Here I go again: Melda vs. Sonible, Oeksund Soothe, Gullfoss?

Official support for: meldaproduction.com
RELATED
PRODUCTS

Post

While many in this Forum are interested in sound creation, I am particularly fascinated by developments in computer-assisted music creation, mixing, mastering, and so on. I hope I am not alone here!

Izotope´s overhyped Track Assistant, Sonible Eqs, Soothe, and a few others, are now joined by a new player (Mac only for the time being, unfortunately): Gullfoss (https://www.soundtheory.com/home)

Are these things we can try and do with MSpectralDynamics or other, or is there no Melda plugin that can compete in this sphere?

Post

for me soothe is like a beautiful but gone wrong child of mspectral dynamics(even though both work diffrent ways i think IMO ..i could be wrong)

track assistant is utter BS i think .its better to belive your ears rather than some other guys algorithm(thats why you have bypass and solo on the MB bands)

sonible eqs are good i think seems like the freeform eq and dynamic eq had a child together

SERUM is like the another smart kid who cared about looks first and knowledge second in the classroom while MELDA took knowledge first and looks second (vojtech please add wavetable and drag and drop)
Win 10 x64 with specs enough to run DAW without bouncing any track
KZ IEM,32-bit 384Khz dac running at 32bit 48Khz
mainly use REAPER, MTotalbundle, Unfiltered Audio TRIAD and LION, NI classic collection,......... ETC

Post

Pretty smart to ask that question in a subforum for the melda fanboys :lol: :lol: :lol: Have you thought someone in here tries to answer that objectively?
Apratim wrote:track assistant is utter BS i think .its better to belive your ears rather than some other guys algorithm
So no need for the big Mixing Revolution anymore? Tell that Vojtech. He's already preparing it :lol:

Post

Apratim wrote:track assistant is utter BS i think .its better to belive your ears rather than some other guys algorithm
I agree with your first phrase and disagree partially with the second. TA was oversold when what it mostly does is recommend a previously PRE-EXISTING preset for your track once it decides what instrument it contains. There is an interesting YouTube by Eric Tarr which shows how TA recommends almost the same settings for a distorted electric guitar and an acoustic guitar, because they are both "guitars"!! I think Izotope was just in a hurry to offer something first, and offered almost nothing but hype, knowing that most reviewers would be too lazy to really check their claims. I am under the impression that their assistant for mastering in Ozone is a bit more sophisticated, but I have not really tested it yet.

Re. your second phrase, there are very interesting things being done, and that can still be developed, with algorithms, machine learning, AI, etc. I find it fascinating and I think it has a lot of potential. Perhaps Vojtech will contribute something new with the Mixing Revolution sometime this century! ;-)
Of course, I am not among those who think robots will replace us completely. These are just very interesting tools to allows us to concentrate on the creative aspects. Anyway, it is a huge topic...

Post

It looks as though that Gullfoss is just providing dynamic equalization based on some pre-defined target curve(s). That's easily done with MSpectralDynamics:

Just run some reference track with a good balance (or several different tracks at the same time) through MSpectral and use the "capture" function to get a snapshot of your own target curve. For this kind of gentle shaping I'd go with a fairly high smoothing amount (10% maybe?). Then just run your own track through MSpectral and that's basically it. For this kind of use I'd probably try a slow attack (maybe with maxed out RMS length) and a super slow release (maybe using the hold function as well) so that it's functioning not so much as a multi-band compressor (and reacting to momentary events) but more as a gentle frequency shaper that reacts to longer term tonal changes. Try a low ratio and then just pull down the threshold until most of the frequency range is getting some amount of compression. Increase the ratio to exaggerate the effect or lower the ratio to make it more subtle. (If some part of the spectrum is constantly getting more compression than other parts it means you have bigger balancing problems that might be better addressed by placing a static EQ before MSpectral.)

I haven't demoed Gullfoss so I don't know its exact behavior but based on the description on their webpage that's how I'd go about doing something similar.
Last edited by Bullfrog001 on Sat Feb 24, 2018 6:08 am, edited 1 time in total.

Post

Good advice Bullfrog001.
I agree, except my bet is that the target curve they are using is the psychoacoustic thingy (the way humans detect certain frequencies) I think it's called fletcher munsen or something like that.
Melda Production & United Plugins
Surface Studio = i7, 32gb, SSD.
Windows 11. Bitwig, Reaper, Live. MTotal.
Audiofuse, Adam Audio monitors + sub, iLoud MTM.
Polybrute, Summit, Pro 3, Tempest, Syntakt, AH2.
Ableton Push 2, Roli Seaboard Block.

Post

Great reply, Bullfrog001, thanks!
I will try your suggestions and settings.
I had never thought of building a reference curve by playing several tracks simultaneously. Ingenious.

Post

jmg8 wrote:I agree, except my bet is that the target curve they are using is the psychoacoustic thingy (the way humans detect certain frequencies) I think it's called fletcher munsen or something like that.
I am curious: why would you bet that the target curve they are using is the "psychoacoustic thingy"? :wink:
You may be right, but what leads you to think it could be so?

Post

Just a guess really. But in the description it mentions a few times "auditory perception".
Melda Production & United Plugins
Surface Studio = i7, 32gb, SSD.
Windows 11. Bitwig, Reaper, Live. MTotal.
Audiofuse, Adam Audio monitors + sub, iLoud MTM.
Polybrute, Summit, Pro 3, Tempest, Syntakt, AH2.
Ableton Push 2, Roli Seaboard Block.

Post

It looks as though that Gullfoss is just providing dynamic equalization based on some pre-defined target curve(s). That's easily done with MSpectralDynamics:
I'm really reluctant to jump into a Melda forum, so tell me to get lost if you will! Just to say - Gullfoss is not a dynamic equalizer. We try very hard to preserve the dynamics - there's a form of loudness conservation in what it does. Part of what Gullfoss can do is to egde the frequency distribution towards an idealised form that helps to maximise the information received by your brain from the audio signal. (Sorry if that sounds cookie, but it is literally what we try to do!) A somewhat cruder way of doing this is to try and 'pink' your signal - and you can use techniques similar to the one you mentioned. Eddie Bazil has a really nice article about this in Sound on Sound a few months back. But beneath the hood Gullfoss is doing a lot more - and the overall aim is to enhance the audiblity of perceptual elements within your mix....
David
Soundtheory

Post

DarkStar, ... Interesting, if true
Inspired by ...

Post

I was also thinking of this more recent one:

https://www.soundonsound.com/techniques ... -made-easy

Post

Well, using the automatic equalization in Melda MAutoDynamicEQ using as a reference signal a pink noise (or, differently, a white noise), which is indeed an interesting and useful technique, seems to give me vaguely similar results, but it's not at all the same, as far as I understand (no psychoacoustics and no loudness preservation involved).

Post

Soundtheory1 wrote:I'm really reluctant to jump into a Melda forum, so tell me to get lost if you will!
Thank you for coming here politely to give us your explanation. I, for one, welcome your appearance. Vojtech has generated a great forum for wide discussions to take place.

Having said this, since you understandably will not give away trade secrets, and us Windows users cannot try your product yet (you told me it will take about six months in an email yesterday), I think there is little more we can discuss with you here in the Melda site.
Questions:
1) Do you plan to create a forum page?
2) Do you plan to create YouTube videos with explanation, tutorial, examples, etc.?

Meanwhile, here we will work hard to replicate and improve on your results using Vojtech's powerful Melda plugins. :D

Good luck!

Post

jmg8 wrote:Just a guess really. But in the description it mentions a few times "auditory perception".
And now he said "maximise the information received by your brain from the audio signal", and "enhance the audiblity of perceptual elements within your mix".

You may be on the right track, Sherlock! :wink:

Post Reply

Return to “MeldaProduction”