If I had the time, I know what kind of test I would do:
I would write an algorithm that generates MIDI data, by randomly pressing and releasing 88 keys, so that 2/3 of all keys are on most the time, but there would be short silence, also. Furthermore, with polyphonic aftertouch data, keys that are on, receive some random volume changes around 2-12dB, in random intervals from a 1/4 second to 3 seconds. These changes would be of two types: ramp/sweep of volume, and step (instant/rectangular volume change). I would perhaps separate ramp and rectangular envelopes into different test runs.

Each key would just create a plain sinus. But I might also switch to a triangle or ramp waveform, in further, different runs.

Then, I would examine input vs. output spectrum, and map this spectral analysis to a certain digital data format, in that each frequency (=key) gets a number, and there are two amplitude values, in and out. Any key press/release, and any change of volume gets a timestamp marker. But dynamic response gets recorded too, in that there is a sampling interval of 50 milliseconds, retriggered by any change in the input sgnal. If nothing happens for a second, we still have data by a 50ms interval.

I would refine this mapping, in that the numbers get doubled, and I will have all the frequency values that sit half-way between two keys, additionally. Just to discover, whether frequency in some way breaks out, and has non-linear treatment, too. (Think analysis of a Leslie rotor cabinet, in a different application.)

Then, I would feed this to a data mining AI system.

I would tune it to listen to harmonic correlation (every 12 keys there is an even harmonic, etc. plus some organ drawbar system for these odd harmonics), and to timed patterns, e.g. if there is some transient shaping. (Therefor, the re-triggering and sync of the sampling process by the time stamp of a change in input, that may come at any random time.)
I would ask it to discover any pattern, that matches input vs. output and resides beyond the linear and trivial behavior of an EQ, or multiband transient shaper.
I would ask it to judge certain aspects of "complexity", and changes of complexity between input and output. (Traveling through realms of entropy analysis.)

With different runs, and different types of these input changes, I would tease it to discover, whether the pattern of reaction is in any way systematically different.

I would tamper with the MIDI randomizer, and overlay it with rhythmic patterns, and harmonic patterns, and gradually reduce the numbers of keys that are _on_ simultaneously. This means reduction of entropy, so I want to know if the plugin is jumping on that fact, and develops a particular reaction, according to the increase of "musical" content in that random input stuff.

...
So, that's naturally a way to get an idea about any blackbox that has inputs, outputs, and a number of knobs...

Whoa, if you ever do that you should record the process. I think that would get a lot of views.

Today I'm just settling for comparative frequency spectrum analysis

Just one end user’s perspective here, but in my experience the answer to the title of this thread is both. Used judiciously, these are tremendous tools when used both individually and collaboratively.

Well, probably depends on your source and clip gain/headroom when editing and mixing. I barely feel any difference on factory settings, anyway - you can quickly tweak Soothe according to your needs and taste, that's for sure!

I'm pretty conservative with tracking levels, I still find myself backing it off to around 11o'clock though! It really depends.

I was working with a really peaky singer yesterday - even on a 47 his 4k is piercing - a notch and some extra help from soothe and he doesn't rip your head off anymore

I made a follow-up video to the previous video I made about Gullfoss where I compare the results of that frequency analysis within Plugindoctor to the curves shown in Gullfoss's own frequency analyzer when it receives audio material. The shapes of the curves are very similar.



I'm very much interested in a specific and accurate description of what this plugin does on a technical level.

One thing I would say about Oeksound Soothe even though I don't own it is that their site is very specific about what it's doing technically. I wish the Soothe manual was available on the website without having to download the plugin first though.

Talk about presumptuous, wilful determination to misrepresent a plugin, and lecture the developer based on your own shortcomings, and crude 'analysis'. They have been very patient with the geeks here trying to bring them down, and talk up themselves. 'Take down your marketing' - what barefaced cheek - how about you take down your amateur efforts to insult them?

Soothe 2 just announced.

Commentary like this only fuels me to want to investigate the subject further.

Whenever there is such sensitivity to reasonable requests for more clear and technical descriptions of the basic operations of a product, that seems to suggest that there is something to hide.

Also, if a leader of any organization calls anyone incompetent for daring to ask for clarification on their statements if that person questions the word-salad marketing language handed to them, it truly inspires me to expose the reality if possible.

Dude there is nothing to "expose"..... They asked you not to analyze with Plug In Dr because it doesn't make sense as GF is an adaptive plug in...... That makes total sense even to a non tech guy like me. Please use your ears to see whether this does what you want it to. Tools like this Soothe Soundbrigade aren't static emulations and react differently to different input like a dynamic eq.....the analyzer isn't made for these......

You're coming off like an ill informed zealot FWIW......they politely explained it to you and you've been using some nasty language attacking them.

My second video shows the similarity of the curves Gullfoss's own analyzer has to the results of a linear analysis that can be done in Plugindoctor.

Anyone can perform the same tests and learn quite a lot about this plugin.

Also, I've never attacked Soundtheory at all, neither in this thread nor anywhere else, so maybe you are confusing me with the other person who was in this thread whom had their comments deleted.

Their marketing however is a different story.

I'd recommend a complete reset on their marketing.

It would be nice to hear an official response on the many questions I've had in this thread.

Look you haven't a clue on why or what they are doing to make their product work, so please can the advice.....

They are doing fine and getting great results in a very high level professional market, so why don't you release whatever product you want to astonish us all with, and move on!.....

What???

I'm just looking for information on what the plugin is doing on a technical level because their marketing does not match nor adequately describe the user experience with it.

I'm not trying to start a plugin company.

However I have started my website to discuss music production, and I'm interested in what this plugin is actually doing.

lower latency! which is my only problem with the original...great!

When I asked by email before buying recently they said it would have lower latency, but still not be usable for real-time.

Still, lower latency is always a plus.

I don’t need it in real time - I don’t really like using native plugins ever in real time, it’s never that great latency-wise.

My issue is that with an hdx rig you can run out of delay compensation quickly, meaning I end up freezing vocal tracks or bass tracks with spiff on them.

Hopefully this helps with that!

introduction price for people who dont have it yet, would be great.

they couldn't be anymore different.Soothe is essential IMO.never quite got the hype surrounding Gullfoss.the way it puts back energy doesn't sound "real"to me.to clarify that i mean that the way it performs the processing does not sound like the source has been enhanced and i always wanted to take it off after a few mins and thus never bothered with the purchase after the demo

My vote goes to Soothe

So did they say how much the latency is in Soothe 2?
In Soothe 1 its over 8000 samples in Pro Tools working at 96k. That's just not usable for recording.

If they could get it down to around 100 samples or less that would be great.
Preferrably I would like to se a zero latency mode.

It’s not the sort of plugin you need for recording. It’s a mix tool. I’d just be happy with halving the delay compensation needed - that’d sort me out!

yeah one is for de-harshing sounds and the other is for improving a mix (sometimes) in my experience.


btw love them both and get a lot of use out of them and sometimes even use them both on a track together!

For anybody wondering the latency is 4096 in 96k sessions.
I'm guessing it's 2048 in 48k.

Nice setup very similar to my workflow! Curious what is the WSD1M3 plug on the master?