Mixing down, when mostly in the box. How do you do it?

Recording Techniques, People Skills, Gear, Recording Spaces, Computers, and DIY

Moderators: drumsound, tomb

User avatar
farview
tinnitus
Posts: 1204
Joined: Tue Aug 31, 2004 1:42 pm
Location: St. Charles (chicago) IL
Contact:

Post by farview » Sat Mar 29, 2008 3:03 pm

@?,*???&? wrote: More interesting would be the question should one Noiseshape their 48KHz mix BEFORE printing to tape. Whereas it will certainly sound better, it will likely be done again at Mastering.
It wouldn't matter. The dithering would be swallowed up by the tape hiss so it's a non-issue.
@?,*???&? wrote:Also, could an argument be made that one should mix sample rates? In a scenario like this, work at 48KHz in the digital session and then commit mix to tape. What resolution will the mastering engineer work at?
It doesn't matter. Once you dump it to tape it's a completely different analog signal. Sending the tape signal through the converters is not the same as doing sample rate conversion. it's a completely separate thing.
@?,*???&? wrote:Double converting at 44.1Khz can sound like ass.
You wouldn't be double converting anything..

Professor
ghost haunting audio students
Posts: 3307
Joined: Wed May 07, 2003 2:11 pm
Location: I have arrived... but where the hell am I?

Post by Professor » Sat Mar 29, 2008 6:01 pm

Earlier, I read this post and then another in the Recording Techniques board and as I typed a reply to the other it seemed to be a really great fit for this one. So please forgive the apparent vanity as I quote myself and post about half of that response here as a thought to consider for this situation.
So that was the technical part of the answer - but I have a conceptual comment here too that relates to another thread I just read asking about 'In the box' mixing...
Folks who mix in the box, and most expecially those who have only ever mixed inside the box, tend to approach reverb in a very different way than those of us who have mixed across consoles, whether live or in the studio. It's really evident in Chris's response. And I've found that it's really evident in the mixes I hear.
See, when we mix across a console, we very rarely have 24, 32, or more individual reverb processors for our system. We might have only one, or maybe a couple of different processors, but rarely more than just a handful. As a result, we don't individually patch a reverb into every single channel... but we certainly want a little reverb on everything in the mix. What we do is to patch an auxiliary send out to a reverb processor (or two, or three) and bring the output of those devices back into the mix on another set of channels. Then we run along the Aux sends for each channel dialing a little or a lot of each instrument into this single, unifying, "room" or "space" we have created. Not much kick or bass, but more snare, overheads, guitar, etc. Maybe we'll put a single reverb on for a special effect, like a floor tom hit ringing for 6 seconds or something. But in general, there is a primary sound for the room so the resulting mix sounds more like a group of instruments performing in the same acoustical space.
In the box, things end up a little different.
Most people seem to decide that an instrument needs reverb and so they click on a plug for that individual channel. Maybe they will use similar presets which at least gets closer to a 'single room' sound. But often enough they will use a reverb on snare that they think sounds good for the snare, and then a single reverb for guitar, for vocals, for keys, for whatever might be in the mix. The result that I hear all the time is a mix that sounds more like a "tossed salad" of acoustic spaces, rather than a smooth, blended, "fondue" of a single space where all of our instruments were played.
We hear endlessly about how ITB mixes don't sound the same, don't "feel" right, are too clinical, cold, inhuman, etc. you can pick your favorite adjective. Well maybe the result isn't to mix across an analog desk or summing network, or to mix to tape, or any of the other attempts to add a unifying single color to a mix. Maybe it's as simple as opening an "Aux Track" on your DAW, assigning an internal aux buss to get there, and putting a single reverb on that bus to which each individual track will submit a little piece of its sound. I would even say that the more close mic sounds, the more direct inputs, the more synth & VST sounds, and the more mono tracks you have, the more critically important this concept is to unifying your mix. In those cases you are trying to combine sounds that never existed in the same acoustic space (or indeed in any acoustic space) and so creating a single space for all the "preformers" to "play in" together is an enormous step in turning that fruit salad into a smoothie.
There you go.
Forgive the apparent blasphemy as I suggest it might not be tubes, transformers, tape, analog boards, stacks of resistors, or some other audio voodoo. And forgive the food references which may still get worse as I go. Sure those things will certainly help individual sounds and overall mixes in their own special ways, but if the goal is to make the mix sound more like a band playing in a room together, then maybe there's another approach to try first.

Imagine that I was trying to record a string orchestra piece. And to do it, I grab a couple violinists and have them play the first parts and second parts in my living room, and I record with a stereo XY pair. Then I find a violist and record three takes of the same part in a practice room with a close mono mic. Then I find a couple cellists and record them in a classroom at school, but for the bass, I can't a player who can do the part so I use a really nice VST instrument. Now what are the chances of making them sound like they are all placed around the stage in a nice concert hall? I'd offer that they will be greater if I'm at least using the same reverb on each part of the group and not creating one sound for each of the four instruments. And I know without a doubt it would have been better if I had at least had everybody record in the same room - even if they were still recorded individually, just having that initial similarity in the sound of early reflections would afford me a better chance later.

So is that really any different than when you might record drums in a friends basement, overdub a DI bass, an acoustic in your living room with a stereo pair, an electric guitar with a single SM-57 on the amp, and the vocals in your bedroom closet?
It still won't be easy to take such disparate sounds and blend them into a single, consistent, and blended sound, but I'm guessing that starting off with chasing down the sound of a single acoustic space will maybe make things a little easier.

-Jeremy

User avatar
@?,*???&?
on a wing and a prayer
Posts: 5804
Joined: Wed May 07, 2003 4:36 pm
Location: Just left on the FM dial
Contact:

Post by @?,*???&? » Sat Mar 29, 2008 6:58 pm

farview wrote:
@?,*???&? wrote:Double converting at 44.1Khz can sound like ass.
You wouldn't be double converting anything..
No.

Indeed you would be. First, converting frequencies into a Pro Tools system, mixing in the box, then running that same mix out to tape.

It is true the mastering engineer might eq and compress in the analog domain, but the signal will have to be converted again to digital. The same frequency content is there- except it already suffers from the first digital generations conversion. The same frequency content will then need to be converted again.

Double nyquist theorem? 20KHz harmonics sampled twice? Yech. They barely get 2 samples as it stands. Add into that the effects of jitter and you've got some high frequency information completely jumping around.

Real fizzy shit.

User avatar
farview
tinnitus
Posts: 1204
Joined: Tue Aug 31, 2004 1:42 pm
Location: St. Charles (chicago) IL
Contact:

Post by farview » Sun Mar 30, 2008 9:29 am

Once the mix hits the tape, it's no longer the same signal. The tape will do it's own thing to the signal including the high highs. The harshness should be dulled by the tape, isn't that the point?

20k harmonics? Are you talking about the 2nd harmonic of a 10k signal?

You only need 2 samples to reconstruct. Most people complain about the low pass filter attenuating below 20k.

Besides, I've never heard anyone say "Wow, that would have been a great album if there was just a little more clarity above 19kHz."

This can't be the only thing keeping your mixes from being great, I just spent the last 45 minutes going through my CD collection with a spectrum analyser trying to get a feel for how much information there is above 15k. The answer is 'not much'.

stompforfuzz
gettin' sounds
Posts: 101
Joined: Tue Nov 01, 2005 2:12 pm
Location: Richmond, CA
Contact:

Post by stompforfuzz » Sun Mar 30, 2008 8:57 pm

Professor wrote:
So is that really any different than when you might record drums in a friends basement, overdub a DI bass, an acoustic in your living room with a stereo pair, an electric guitar with a single SM-57 on the amp, and the vocals in your bedroom closet?
It still won't be easy to take such disparate sounds and blend them into a single, consistent, and blended sound, but I'm guessing that starting off with chasing down the sound of a single acoustic space will maybe make things a little easier.

-Jeremy
Totally - I've been guilty of the million reverb march, and recently I read a post pertaining to using an aux send for effects mainly for saving cpu cycles, but the reason you gave seals the deal. I've tried it on a few of my old mixes and it does sound good - I used to use aux sends for groups of instruments (bass and drums together, then guitars, then vocals etc) .

Thanks for the enlightenment - I really should get to know a real console - the only time I ever used on was in college for live sound. Never used effects in that situation.
tear out your heart and make them wear it around their neck so they'll understand

MoreSpaceEcho
zen recordist
Posts: 6373
Joined: Wed May 07, 2003 11:15 am

Post by MoreSpaceEcho » Mon Mar 31, 2008 7:54 am

@?,*???&? wrote:
No.

Indeed you would be. First, converting frequencies into a Pro Tools system, mixing in the box, then running that same mix out to tape.

It is true the mastering engineer might eq and compress in the analog domain, but the signal will have to be converted again to digital. The same frequency content is there- except it already suffers from the first digital generations conversion. The same frequency content will then need to be converted again.

Double nyquist theorem? 20KHz harmonics sampled twice? Yech. They barely get 2 samples as it stands. Add into that the effects of jitter and you've got some high frequency information completely jumping around.

Real fizzy shit.
OMG WHAT are you talking about? i bet almost every Professional Legitimate record recorded to digital in the past 20 years has been run out to analog and back to digital in mastering and you think they all sound "real fizzy"? c'mon man. maybe double converting with your behringers sounds shitty, but most mastering folks have slightly better converters than that. jitter is also a non-issue in any reasonably professional mastering setup.

User avatar
@?,*???&?
on a wing and a prayer
Posts: 5804
Joined: Wed May 07, 2003 4:36 pm
Location: Just left on the FM dial
Contact:

Post by @?,*???&? » Mon Mar 31, 2008 9:03 am

MoreSpaceEcho wrote:
@?,*???&? wrote:
No.

Indeed you would be. First, converting frequencies into a Pro Tools system, mixing in the box, then running that same mix out to tape.

It is true the mastering engineer might eq and compress in the analog domain, but the signal will have to be converted again to digital. The same frequency content is there- except it already suffers from the first digital generations conversion. The same frequency content will then need to be converted again.

Double nyquist theorem? 20KHz harmonics sampled twice? Yech. They barely get 2 samples as it stands. Add into that the effects of jitter and you've got some high frequency information completely jumping around.

Real fizzy shit.
OMG WHAT are you talking about? i bet almost every Professional Legitimate record recorded to digital in the past 20 years has been run out to analog and back to digital in mastering and you think they all sound "real fizzy"? c'mon man. maybe double converting with your behringers sounds shitty, but most mastering folks have slightly better converters than that. jitter is also a non-issue in any reasonably professional mastering setup.
Another excellant, mature response from someone who does not know how digital conversion works.

MoreSpaceEcho
zen recordist
Posts: 6373
Joined: Wed May 07, 2003 11:15 am

Post by MoreSpaceEcho » Mon Mar 31, 2008 9:21 am

i think you are the one not understanding how conversion works, jeffrey.

by your logic, everything that was tracked to digital and then run through an A/D/A loop at mastering would sound terrible. pretty clearly this is not the case.

i record to digital. quite frequently i will run the individual tracks out of the computer into my analog eq and compressors, or the space echo, or whatever, and back in. they still sound fine. then the mix will get run out through the analog loop AGAIN in mastering, and lo and behold, it still sounds fine.

so tell me this...you have your mix on the computer. would you rather master it using all plug ins so you avoid the dreaded d/a/d conversion, OR would you rather master it by running out of zillion dollar converters into a bunch of top shelf analog eq and compressors, and back in via another megabuck converter?

User avatar
fossiltooth
carpal tunnel
Posts: 1734
Joined: Sat Mar 24, 2007 3:03 pm
Location: Brooklyn, NY
Contact:

Post by fossiltooth » Mon Mar 31, 2008 9:38 am

MoreSpaceEcho wrote:i think you are the one not understanding how conversion works, jeffrey.

by your logic, everything that was tracked to digital and then run through an A/D/A loop at mastering would sound terrible. pretty clearly this is not the case.

i record to digital. quite frequently i will run the individual tracks out of the computer into my analog eq and compressors, or the space echo, or whatever, and back in. they still sound fine. then the mix will get run out through the analog loop AGAIN in mastering, and lo and behold, it still sounds fine.

so tell me this...you have your mix on the computer. would you rather master it using all plug ins so you avoid the dreaded d/a/d conversion, OR would you rather master it by running out of zillion dollar converters into a bunch of top shelf analog eq and compressors, and back in via another megabuck converter?
I'm with MoreSpace on this one.

If an ELOP/La2a/ADL/1176/radioshack-eq/whatever, is going to sound great on your vocal/bass/jaw-harp/whatever, then go out of you digital recording device, into the unit, and then eventually back to digital. Or Tape. Or whatever.

Yeah, those Beatles records sound so awful because everything is like 90th generation tape. Yeah, "Degradation really kills music. F* it. I spend 90% of my time on digital mixes trying to find fun and pleasing ways of "degrading" audio.

So basically..... whatever. Do what works for your ears. If you're saying that going through one extra step of D/A to use an awesome piece of gear makes things sound awful, then you either:

A) Need better converters.
B) Are listening with your brain rather than with your ears.

I blame neurotic Gearslutz for making amateurs so nervous about "extra" conversion. Let them have their specs. Let the rest of us make records.

When I'm on a DAW, digital is my tape machine. Unfortunately, sometimes it's my mixer too, but hey, what can you do? I make it work!
Last edited by fossiltooth on Mon Mar 31, 2008 9:42 am, edited 2 times in total.

User avatar
farview
tinnitus
Posts: 1204
Joined: Tue Aug 31, 2004 1:42 pm
Location: St. Charles (chicago) IL
Contact:

Post by farview » Mon Mar 31, 2008 9:40 am

Even besides all that, the concern about using the same sample rate (or different sample rates) is only valid when keeping the signal all digital.

Once the signal is reconstructed, it is an analog signal. Generation loss is neglegable on anything but the cheapest, worst, soundblaster-type converters.

Everything that people complain about 44.1k will be softened and/or eliminated when the signal hits the tape. Due to the smearing, compression, and noise floor of the tape, all that stuff at 20kHz that is essentially 70db below line level will not hurt anything.

Mastering engineers will send the signal out into the analog world to avoid SRC and to capture the signal with top shelf converters.

User avatar
Nick Sevilla
speech impediment
Posts: 4959
Joined: Mon Mar 03, 2008 1:34 pm
Location: Lake Arrowhead California USA
Contact:

Post by Nick Sevilla » Mon Mar 31, 2008 9:52 am

MoreSpaceEcho wrote:
@?,*???&? wrote:
No.

Indeed you would be. First, converting frequencies into a Pro Tools system, mixing in the box, then running that same mix out to tape.

It is true the mastering engineer might eq and compress in the analog domain, but the signal will have to be converted again to digital. The same frequency content is there- except it already suffers from the first digital generations conversion. The same frequency content will then need to be converted again.

Double nyquist theorem? 20KHz harmonics sampled twice? Yech. They barely get 2 samples as it stands. Add into that the effects of jitter and you've got some high frequency information completely jumping around.

Real fizzy shit.
OMG WHAT are you talking about? i bet almost every Professional Legitimate record recorded to digital in the past 20 years has been run out to analog and back to digital in mastering and you think they all sound "real fizzy"? c'mon man. maybe double converting with your behringers sounds shitty, but most mastering folks have slightly better converters than that. jitter is also a non-issue in any reasonably professional mastering setup.
+1

MoreSpaceEcho
zen recordist
Posts: 6373
Joined: Wed May 07, 2003 11:15 am

Post by MoreSpaceEcho » Mon Mar 31, 2008 6:45 pm

farview wrote:Generation loss is neglegable on anything but the cheapest, worst, soundblaster-type converters.
this thread piqued my curiosity, so i ran a track through my D/A/D loop to compare. if anyone's interested:

www.rohdelikat.com/crag/oatmealraisin.wav
www.rohdelikat.com/crag/chocolatechip.wav

of course it would be better to use a less shitty sounding track for such a test, but everything else i have on my computer is clients stuff, so we're stuck with one of my stupid jams from like 5 years ago, recorded in my bedroom in about 20 minutes...

User avatar
@?,*???&?
on a wing and a prayer
Posts: 5804
Joined: Wed May 07, 2003 4:36 pm
Location: Just left on the FM dial
Contact:

Post by @?,*???&? » Mon Mar 31, 2008 6:59 pm

MoreSpaceEcho wrote:
farview wrote:Generation loss is neglegable on anything but the cheapest, worst, soundblaster-type converters.
this thread piqued my curiosity, so i ran a track through my D/A/D loop to compare. if anyone's interested:

www.rohdelikat.com/crag/oatmealraisin.wav
www.rohdelikat.com/crag/chocolatechip.wav

of course it would be better to use a less shitty sounding track for such a test, but everything else i have on my computer is clients stuff, so we're stuck with one of my stupid jams from like 5 years ago, recorded in my bedroom in about 20 minutes...
TapeOp could really use more experts...

RefD
on a wing and a prayer
Posts: 5993
Joined: Fri Aug 27, 2004 9:10 pm

Post by RefD » Mon Mar 31, 2008 7:08 pm

@?,*???&? wrote:
MoreSpaceEcho wrote:
farview wrote:Generation loss is neglegable on anything but the cheapest, worst, soundblaster-type converters.
this thread piqued my curiosity, so i ran a track through my D/A/D loop to compare. if anyone's interested:

www.rohdelikat.com/crag/oatmealraisin.wav
www.rohdelikat.com/crag/chocolatechip.wav

of course it would be better to use a less shitty sounding track for such a test, but everything else i have on my computer is clients stuff, so we're stuck with one of my stupid jams from like 5 years ago, recorded in my bedroom in about 20 minutes...
TapeOp could really use more experts...
always good to see a constructive remark from you, champ. :roll:
?What need is there to weep over parts of life? The whole of it calls for tears.? -- Seneca

MoreSpaceEcho
zen recordist
Posts: 6373
Joined: Wed May 07, 2003 11:15 am

Post by MoreSpaceEcho » Mon Mar 31, 2008 7:13 pm

jeff, you're an expert, which one of those files is the original?

Post Reply

Who is online

Users browsing this forum: No registered users and 8 guests