Filling the gap- community mixing/mastering fx project(s)?

[noisenerd]

Branching off from a discussion in Spogg's Flowstone music thread so as not to derail...

I would first like to mention that I mean no disrespect to existing products, such as the awesome Variety of Sound stuff, but there are still a few gaps to be filled, and, well, variety never hurts either, eh?

So, this is an attempt to see what we (as individuals or a community) might be able to do to increase the number of quality mixing and mastering tools built with FlowStone, for the good of the FS community and the world! Or something like that...

For my part: know a lot of theory, but am still in the process of learning the necessary DSP for such tasks. Right now I'm concentrating on this "circuit-modelling" stuff, specifically filters, largely because I'm pretty familiar with the analog counterparts. I have built a lot of stuff from others' components in SynthEdit and other similar packages, so I'm pretty familiar with the "visual programming" paradigm and how a lot of stuff is made... but I couldn't make anything at the level of the VoS stuff at this point, for example.

Some general thoughts on the subject (none of these are new ideas, just sharing my perspective):

- I think the "analog magic" so many people talk about is largely hype. There is no denying that certain analog devices sound great, but if you study these things, it's pretty much down to two things: distortion and noise (often subtle, but still). The VoS blog actually has some great articles on this, and if you look around you can find more (I can rustle up some links if anyone's interested).

- Along those same lines, I learned in my analog EQ experiments that a lot of the magic there is down to the shape of the curve (this is something I've seen discussed elsewhere too, can't recall where offhand though). My DSP experiments so far (and info all over the net) have shown me that this can be tricky in certain situations, but apparently there are ways to cope.

- I think tubes are overrated (not that they're bad, mind you), but apparently putting them on your GUI and making them glow brighter as you increase the gain (or similar parameter) makes your plugin sound 31% warmer.

Anyway, it's very late here and I may be rambling. So, anyone else interested in getting involved?

[martinvicanek]

- I think the "analog magic" so many people talk about is largely hype. [...]
- I think tubes are overrated (not that they're bad, mind you)

+1.
Nowhere else will you find so many notorious believers in BS that some self-proclaimed authorities continue to spread, ignoring solid evidence of the contrary. I used to own a Les Paul and a VOX AC30, and yes, if you cranked up the volume to the max, suddenly the world turned into an exciting place. But I am convinced we can reproduce this genuine vintage sound with today's DSP HW and algos, and go even way beyond, explore endless possibilities we could not even imagine in the analog era.

Thanks also for the VoS link! And if you have more of that sort, please do post them. I do not have much experience in mastering, but I am willing to contribute to this project if I can.

[tulamide]

This is the area where I can contribute close to nothing. I know what I need when mastering. But I don't know how to build that.
For example, on the master track I use a Maximizer/Limiter. I don't use it to maximize loudness, most of the time it's just driven very soft. But it needs to be there to get that final punch, that clear definition of the main sounds. Such a maximizer should have a 4ms look ahead. And here I'm out. I don't know what this does technically. Probably making sure inter-sample peaks are caught. However, it is this setting that helps me much to achieve that final sound I'm after. Together with the soft clip which I can turn on and off and drive it from 0 to full, and it is a subtle change in characteristics. Again, no clue what it does technically.

Prior to mastering I often use a sidechain compressor while mixing, for the ducking. Again not to get that over-exeggerated pumping of EDM, but to create more room for the kick drum.

Those are effects I miss as SM/FS effects. There is a sidechain compressor, but it is working ...what would be a word to correctly describe it... unmusically. It's harsh and metallic, not warm and soft. Ah, hard to explain.

Finally I am also missing a really good stereo processing effect. Whatever technique it would use (mid/side, de-phase, etc.), it needs to have musically supporting effect on the music, not just setting sounds wider. I am currently playing around (well, was playing around, it is on hold while doing the Ruby modules for someone else) with a mid/side stereo processor working with 3 bands for greater flexibility (for example, low towards mono, mids subtle widening, highs full stereo field), but the first tests did still sound ... technical more than musical.

A wall of text already, I will stop here

[noisenerd]

Nowhere else will you find so many notorious believers in BS that some self-proclaimed authorities continue to spread, ignoring solid evidence of the contrary. I used to own a Les Paul and a VOX AC30, and yes, if you cranked up the volume to the max, suddenly the world turned into an exciting place. But I am convinced we can reproduce this genuine vintage sound with today's DSP HW and algos, and go even way beyond, explore endless possibilities we could not even imagine in the analog era.

I couldn't agree more.

Thanks also for the VoS link! And if you have more of that sort, please do post them.

Cheers! Coming right up...

Here's an older one that's still one of my favorite, showing how you can mimic any "classic" analog EQ with a standard digital parametric EQ + saturation: Digital EQ fact & myth

I think that concept applies broadly. For instance, with compressors, it might be getting the envelope/timing right + saturation.

Some good stuff linked on this guy's site: http://www.timstinchcombe.co.uk/index.php?pge=papers

The rest of these are less directly about the subject at hand, but contain info that may be useful:

First off, a video for anyone that still argues with you about digital audio being crap: http://xiph.org/video/vid2.shtml

Here's some stuff that's more about distortion modelling for guitar purposes, but some info is broadly applicable:

http://www.simulanalog.org/

https://ccrma.stanford.edu/~dtyeh/papers/DavidYehThesissinglesided.pdf

This is an old page with one of the better guides to analog EQ that I've come across. Probably only really helpful if you have no idea how these things work, but a good confirmation that there's nothing magic about analog EQ if nothing else (not that this is every circuit, but seeing the basics if you don't know them should explain what I mean):

Wayback machine link

Here's some tube-based synth circuits. Warning: author is the guy from Metasonix, an unapologetic tube dude. Still, there's some good info here (he gets pretty in-depth), and like the EQ thing, a good example of how tube circuits aren't magic, so if you can model a tube to a convincing degree, you're pretty much there:

http://cgs.synth.net/tube/index.html

General circuits, useful for reference purposes:

http://experimentalistsanonymous.com/diy/index.php?dir=Schematics
http://www.emusic-diy.org/Schematics

I'll post more if as I remember/come across 'em.

Such a maximizer should have a 4ms look ahead. And here I'm out. I don't know what this does technically. Probably making sure inter-sample peaks are caught.

Here's a pretty decent article that might help:

https://www.soundonsound.com/sos/dec00/articles/adcompression.htm

Sounds like we have the same basic idea though, keep things subtle but add some oomph. I'm not saying you necessarily want to follow my advice, but I find you can get some nice added (perceived) volume without getting squishy by chaining 3 or more stages of compression, but using slowish attack times on the first ones so as not to lose transients. I usually try to not have more than 3 dB of gain reduction per stage.

Finally I am also missing a really good stereo processing effect.

Just my two cents worth: "mono-izing" the bass definitely has value, but I'm not a big fan of stereo wideners. I've just never been impressed with the results I get. I find I get better results by doing stuff in the mix like hard-panning two copies of a track (or two different takes of a guitar or whatever) with different EQ settings and a slight delay on one, and/or good stereo chorus/phaser/etc. I could post an example if you like.

Of course, this stuff is highly subjective, and also dependent on the material in question.

Heh, my turn for the wall of text. More later...

[Spogg]

I would like to mess around with "Mastering" at some point.

I would like some music files in wav format that are un-mastered stereo that I could fool around with and see what results I could achieve using Flowstone only.
If guys could link to such files I would really appreciate that. Maybe we could do some comparisons and maybe I could learn something...?

Just an idea.

Cheers

Spogg

[nix]

Sure-
I can do that, if you can put up with my music hehe

I have 2 mastering methods,
with the intent of reducing dynamic range and increasing loudness.
One is to use AudioDamage RoughRider compressor to raise low threshold sounds,
then brickwall limit with a FS limiter and normalize.
Just 2 weeks ago I learnt DiscoDSP Thrillme, and can master in one pass now in a little mellower fashion,
setting ratio, threshold and spectral balance. It compresses and limits in a charming way of interactive devices.

edit- I forgot to mention my third mastering element, cassette tape. I use this to saturate and remove any dc.
Sometimes I seem to get pretty bad dc errors, but tape will fix that better IMO than a semi-automated audio editor.

[tulamide]

I would like some music files in wav format that are un-mastered stereo that I could fool around with and see what results I could achieve using Flowstone only.
If guys could link to such files I would really appreciate that. Maybe we could do some comparisons and maybe I could learn something...?

Here you are: yeah.wav
This is the un-mastered stereo of this song: Yeah!
The wav file is in 24-bit int format.

Legal notice:
copyright 2016, Sevenfifteen Music.
All rights reserved.
You are allowed to use the wav file to train your mastering skills. All other uses must have a written permission.

[Spogg]

Thanks guys!

The wav file is in 24-bit int format.

Before I even start tulamide, do you know if I make, say, a multiband compressor in FS and run it in Reaper as a VST, will it work with a 24 bit track? Or do I have to down-sample it first so the VST will work?
Sorry for a basic question but I'm a beginner at this kinda thing...

Cheers

Spogg

[tulamide]

Thanks guys!

The wav file is in 24-bit int format.

Before I even start tulamide, do you know if I make, say, a multiband compressor in FS and run it in Reaper as a VST, will it work with a 24 bit track? Or do I have to down-sample it first so the VST will work?
Sorry for a basic question but I'm a beginner at this kinda thing...

Cheers

Spogg

The only thing important is to differentiate between working within Flowstone and working with an exported VST. Reaper has no issues with this wav format (I know because I do the mastering with Reaper), and Reaper will deliver its own format to the VSTs (raw audio data, no sound format headers). So that should work. Only when using the stream or wav player in Flowstone you have to consider the format. At least that's my experience so far.

[noisenerd]

Here's another one for ya. 24 bit, 44.1kHz. It's the pre-master for the track I posted in the "Flowstone music" thread. Use it however you like.

My basic mastering setup goes like this (in order of routing):

Gain (if needed)
-> HPF at 20 Hz or less
-> basic tonal EQ (hi/lo shelf, etc., again if needed)
-> collapse bass to mono (say, < 100 Hz, varies per track)
-> multiband compression
-> precision EQ (tweaking specific areas)
-> slight "flavorless" compression
-> slight "flavored" compression
-> psychoacoustic dynamics processing (like an exciter, I use a thing called CLAS which seems to have disappeared, but there are similar things around)
-> VoS Ferric (more flavor)
-> brickwall limiter (limit to -0.5 dB for safety)
-> dither

It looks more complicated than it is. The reason I use so many compression stages is because I find I can get more relative loudness without sounding squashed by only compressing a few dB each time. The "flavor" stages are generally analog emulations that I think add just a touch of pleasant distortion. I go easy here, any heavy stuff is done on individual tracks in the mix. This is maybe the ultimate "ymmv" topic though, and as I've said before, I don't claim to be an expert. I have generally had good comments about it though, and I'm generally happy(ish) with my results.

There's also at least a couple of debateable items in there, like the "psychoacoustic" deal, and the dither. Google it, read both sides, experiment, and make your own choices there. I'm generally torn on exciters and their ilk, but I do like the results I get with CLAS. It was a freebie from a company that seems to have gone under. Dithering can be explained better by people other than myself.

There's some good info on this subject out there in the interwebs, and if it's confusing at first, it will make more sense as you play around. Be warned, there's a lot of audio woo being tossed around in this area... but you should be able to filter it out, especially since you have an electronics background.