The No.1 Website for Pro Audio
AI MASTER - Fully Automatic Mastering Plugin
Old 19th September 2020
  #211
Lives for gear
 

I'd definitely consider purchasing this purely for a frequency response reference.. but that means I really don't want it to do any dynamics stuff at all. For frequency response reference this thing would be amazing when working long hours on stuff like film/tv scores and similar stuff, where you are just grinding through tons of material and easily get affected by ear fatigue.

So yeah.. a selfish feature request (as I'm not even a customer yet): The ability to remove the limiting/compression from the equation. I'd also like to be able to remove any widening that happens.
Old 19th September 2020 | Show parent
  #212
Lives for gear
 
zeljkom's Avatar
Quote:
Originally Posted by bmanic View Post
I'd definitely consider purchasing this purely for a frequency response reference.. but that means I really don't want it to do any dynamics stuff at all. For frequency response reference this thing would be amazing when working long hours on stuff like film/tv scores and similar stuff, where you are just grinding through tons of material and easily get affected by ear fatigue.

So yeah.. a selfish feature request (as I'm not even a customer yet): The ability to remove the limiting/compression from the equation. I'd also like to be able to remove any widening that happens.
Or even better : Separate slider (wet / dry) for limiting/compressor, eq and widening. Then it's a winner.
Old 20th September 2020 | Show parent
  #213
Gear Nut
 
Guy Rotem's Avatar
 

Thanks everyone for the great suggestions and feature requests, keep'em coming!
Old 20th September 2020 | Show parent
  #214
Lives for gear
Just discovered this. Very intrigued. Guy - I have an idea/request for your future project. Perhaps you could make a separate version of this for film post people. A dialog processing plugin. Something that you put on your dialog bus and it analyses incoming dialog/voice/clips/phrases in a film scenario. The plugin puts the dialog somewhere in the ballpark frequency and level wise. This would be ideal for client preview purposes. Say, when you're editing dialog and sending out previews to clients, you don't always have time to EQ and clip gain everything - so you get boomy lavs on one phrase, nasal boom recordings, too loud, too soft etc. This makes it difficult for directors or producers to evaluate the overall mix. Slamming a compressor or dynamic EQ is crude and hardly works. So, something that levels it out would be such a time saver. Not for final mixes, but for such preview situations, or for reality/TV/interview/promo productions. Make it AAX Pro Tools compatible and many will buy it.
Thank you an keep up the great work!
Old 20th September 2020 | Show parent
  #215
Gear Nut
 

Off course there should be more controls but until then there's a way to go around that. If AI Master compresses to much then you just have to lower the reduce a dB or two in to the plugin when the analyze is done. If its too bright, reduce high frequencies.
Old 20th September 2020 | Show parent
  #216
Lives for gear
 
zeljkom's Avatar
Quote:
Originally Posted by MickeUppsala View Post
Off course there should be more controls but until then there's a way to go around that. If AI Master compresses to much then you just have to lower the reduce a dB or two in to the plugin when the analyze is done. If its too bright, reduce high frequencies.
Yeah, but dealing with it on the fly would be awesome. There is no way to reduce that stereo spread now. It would be redundant if I have to do it afterwards, no?
Old 20th September 2020 | Show parent
  #217
Gear Nut
 

Quote:
Originally Posted by zeljkom View Post
Yeah, but dealing with it on the fly would be awesome. There is no way to reduce that stereo spread now. It would be redundant if I have to do it afterwards, no?
Yes that's a clear exception. The more stereo spread the more phase issues, so there's no good way to undo that.
Old 23rd September 2020
  #218
Gear Addict
 
pipelineaudio's Avatar
 

So will this thing act different depending on the level you send it? Seems like that is something it should be able to detect and deal with.
Old 23rd September 2020 | Show parent
  #219
Lives for gear
 

Quote:
Originally Posted by pipelineaudio View Post
So will this thing act different depending on the level you send it? Seems like that is something it should be able to detect and deal with.
I think the dev recommends hitting AI Master at -6db during the analysis and then backing of input to taste once in Master mode.
Old 23rd September 2020
  #220
Lives for gear
 
dirtROBOT's Avatar
Although there's alot of warning about running the signal into it too hot, it has always handled this pretty well for me so far. Not that I'd recommend it but it doesn't go flying off into the sun or anything.
Old 23rd September 2020
  #221
Lives for gear
 

What happens when you send the audio through twice? Does it not make any changes the second time? Haven't demo yet
Old 23rd September 2020 | Show parent
  #222
Lives for gear
 
dirtROBOT's Avatar
Quote:
Originally Posted by trevon View Post
What happens when you send the audio through twice? Does it not make any changes the second time? Haven't demo yet
it quotes satanic blasphemies
Old 23rd September 2020 | Show parent
  #223
Lives for gear
 

Quote:
Originally Posted by trevon View Post
What happens when you send the audio through twice? Does it not make any changes the second time? Haven't demo yet
I've used it in the mix process as a second opinion. Checked the before and after EQ curve and if I agree and make changes in the mix I've noticed that AI Master will apply less processing as far as EQ is concerned.

I haven't run a mix twice without changes to check consistency in the processing. That is a rabbit hole I don't wanna fall into (because I know I would...)
Old 23rd September 2020 | Show parent
  #224
Lives for gear
 
dirtROBOT's Avatar
Quote:
Originally Posted by JfromRVA View Post
I've used it in the mix process as a second opinion. Checked the before and after EQ curve and if I agree and make changes in the mix I've noticed that AI Master will apply less processing as far as EQ is concerned.

I haven't run a mix twice without changes to check consistency in the processing. That is a rabbit hole I don't wanna fall into (because I know I would...)
My guess is this is a neural net style where the program is fed a bunch of reference material, generates a network of solutions and then does a closest match to the input sound and figures out the best solution to achieve closest match. So sending it in once, solution x is determined, sending it in again would have it determine solution x with some variants due to how neural nets work.

I'm actually curious if the AI is dynamic enough to change things up on the fly or if it can only apply changes once to the whole song?
Old 23rd September 2020 | Show parent
  #225
Lives for gear
 

Quote:
Originally Posted by dirtROBOT View Post
My guess is this is a neural net style where the program is fed a bunch of reference material, generates a network of solutions and then does a closest match to the input sound and figures out the best solution to achieve closest match. So sending it in once, solution x is determined, sending it in again would have it determine solution x with some variants due to how neural nets work.

I'm actually curious if the AI is dynamic enough to change things up on the fly or if it can only apply changes once to the whole song?
Would be nice to know if you send it through again after it processed the master if it would leave it unchanged or if you send in a raw mix through two times if it give two different versions or the exact same.
Old 23rd September 2020 | Show parent
  #226
Gear Nut
Quote:
Originally Posted by dirtROBOT View Post
My guess is this is a neural net style where the program is fed a bunch of reference material, generates a network of solutions and then does a closest match to the input sound and figures out the best solution to achieve closest match. So sending it in once, solution x is determined, sending it in again would have it determine solution x with some variants due to how neural nets work.

I'm actually curious if the AI is dynamic enough to change things up on the fly or if it can only apply changes once to the whole song?
I've been toying with the plugin nonstop since purchase and one thing I'm not seeing anyone else mention is to make changes in the chain AFTER AI master processes the track.

I've been adding additional saturation and eq adjustments once I like what AI master does. These go before AI master and it does seem to react to these changes in a dynamic way that generally remains faithful to the original pass it took, if that makes.

Something like adding in Bassroom and making changes there while AI Master is on master mode. A game change in this thread that I have incorporated that does work a bit better is Guy's mention of lowering the level of the track before it hits the AI if you don't favor the processing intensity of it's dynamics choice. This application of level adjustments yields a much subtler sound out of it, but still imbues some flavor.

Honestly, I've been running my tracks through some of my outboard gear before they hit AI Master as well and am a bit suprised how close it can get to the analog sound with and without the HW.

My biggest gripe is it tends to go a bit heavy on the stereo spreading every time. To combat this, I use a stereo editor to compensate where AI widened too much.
Old 23rd September 2020 | Show parent
  #227
Lives for gear
 
dirtROBOT's Avatar
Quote:
Originally Posted by Sance21 View Post
I've been toying with the plugin nonstop since purchase and one thing I'm not seeing anyone else mention is to make changes in the chain AFTER AI master processes the track.

I've been adding additional saturation and eq adjustments once I like what AI master does. These go before AI master and it does seem to react to these changes in a dynamic way that generally remains faithful to the original pass it took, if that makes.

Something like adding in Bassroom and making changes there while AI Master is on master mode. A game change in this thread that I have incorporated that does work a bit better is Guy's mention of lowering the of the track before it hit the AI if you don't favor the processing intensity of it's dynamically choice. This application if level adjustments yields a much subtler sound out of it, but still imbues some flavor.
interesting, so you didn't re-analyze when you added to the input chain?
Old 23rd September 2020 | Show parent
  #228
Gear Nut
Quote:
Originally Posted by dirtROBOT View Post
interesting, so you didn't re-analyze when you added to the input chain?
Sometimes I create the chain I like before it Analyzes, but I tend to be ok with its initial pass and make adjustments from there.

One issue mentioned earlier in the thread was the loudness not being enough after it masters. I've put a limiter/clipper after AI master in the chain and that gets me where I need to be, if necessary, without problem.
Old 23rd September 2020 | Show parent
  #229
Lives for gear
 
dirtROBOT's Avatar
Quote:
Originally Posted by Sance21 View Post
Sometimes I create the chain I like before it Analyzes, but I tend to be ok with its initial pass and make adjustments from there.

One issue mentioned earlier in the thread was the loudness not being enough after it masters. I've put a limiter/clipper after AI master in the chain and that gets me where I need to be, if necessary, without problem.
yeah same, I'm targeting -8db not -12db which iirc is what ai master is always trying to achieve
Old 23rd September 2020
  #230
Lives for gear
 

Just so I understand. Are you guys following a brickwall limiter, which is what I hear AI Master applying, with another brickwall limiter? I've used a true peak or brickwall after a limiter without those features. A clipper as well. But not brickwall after brickwall. Am I miss understanding?
Old 24th September 2020 | Show parent
  #231
Lives for gear
 
dirtROBOT's Avatar
Quote:
Originally Posted by JfromRVA View Post
Just so I understand. Are you guys following a brickwall limiter, which is what I hear AI Master applying, with another brickwall limiter? I've used a true peak or brickwall after a limiter without those features. A clipper as well. But not brickwall after brickwall. Am I miss understanding?
I'm using event horizon, which is a transparent lookahead clipper but I'm barely touching it, I"m just tweaking the threshold about 3db.
Old 26th September 2020 | Show parent
  #232
AI FEEDBACK, COMPARISONS and FEATURE REQUEST

Hi Guy,

thanks for creating an interesting plugin. I bought Exonic AI after a brief test and have been testing it more thoroughly alongside my usual UAD setup and did a brief comparison with Ozone 9 suite.

An interesting phenomenon is to have SoundTheory GULLFOSS AFTER your plugin and then switch off the AI at intervals. Gullfoss sure has a lot less to tame and recover when your plugin is active.

I do however find the limiter a bit aggressive , especially on acoustic material, and therefore would join the previous posts in asking for a possibility to either choose an IRC and "profiles" of sorts like True Peak or be able to bypass that aspect.

The way the Exonic AI deals with the stereo field is also interesting : It reminds me of certain settings in the Ambience Recovery tool K-Stereo from Bob Katz, which I use on almost every session with acoustic material.

I do think it would be nice to be able to show and hide the robot please. It is distracting on the screen after the novelty wears off. Also strange to see it do a "beat" thing on classical guitar with Cello and some strings ;-) The numbers during analysis, while funny, might also be replaced by something more useful like an accumulation of the "red collars" of the robot i.e. "overloads" with a timeline indicating how much to temper the Mix-Stage.

Over all : Glad I bought it. I am setting it up as a reference next to my usual mastering AUX track in Logic Pro X and can switch on the fly between them with option-solo. It is frighteningly close to all the hard work done in excellent plugins from UAD and Fabfilter... In certain cases sofar maybe a tad more "muffled" or with less prominent mid-highs for acoustic material. But that is a short term observation that needs to be tested against much more audio.

Anyway. Hope this helps. And I hope you have the time and financial success to listen to the feature requests from your fans. Excellent choice of Price Point too BTW.

kind regards

Frans
Old 26th September 2020 | Show parent
  #233
Company Rep
 
Nonlinear's Avatar
 

The problem I’ve seen with products like this is that users immediately begin requesting all kinds of “tweakability” features. Adjustable limiting/loudness, adjustable EQ contours, adjustable this and that.

When all of that is incorporated what you end up with is essentially MANUAL mastering, i.e., the user is making all the choices. So the “AI” concept makes little sense. At best it can be “AI assisted” like Ozone, etc.
Old 26th September 2020
  #234
Lives for gear
 
robert82's Avatar
Well what the heck kind of product rollout would it be if Gearslutters didn't immediately chime in with advice/complaints/suggestions and of course, compatibility issues and requests for intro deals!
Old 26th September 2020 | Show parent
  #235
Lives for gear
 

Quote:
Originally Posted by Nonlinear View Post
The problem I’ve seen with products like this is that users immediately begin requesting all kinds of “tweakability” features. Adjustable limiting/loudness, adjustable EQ contours, adjustable this and that.

When all of that is incorporated what you end up with is essentially MANUAL mastering, i.e., the user is making all the choices. So the “AI” concept makes little sense. At best it can be “AI assisted” like Ozone, etc.
If the dev considers incorporating some user control and allows the AI to learn our preferences then it will become a must have tool.
Old 26th September 2020 | Show parent
  #236
Lives for gear
 
dirtROBOT's Avatar
Quote:
Originally Posted by Nonlinear View Post
The problem I’ve seen with products like this is that users immediately begin requesting all kinds of “tweakability” features. Adjustable limiting/loudness, adjustable EQ contours, adjustable this and that.

When all of that is incorporated what you end up with is essentially MANUAL mastering, i.e., the user is making all the choices. So the “AI” concept makes little sense. At best it can be “AI assisted” like Ozone, etc.
lol agreed, I just want to turn the limiter/clipper on/off and the stereo imager off, I don't want to adjust anything tbh
Old 27th September 2020 | Show parent
  #237
Lives for gear
 
dirtROBOT's Avatar
Quote:
Originally Posted by JfromRVA View Post
If the dev considers incorporating some user control and allows the AI to learn our preferences then it will become a must have tool.
I don't want it to learn from my preferences because I"m not very good at mixing. Maybe your idea of preferences is different from mine tho
Old 27th September 2020
  #238
Lives for gear
 
Space1999's Avatar
 

Attention Borg Collective!

I really do think you have something here, especially for those who could not afford mastering and need help. The reviews here look positive and I am one to embrace new technology.

But OMG you have to change your Uber Creepy AI assistant. Seriously, that Avatar gives me night terrors.

Couldn’t it just be a human looking Avatar?

Sincerely,

Captain Picard
Old 27th September 2020
  #239
Lives for gear
 
dirtROBOT's Avatar
if you get rid of the robot then NO ONE will dance to my music
Old 27th September 2020 | Show parent
  #240
Gear Nut
 
Guy Rotem's Avatar
 

I made this exhaustive tutorial that demonstrates how to remove the robot graphics after you've switched from 'ANALYSE' to 'MASTER'.



Hope that helps
📝 Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 59 views: 8758
Avatar for murphythecat87
murphythecat87 30th August 2018
replies: 87 views: 18817
Avatar for Jerry Tubb
Jerry Tubb 3rd January 2019
replies: 56 views: 5289
Avatar for BIG BUDDHA
BIG BUDDHA 1st June 2019
replies: 7 views: 1135
Avatar for Guy Rotem
Guy Rotem 16th October 2020
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
🖨️ Show Printable Version
✉️ Email this Page
🔍 Search thread
🎙️ View mentioned gear
Forum Jump
Forum Jump