I mean, there's a meter on ReaComp and on the tracks and on the master. I do use Voxengo SPAN usually as a sanity check, and when I get to the point of mastering I will often run the SWS Analyze Loudness action to get an idea of how things compare to each other and whatever target I'm shooting for.

Too many people worry way too much about those things though. Get all upset about the difference between RMS and LUFS and caught up in that -18 rule of thumb as though it's some carved in stone god's going to smite you kind of doctrine. All that really matters is how it sounds and how it translates to other playback systems, and the numbers are just ways to help predict some of those things.
ashcat_lt is offline   Reply With Quote
Old 08-16-2019, 06:38 AM   #47
3buddhas
Human being with feelings
 
Join Date: Oct 2018
Posts: 54
Default

Fine answer, thanks. Exactly what I was fishing for. "Don't worry so much."

Quote:
Originally Posted by ashcat_lt View Post
and when I get to the point of mastering I will often run the SWS Analyze Loudness action to get an idea of how things compare to each other and whatever target I'm shooting for.
This sounds like the sort of thing I was reading about. Analyzing loudness. I've not tried anything SWS. For we programming-shy, might you instruct on how to use such a tool?

These lessons are helping me accomplish what I'm trying to accomplish.
Cheers
3buddhas is offline   Reply With Quote
Old 08-16-2019, 11:26 AM   #48
ashcat_lt
Human being with feelings
 
Join Date: Dec 2012
Posts: 7,271
Default

Download and install SWS extensions. Select audio items. Open the Actions List. Type “anal” into the search box. I find that amusing as hell every time.
ashcat_lt is offline   Reply With Quote
Old 08-16-2019, 12:20 PM   #49
3buddhas
Human being with feelings
 
Join Date: Oct 2018
Posts: 54
Default

Quote:
Originally Posted by ashcat_lt View Post
I did not invent this pre/de emphasis technique. I’ve been using it one way or another for almost ever, but Dan Worrall made great vid about it recently:
https://www.youtube.com/watch?v=H-Gs-o39C5o
Will look into this.

Quote:
Originally Posted by ashcat_lt View Post
But I do things a lot differently from most other folks. If you use a multi-band compressor (ReaXComp), it pretty much just does all that or at least kind of defeats the purpose.
Maybe this is saying that both multiband compressors and pre/de emphasis methods are ways to use EQ to direct compressors?

Quote:
Originally Posted by ashcat_lt View Post
It’s only recently (since they fixed the knee, make sure Weird Knee is NOT checked) that I could use ReaComp as the saturator at the end. I was using a JS plug I’d written, but I’m pretty happy to dump that and use the native solution. It’s a little less convenient to get asymmetrical saturation, but for this usage it’s not really necessary and maybe not really advisable. I’ve bought a couple plugins, but I really try to keep it free and Reaper native when I can. I definitely don’t need any othe emulations. I don’t need my EQ to also distort and I don’t need my compressor to have its own EQ curve. What if I don’t like the decisions the coders made? I just build my own with chains like this.
Chewing on this. What I gather from it, when a compressor squashes a peak, there's distortion. Maybe distortion and saturation are the same thing? (Would love a blurb on what is "asymmetrical" saturation.) Compressor and DAW squashing-algorithms vary, so amount and type of distortion/saturation are uncertain. Wanting to control that uncertainty, you cherry-pick components that make distortion. Seems confirmed in following paragraph.

Quote:
Originally Posted by ashcat_lt View Post
The name of this preset is “rail” because analog gear cannot put out any voltage greater than its power supply. We often call that power supply the voltage rail or just rail. Reaper itself has no such thing. It will very gladly (try to) put out voltages billions of times greater than your DAC could ever create. When it does that, what happens is basically out of your hands and could be different on different playback devices. There will be distortion, but you won’t ever know for sure what the distortion will sound like for your individual listeners. So I’m taking control of that distortion. Many of us who used to mix analog used to rely on this sort of rounding off and limiting of the peaks. I actually used to choose boards based on how they sounded when they were slammed. I do use this on a lot of tracks/busses throughout a mix. Again, handling it at the lower level so there’s less to do at the master. On the mix bus this won’t usually do much. If you start to hear it, you’ve gone too far.
Wanting to understand this. I'm thinking a Digital Analog Converter (DAC) takes ones and zeros and turns them into audible signals. Something you can hear on your monitors. If the ones and zeros say sonic boom, Reaper's DAC will try to create a huge noise, with possible ugly results on playback systems that can't handle huge noises. Maybe no problem on playback systems that can handle them. To avoid the worst case, upstream you try to avoid slamming the DAC with sonic boom commands. So you round off and limit. And try to use rounding/limiting tools that give consistent, desired results. And you seem to say, apply these even at the multitracks' individual track stage where needed to tame peaks. This last seems confirmed in next paragraph. And be conservative with the squashing on the individual project 2-busses.

Quote:
Originally Posted by ashcat_lt View Post
In fact, this whole chain is a lot like a channel strip for me. I drop it in all over and tweak the values to suit the situation. The lookahead leveler with a bit more ratio and a more aggressive threshold is great for taming vocals, kick drums, bass guitars, or pretty much anything that has a bit too wide of a dynamic range. In those cases there will often be a more traditional compressor afterwards, and the leveler helps that next one work more consistently.
Noted.

Quote:
Originally Posted by ashcat_lt View Post
But of course most of this relies on the fact that you are recording, mixing, and mastering, and can fairly easily go back and fix things at the more granular level so there’s nothing to do at the master. When you’re working with other people’s stereo mixes, it can sometimes take a whole lot more. It starts to lean more toward forensic restoration which is not something I particularly enjoy doing.
I expect to be bouncing between mix-projects and mastering project for a spell. The two stages will show up any shortcomings in each other, I'm guessing.

Many thanks for these little forum-inars. I find them a great, paced way to get a toe in these complex subjects.
Cheers
3buddhas is offline   Reply With Quote
Old 08-19-2019, 09:51 AM   #50
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

Quote:
Originally Posted by serr
on DA converters that run cleaner at HD vs SD
Is that really still a thing?

There is nothing that software upsampling can do with a given bit rate and depth format that can't be done in the DAC converter.

I guess some older DAC chips might work better fed like that, but these days how many of them are in rooms of sufficient quality, with sufficient quality speakers and amps for it to matter at all?
drumphil is offline   Reply With Quote
Old 08-21-2019, 09:52 PM   #51
ashcat_lt
Human being with feelings
 
Join Date: Dec 2012
Posts: 7,271
Default

Sorry it took a while to get back on this. I scanned your post, didn't see anything that I thought was dangerously wrong , and put off replying. Then I threw a 3 day music festival at my house/studio and I'm kind of just getting back to the "real world".



Quote:
Originally Posted by 3buddhas View Post
Maybe this is saying that both multiband compressors and pre/de emphasis methods are ways to use EQ to direct compressors?
Yeah. EQ in the sidechain does much the same thing. Each is different. In fact, they're only kind of the same if you squint, but yeah.


Quote:
...when a compressor squashes a peak, there's distortion.
I think what I was actually trying to say in the quote section is that compressors (analog and emulations thereof, at least) distort. If the compression was perfect (like we can do in digital), it wouldn't sound like distortion to us. But physical components have physical limits. If nothing else, that "rail" I was talking about. No matter how much you ask it to put out, it can't give you more than it's power supply gives it. Most of the time it can't even really give you all of that. So it gives you what you can, and if that's not as much as you asked it for, it's going to sound like distortion. So the compression circuit inside that box is wrapped between a couple amplifiers, and is also an amplifier itself, and each of these have real output limits.



Quote:
Maybe distortion and saturation are the same thing?
Distortion, overdrive, saturation, breakup, crunch, fuzz... Everybody has a different definition of every one of them. What I mean is that there is some sample-to-sample non-linearity.



I tried to multiply sample a by x gain, and I got a * x
I tried with sample b which is bigger than sample a, and I got b * something less than x.

Quote:
what is "asymmetrical" saturation
Saturation that's not symmetrical If the top of the waveform hits the limit before the bottom, or vice versa. People will tell you all kinds of things about even and odd harmonics, but we've had enough math for one post. In analog, the signal wiggles between and upper and a lower limit, and it is very often not exactly centered between them for one reason or another. People are used to hearing that sort of distortion, and it kind of is just a little more musical and interesting.



There's a bunch of these all through the signal path, and if you do it right you never really come near any of those limits unless you like the way it sounds when you do. The fun part is, though, that most "saturation curves" start rounding off at about -6db. That means that by the time you're half as loud as you can be, you're already not quite as loud as you should be.



Quote:
If the ones and zeros say sonic boom, Reaper's DAC will try to create a huge noise,
Reaper doesn't have a DAC. Your interface has a DAC. It has a physical limit as to how much voltage it can create. We don't care how many volts that is. All we need to know is that's all you get. We call that 0dbFS. Your interface (driver?) just doesn't even care about anything bigger than that. Reaper can conceive of numbers billions of times bigger than that, but whatever.

Quote:

with possible ugly results on playback systems that can't handle huge noises. Maybe no problem on playback systems that can handle them.
Every system will fail in a different way. Some are more pleasing than others.



Quote:

To avoid the worst case, upstream you try to avoid slamming the DAC with sonic boom commands. So you round off and limit. And try to use rounding/limiting tools that give consistent, desired results.
In order to be reasonably sure that we know how it fails, we fail before they do.
Quote:
And you seem to say, apply these even at the multitracks' individual track stage where needed to tame peaks. This last seems confirmed in next paragraph. And be conservative with the squashing on the individual project 2-busses.
We've said that a few times on this thread. If one thing sticks out a little too far at some point, we'd usually rather squash that one thing than just like everything.
ashcat_lt is offline   Reply With Quote
Old 08-22-2019, 10:01 AM   #52
3buddhas
Human being with feelings
 
Join Date: Oct 2018
Posts: 54
Default

Many thanks for the follow-up. It's my opinion that this subject matter is difficult. People who think it's easy maybe haven't delved, or maybe were born to a long lineage of delvers. Or some just worked at it a long time to get it. Seems you can accomplish recording tasks without understanding this theory maybe, but at some point it's probably better to understand it. Then if you're where I'm at you get to scratch your head a bit. As example ...

Quote:
Originally Posted by ashcat_lt View Post
The fun part is, though, that most "saturation curves" start rounding off at about -6db. That means that by the time you're half as loud as you can be, you're already not quite as loud as you should be.
Shades of Tolkien.
Bilbo was doing splendidly. This is the sort of stuff they liked: short and obvious. "I don't know half of you half as well as I should like; and I like less than half of you half as well as you deserve." This was unexpected and rather difficult. There was some scattered clapping, but most of them were trying to work it out and see if it came to a compliment.

But I'm quite sure there's a nugget in there.

Quote:
Originally Posted by ashcat_lt View Post
I tried to multiply sample a by x gain, and I got a * x
I tried with sample b which is bigger than sample a, and I got b * something less than x.
Grasping the algebra. Pondering the application of it.

Quote:
Originally Posted by ashcat_lt View Post
Saturation that's not symmetrical If the top of the waveform hits the limit before the bottom, or vice versa. People will tell you all kinds of things about even and odd harmonics, but we've had enough math for one post. In analog, the signal wiggles between and upper and a lower limit, and it is very often not exactly centered between them for one reason or another. People are used to hearing that sort of distortion, and it kind of is just a little more musical and interesting.
Smiling at the joke, appreciating the serious definition.


Quote:
Originally Posted by ashcat_lt View Post
Reaper doesn't have a DAC. Your interface has a DAC.
Obviously I didn't quite know that. Thanks.

Quote:
Originally Posted by ashcat_lt View Post
Every system will fail in a different way. Some are more pleasing than others.
Makes for rather an elusive target, yes? I suppose as in any art you have to get a feel for the sweet spot.

I've been busy elsewhere too and am just now applying the suggested mastering fx chain. Early days, but feeling encouraged about it. Wondering about a practical matter. How do I indicate for cd duplication where to insert the digital markers for song beginnings? So cd players can tab to them. The rendered master will have the silences between song tracks, but those can vary, short to long. Would seem a more precise marker is needed. And what if I run two songs together but want them indexed as separate tracks?

I very much appreciate this tutorial. Your time and expertise, ashcat_lt, is valuable.
Cheers
3buddhas is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 06:47 AM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.