Mixing in or out?

Think I’ll be looking to get these sort of mixers. Most likely will sit in front of ebay for months moaning about the price of things these days.

http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=7376915306&rd=1&sspagename=STRK%3AMEWA%3AIT&rd=1



http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=7376423878&rd=1&sspagename=STRK%3AMEWA%3AIT&rd=1

Obviously its all opinions, but....


Over the years I've tried just about every approach to mixing I could think of and have always ended up drawing the same conclusion - mixing "outside the box" gives significantly better reults than mixing "inside the box".

A couple of years back, whilst mixing "Angelic Particles" for Hallucinogen in Dub, I decided that the mix I had going inside Logic was sounding good and that there was no need to split the outputs and mix it on the desk - I could just finish it off internally in Logic.

So, I got it as good as I could and burned it onto a CD. A couple of days later, Simon came round for a spliff and a cup of tea and I decided to play him the mix to see what he thought.

He listened and then said "Ok - good rough mix - when you doing the proper one?" and having listened to it again I could see what he meant. It sounded kind of dead and flat compared to all the other mixes I'd done.

I decided not to be such a lazy twat and to split the outputs and mix it again - which I did over the next few days. It was essentially the same mix - same arrangement, same approach, same balance etc etc but when it was done, the difference was amazing. The highs sparkled and the lows growled and the mids were defined and organic sounding. By comparison, the "in the box" mix sounded like a demo; everything was there but it had no "vibe" or "specialness".

Since then I've tried many times to mix a tune "in the box" and I just can't get it to sound like I want it to.

The difference for me is like the difference between large format cinema film and cheap camcorder video; mixes done "in the box" lack richness and detail and excitement.

I have a few theories about why this should be:



1. The "Painting the hallway through the letterbox" effect.

A mouse is a really tedious instrument for mixing music with. For me the ergonomic equivalent of mixing in Cubase or Logic would be shrinking your 64 channel analogue mixer down to 18 inches and turning the knobs with a knitting needle. Its fiddly and its dull, and it takes all the fun out of mixing.

I did a little experiment a while back. With a stopwatch i timed how long it took me to tweak a bit of top end eq onto a snare drum sample using a Waves REQ plugin in Logic. I had to open the mixer, find the channel, dbl click the plugin slot, find the top end "knob" with the mouse and give a it a tweak. It took me an average of 9 seconds to complete.

Then I did the same using my analogue mixer. It took 1.5 seconds.

When you consider the thousands of tweaks performed in the average mix, the fact that such a simple task takes me six times longer is, I think, significant.

2. Summing in the digital domain.

I have absolutely no idea of the science behind digital summing algorithms whatsoever, but I have performed many comparisons between the inherent sounds of the various audio sequencers and this is what I found.

Logic 5.5 pc sounds terrible compared to any of the others. Mixes done inside Logic sound scrunchy and grainy and dirty and not appealing to my ear. The denser the mix the worse it sounds. A real shame because I've yet to find any piece of PC software as reliable and bug-free as Logic.

Cubase VST 32 [long since obsolete] sounds better - clearer and cleaner.

Cubase SX/Nuendo better still - big and clear and sparkly and airy - the denser the mix, the bigger it sounds. Shame they are so buggy on my system.

Magix Samplitude/Sequioa - sounds amazing - like a high-end digital console.

Ableton Live 5 - Don't ask me how but definitely the best of the lot.

99% of music I have released has been made with Logic 5.5 pc, but on a PC with 50 analogue outputs which feed a 64 input Soundtracs Solo Logic analogue mixer. When used like this Logic sounds fine - as good as any other DAW. Its only when the audio is summed internally that this difference in quality becomes apparent.


The fact that there is such a high degree of variance between the different DAWs suggests that the summing of many signals in the digital domain is a really complex science, and suggests to me that this is where the weak spot is.

With all the talk about this mythical "analogue warmth" we hear so much about, i suspect that it is at the signal summing stage that it is most crucial.

3. A brickwall cutoff frequency of 22.050khz throughout the entire mix path.

Supposedly, the upper range of human hearing cuts off at about 20khz. It is for this reason that a sampling frequency of 44.1khz was considered adequate for CDs and other consumer digital audio products.

When you run your DAW at 44.1khz, the highest frequency that it can possibly generate is 22.050khz. Anybody who needs to know why should Google "Nyquist Theorem".

My cheap old mixer can quite happily deal with frequencies way beyond that, and significantly, its frequency response tails off gradually at the upper end. A digital mixer running at 44.1khz, on the other hand, cuts off abrubtly at 22.050khz.

To cut a long story short, inside my analogue mixer there are all sorts of ultrasonic signals flying about that I can't actually hear but that my brain can perceive, owing to the fact that these high frequency signals generate lower harmonics that i can hear and which alert my brain to the existence of the higher frequency signals.

In a digital mix [performed on a system at 44.1khz sampling rate] these ultrasonic signals are never generated and so the lower harmonics that point to them are never present - which, I suspect, is what gives digital mixes that lifeless, plastic quality. IMO of course.


The good news is that there has never been a better time to buy an analogue mixer. The small ads are full of lovely old Soundtracs/Soundcraft/etc etc mixers going incredibly cheap which would have cost an arm and a leg ten years ago.

Recently I was paid £80 to take away a beautiful old Soundtracs M Series 24 channel mixer - gorgeous EQ, loads of headroom, 8 fx returns, midi mute automation, amazing sound - because the owner wanted it out of the house. I subsequently tried to give it away and the conversation went something like:

[Me] - "I've got this lovely old desk if you want to use it. Sounds Lovely."

[Warwick Bassmonkey] - "Oh really? Brilliant - yeah - I'd love it. where is it?"

[Me] - "Its here under my bed - let me get it out and show you..."

[Warwick Bassmonkey] - "Oh bloody hell - look at the size of it - where would I put it? That would take up my whole studio. The missus would kill me... No - I'd better not."

And so there is stays until I find a use for it. I've decided to keep it now because it is so lovely.

But my point is, the world is full of unwanted old analogue desks and I have absolutely no doubt that, despite the convenience afforded by mixing inside a computer, if you care about sound then a real mixer is the only way to go.



NB. Having spent a lot of time listening to "Psychedelic Trance" over the last few years I am fully aware that rich, sumptuous sound is not exactly top of the list for most trance makers.

This is not a dig btw - it occurs to me that the soulless plasticy sound that i hate so much actually works within the context of psy-trance.

Thank you Ott for a very in-depth reply. This was just the kind of info and reasoning I was looking for.

I just had a little experiment and bounced of a small section track by track of a tune I’m working on in Logic and imported the audio tracks in to cubase SX3 and the difference was obvious. Especially when the two resulting loops where played one after the other in cooledit. The bass in logic sounds duller and further away and the high end sounds flat in contrast to the results in cubase. But knowing that if the signals were being dealt with in a good mixer the results could be better still.

Fingers crossed I can find my self an barging on ebay and get hold of some more Gina’s which are going really cheap on ebay now. One went for £10.00 a few months ago.

Of course if you get board with housing that beautiful old Soundtracs M Series 24 channel mixer under your bed I could be able to lighten your burden.:irolleyes "Cheeky bugger psyfi."

I’m going to try the audio through Ableton live 5 now and see what happens there.As far as the plastic sound it concernd I recently posted a tune on psynews.org for review and one of the comments pionted out the plastic sound issue. anything that helps my music sound like "sound" rather than a process resulting in a sound that carries the signiture on the process in it is something I've been looking for. Does that last bit even make sense?

I'm warning you... it really is ***ing enourmous. But gorgeous.

Everytime I come over feeling regretful about not taking it, I look around my 7' x 7' "studio" and realise the hopelessness of it all. Plus the fact that I probably won't actually ever have to mix anything all by myself.

That Mackie Onyx 1220 looks handy though...

Perfect sense.

A digital bounce down is just the "answer" to a very large mathematical calculation, whereas a mix performed on an analogue mixer is a waveform of fluctuating voltage, shaped by a collection of discrete components, all of which are subject to environmental influences such as temperature and humidity.

If you drive the mix buss of my Soundtracs, the mix gets fatter and smoother.

If you drive the mix buss of Cubase SX... well, actually you can't "drive" the mix buss of Cubase SX.

That is the difference.

Ott I understand where you're coming from, but for home production with limited space etc. I think mixing within a DAW is fine. If someone's desperate to improve an almost there mix via the traditional analogue route, then he/she can always hire or borrow a studio.

BTW- if you're using tracks from the DAW and driving them individually through an analogue mixer, then the source tracks from Logic will still all be clocked at 44.1Khz (unless you clock at a higher rate). Summing happens at a sample point, otherwise we're talking about oversampling where some lateral interpolation is performed?

I wonder if any of the DAWs do oversampling when they mix down. Anyone know? If they do, then I'd have thought the *only* place to do that (without extrapolating what the wave should look like) would be in the DAW's engine.

Not looking for an argument here BTW anyone (knowing that this topic quite often ends up in one .

BTW - Pascal always uses his huge 32 channel Spirit mixer and *no* VSTis. He has a rack of old analogues all CVd up and driven by Cubase.

So he should certainly be getting plenty of warmth!

I totally agree and it really is a question of whether you are happy with "fine".

I don't think I'd ever be happy with an "almost there" mix.

Yes, but as soon as i switch in my channel EQ and start tweaking I am generating all kinds of random upper and lower harmonics and artifacts that my plugin EQ doesn't. Sometimes i will find myself boosting 2khz by 5 db on the upper mid and then cutting 2khz by 5db on the lower mid - a pointless activity both on paper and also on Cubase SX channel EQ.

On my desk EQ, however, it results in a kind of halo of fizzing energy around the sound in question that i can't get any other way. This distortion is crucial for me and functions as a kind of glue or laquer - to smooth everything together.

I'm out of my depth already with "lateral interpolation."

Outside you bastard..! :imad:

Ah but my room is bigger than that. I got a piano in here. One I'm thinking of getting rid of as it has a split down the tuneing board, bugger. Then again what do you expect for 99p of ebay. Plus there isn't anyone who would kill me if I had it. Well Ott might cus he loves the mixer and the only way I would have gotten hold of it is by digging a tunnle under his bed and dragging it all the way back here. Which I am prepard to do. Now where is my spoon?

Who is Pascal?

Don't get me wrong, I'm not on some bullshit anti-digital "analogue is best" crusade here. If i could get the results I want using just a laptop and a sound card I would.

In summer the mixer chucks out so much heat it can get unbearable. The insert points crackle if i don't clean them regularly and every time I move house i end up cracking a PCB and having to spend £400 on repairs.

Its a fucking pain in the arse to be honest but when it comes to mixing I can't get round the fact that I need a mixer to do it.

Me and Simon have discussed this for hours and hours - going over digital mixers [D8B, DxB etc etc] control surfaces [Mackie Control etc] and always end up screwing up our noses and going "Nahhh!"

Not yet anyway. Maybe when processor speeds have quadrupled and sample rates have doubled again and modelling technology has improved to the point where you can't hear the difference.

Take, for example, the way soft synths have improved. Listen to the Steinberg Neon or one of the other early VSTi's and then compare to GMedia Minimonsta or one of the other latest generation. The difference is stunning.

Or compare early Cubase VST Eq and dynamics to the new Sonalksis stuff. No comparison.


If it keeps on advancing like this maybe in a few years I'll be mixing on a laptop the size of a Shreddie with a VR helmet and gloves.

In the meantime I'll be happy to slave over hot capacitors.

Ah - fair enough, I understand. We're talking about the external EQing etc. Effectively analogue effects applied to the track which are then summed over the set of tracks in the analogue domain.

Pascal Eloy a.k.a. OVNI.

Someone I collaborate with quite regularly ('specially live).

Hope I'm not getting you wrong - my comment wasn't a dig at all. Just saying that Pascal's also taking the "traditional" route and getting really good results in terms of quality.

Well taking all that in to consideration will the move to 64bit from 32bit improve the areas of digital sound we have discussed. Will it be possible to produces digital mix downs of a high enough standard or is the bit depth of a sound not in question here when the sound it's self is still produced in a purely digital manner?

Thassit.

I didn't take it as one.



As it goes I'd rather use VSTi's to generate sounds and there are things I can do with plugin fx that i couldn't dream of with outboard. I've got a lot of hardware synths gathering dust in my studio, while I spend all my time with Kontakt, Battery and a handful of VSTi's.

Mixing is the only part of the process that [for me] still requires a big lump of analogue junk to get right, and the day I can get the results i want without it, its going in the fucking skip i promise you.

I'm not so sure.

The resolution is effectively the set of potential values at a given sample point.

You can increase that as much as you like, but if you're running at 44.1Khz, you're always going to be guessing what the potential values between the sample points might be.

If you consider a sine wave at high frequency (say 22Khz-ish), then the resultant digitally transformed wave will necessarily come out more like a square wave. That's unwanted harmonics.

Now, with a clever bit of guesswork, you could interpolate between the values and guess that you're trying to create a sine wave, but of course you might very well not be.

So - you either fake (should I say 'guess' the most likely) the harmonics at high frequency or you increase the sample rate.

Bitrates and sampling rates are only half the story for me. As important is the ergonomics of virtual mixing.

Also - as Ott says, there are properties of the components of an analogue mixer that add to the character of a sound.

If you want to 'drive' the signal in the digital domain, you need to add an insert plug-in to attempt to recreate the distortion.

I think the other thing to bear in mind in the analogue vs. digital debate, from a synth point of view at any rate is (apologies if this is offtopic):

In my opinion, analogue and digital complement each other very well.

For example, I have a Z1 which by todays standards doesn't do a fantastic analogue emulation since it has horrible aliasing / digital artifacts (especially at high filter resonances).

However, that 'horrible'ness can actually be really useful. It can produce sounds that really cut through in a mix and have good contrast with analogue warmth.

NI got it right with Absynth 3 in that they let you turn the anti-aliasing on or off to taste.

Yeah piont taken. It would be move the case of Vst emulators like tube or psp vintage warmer getting that good. If they can?
Another question. When you have made your mix down through an analogue mixer. Once it is digitally recorded and on CD what will have happened to the ultra high sounds from the mixer? Will they be lost? I would imagine so but is the difference that the sound that was recorded was not hemmed in by the sample rate restrictions of digital? Or are these ultra high sounds only of add to the producer when mixing and it's the more audible sound of the mixer that make the final digital recording sound good?