Hello everyone, I hope you're all passing a good Sunday.
I come towards you in a hope of advice from wise minds.

Most of my mixes sucks. I'm really into it (reading books and tutorials) since not so long, as the quantity of academic information tears appart my creative side.
In my quest of getting an "out of the box", "pro-like", "thick", and most of all "authentic" sound, with a sense of solidity, an identity, I clearly keep the faith of digital tools. I try to keep farseeing beyond the gearsluts wars, searching the balance between rational technology-wise facts and hardware worship arguments.

Once, playing cross-modulation with saws I got the idea to oversample the virtual modular synth. It was with Arturia's one. By the way, softube is a complete monster compaired to, can't stop using it since a the last two weeks...
Anyway, deadmau5' example in his masterclass hyped me to test those digital audio limits; and with an hazardous and unstable setup (a wrapper for plugins with a x2 option) I've been able to oversample what technically you cannot.

And it sounded different. But not like shit.
The crossmodulations where more complex, the dragonball z-ish artefacts sounded more close. I tried to draftly write basslines and basic leads, and everything seemed much closer.

But in was a pain in the neck, too much loading UI loading, anyway my CPU was ok so it was cool then.
I abandoned the idea when I tried an x4 experiment, wrapping x2 a wrapper containing the x2 synth... Too much instability.
With this principle applied, notably propped by other examples (UAD sound, option to oversample on some distortion/dynamics plugins, ect...) I decided to "oversample" my daw. So I could run all my plugins in a oversampled utopia where everything would sound better.
And to my ears... everything seems to sound much better. Filter resonance on synths is just wow, compressors seem to bring the sound so close, ect...

I heard about sample rates earlier, I had my simple postulate (useless, in the end it will be compressed) but after this experiment I thought differently. I knew I'll have to be cautious because it may not work with everything, but, I really think it makes live production sound better. Because all the DSPs involved will run 4 times faster. It will not sound more analogic, because it's in the box, but less digital, I guess.

Nor with a better quality material, in terms of quality standards nowadays digital whiped the ass up to analog, but with a better Sounding material, downgrading a 176.4khz print to a lesser sample rate is much more appreciated than just creating / recording and printing in 44.1.

Don't worry I'm almost done.

This is just my opinion. I wanted to share this to get reactions and advice against it (not necessarily negative) because I want to progress and get wiser.
My questions are "is oversampling my daw will make my overvall dsp acruacy / work better (dynamics, eq...) ?"
Will my new 176.4khz calibrated daw will bring to my ears better mixes ?

Secondly, I want to switch to analog mix. I'll have questions about it because I'm a total noob on analog mix but I may write something else next time.

Thank you for your attention.

 

You really need to get technical about cons and pros of it (try to read scientific papers, not some "wild ideas"), but in the end it is NOT what makes "mixes better".

 

No.
You can make awesome mixes at 48kHz, as well as 96kHz (and 192kHz).
Good mixing and mastering is not really dependent on SR and BD. It's more about knowledge, experience, talent, etc.

I hear differences though regarding higher frequencies captured, fold-over/aliasing, etc (different AD/DA converters and different SR), but a good mix is still a good mix, regardless of SR.

Shoot your analog questions. We are several here who came from that world. At the moment I have access to SSL 4000G console, API 1608 console, awesome outboard gear (racks and 500-series processors), etc.

 

Last edited: Jun 11, 2017

For proper work from start to end, higher samplerates, oversampling, bit depth are for less digital shit to be out and more precision for analog / digital sound to be achieved, to be more digitally cleaner. It's all about digital artefacts mostly.
The main are performance, arrangement, music, how you record, edit, mix, restore bad stuff, master. Your gear, your soft, monitoring and room, skills and knowledge. Your ears, brains and hands.

Oversampling and high sample rates are for less aliasing, to minimize or almost kill them. Usually if you have a lot of highs and a lot of distortion and/or saturation, aliasing can be heard, or not. So if hear aliasing, use oversampling built in your vst, or/and use higher samplerates. Usually, intermodulation distortions is bad too, but you can avoid it by not using distortion or saturation. On simple sines and sweeps you will see harmonics created from saturation, even, odd ones. But if use saturation on sine+sweep or more complex sounds, there will be harmonics and intermodulation distortions. Synths create aliasing too as well.
Also equalizers with oversampling or at high samplerates work extremely much better in high and mid-high range areas at 44.1, so near nyquist it works without cramped and decramped filters approximations. or at higher samplerates, it will have extended frequency range, so your highs will be processed as with analog eq shapes/curves, not digitally limited. Also, phase shifting near nyquist will be much closer to analog. So, at 44.1 we need 2x or 4x oversampling to avoid this digital approximation near nyquist. More info about this, eqs, can be found on vladg page. Btw, decramped filters are much better than cramped ones, but oversampled or at high samplerates it is better than decramped. Decramped is a kind of compensated or so.
Yes, your audio will be represented closer to analog, but less digital, so to say less digital artefacts.
Also, if you work with 96khz, for example and you have 44.1 kHz audio samples, you must resample them to 96 kHz with excellent src, such as rx, sox, dbpoweramp etc. You achieve almost ideal filter, with steepest high-cut closer to nyquist and linear phase. Pre and post ringing due to steepest filter and linear phase is not noticeable at nyquist, because it is ultrasonic, and it happens too fast and not so loud. Worse effect with minimum phase. Your transient and all highs become sounding later and with long ringing, and you smear more highs below nyquist. So linear phase is more safe. Rx tests may prove this.
Also linear phase give less phase shift, so you may clip your sound much less than with minimum phase (less peaks grow due to linear phase). But before upsampling, first do gain down if this sample is already normalized, it will not clip after resampling.
After all the work is done, render your master without limiters and not clipped, resample down to 44.1 (sox, dbpoweramp, finalcd 0.24 with sharp filter 32 bit, or rx advanced) and then use final limiter with oversampling / intersample / truepeak detection, output ceiling (-0.3dB or even -0.5dB for wav/cd/flac/other lossless and 16 bit dithering, standard tpdf or noiseshaping such as mbit+) or output ceiling (-1dB or so, no dithering, render any lossy stuff as mp3, AAC, ogg etc).

These are true tips and tricks for how to work with more quality and less shit out. Other methods are worse and logically not proper. I work at 96khz with bit depth higher than 24 bit. 24 bit is for analog to digital recording, because audio interfaces usually work as 19-22 bit practically, but for posterior editing and restoration I record at 32 bit and higher, I never choose 24 bit, because after ADC you need your audio to be processed higher than 24 bit, less shit appears, less truncation, less quantization errors etc. Digital audio created on pc internally is much more than 24 bit.

And more info: working at high samplerates is better than always oversample up and downsample at lower samplerates. So to say, less often oversampling procedures happens.

 

Last edited: Jun 12, 2017

Oversampling will do nothing if your mixes suck. At the end of the day, wherever they are uploaded after mastering they will be 320kbps mp3. Fix your mix or hire a mixing engineer and focus on creativity only.

#1 cause of bad mixes is bad mixer and bad production.
#2 is acoustic environment
#3 bad monitors

That's all

 

I am to lazy to write. Watch Mr Craig Anderton talk about oversampling/upsampling at minute 15.24 ...

I personaly record a lot of instruments in my music. And i am using some hardware while mixing so i am not totaly in the box which means 48khz and i give (most of the time) a f**k for oversampling/upsampling.

 

Last edited by a moderator: Jun 11, 2017

Thank you so much for all the responses, that fast.

I want to get an Signature 22MTK from Soundracft actually, the product seems promising. My accession to analog mix.
Just one thing, being such a noob about this (I never approached an analog mix console, only some digital ones) and producing with ease and digital freedom on my Ableton, I've noticed that hardware channels are actually Mono. I just wanted to know if that's a console standard, because it may become such a pain in the veins for bouncing stereo tracks, with two faders... no ?
It makes 11 stereo tracks, leaving let's say one for the Kick and one for the bass, so 10 stereo tracks... this become quite limited isn't it ?

 

Yeah, that type of console is pretty standard. I still own a beat up Soundcraft live mixer and have worked on many of them. Usually 24 channel ones and usually for live gigs, but also in the studio recording bands.

Yes, console channels are mono. If you want stereo you use two channels and counter-pan these out. Some mixers have stereo line ins, like dedicated tape/CD/aux/laptop ins, sometimes with just a volume knob.

You can use group outs from your DAW and audio interface, so that you group things like basses, drums/percussion, synths/pads, vocals, effects, etc. That's how you keep the channel amount down on your HW-mixer.

 

Last edited: Jun 12, 2017

I think it's been said before. Just stick with one sampling frequency all the way. The plugins that really benefit from oversampling (usually those who introduce high distortions/non linearities) have already that option inside the box. You may wanna check these specially for the final mixdown.

 

Just pick a sample rate and stick with it, focus on EQ, balance and the quality of the input sound and less on dynamics and distortion. Be creative, but also limit yourself, so that you can understand what interactions are taking place. It isn't as formulate as some people make it out to be, mixing a song. When after a few years you think you've got that down, start adding more flavors and into the mix.

You do not expect to play like Joe Bonamassa when you pick up a guitar for the first time, why do you expect to be able to mix like someone who's been at it for 20 years in a month or even a year later.