This really has almost nothing to do with the question, which itself had almost nothing to do with "older audio formats". On Thu, Oct 13, 2016 at 2:29 PM, pierre jocelyn andre <temps.jo(a)gmail.com> wrote:

On Thu, 13 Oct 2016 20:39:02 +0200, Markus Seeber wrote: On Thu, 13 Oct 2016 14:39:24 -0400, Paul Davis wrote: At least proprietary software claims to connect audio, MIDI and sync of the provided different formats, by newer hosts with more precision and better assignment abilities, than it is possible with established apps. Today I read about http://musicappblog.com/aum-review/ , https://itunes.apple.com/us/app/aum-audio-mixer/id1055636344?mt=8 . I don't know if this is just another host/mixer with some gimmicks, IOW just something we could do with the provided software on all platforms, within the limitations of the supported formats, or if it introduces something new and useful. Perhaps I purchase it tomorrow, OTOH I also decided to install Carla a while back and in the end I never used it. [rocketmouse@archlinux ~]$ grep carla /var/log/pacman.log [2015-08-28 10:17] [PACMAN] Running 'pacman -U carla-git-1.9.6.r239.g876cf81-2-x86_64.pkg.tar.xz' [2015-08-28 10:17] [ALPM] installed carla-git (1.9.6.r239.g876cf81-2) [2015-08-28 10:47] [PACMAN] Running 'pacman --color auto -U /tmp/yaourt-tmp-rocketmouse/PKGDEST.I9q/carla-bridges-win-2.0beta4-1-x86_64.pkg.tar.xz' [2015-08-28 10:47] [ALPM] installed carla-bridges-win (2.0beta4-1) Maybe the proprietary, as well as the free Linux formats already provide more interconnectedness than we expect, resp. need. Regards, Ralf

On Thu, 13 Oct 2016 20:54:54 +0100 Will Godfrey <willgodfrey(a)musically.me.uk> wrote: This is absolutely true and also makes the biggest challenge. OTOH, the networking of plugins in order to enable them taking decisions collectively, is geared at the mixing process. There also, even though the bird's song was collected not too badly, there is also room for creativity. It would then be to lighten up the mixing task to better concentrate of artistic expression, that the plugins would act as a group. I'm not saying I approve of this.

On Thu, 13 Oct 2016 14:39:24 -0400 Paul Davis <paul(a)linuxaudiosystems.com> wrote: For instance all plugins within a track would communicate between themselves to take decisions. Likewise when they're on a stem or the master buss. Hmmm... Network sockets, to take a common parallel, could be created by plugins. After all this is about communication between plugins. Not sure that a dedicated DAW would be needed.

On Thu, 13 Oct 2016 16:40:15 -0400 Paul Davis <paul(a)linuxaudiosystems.com> wrote: Good. That's how things can go forward :)

On Thu, 13 Oct 2016 15:43:56 -0500 "Chris Caudle" <chris(a)chriscaudle.org> wrote: The closest practical example would be iZotope's Track Assitant/Neutron released last week.

On Thu, 13 Oct 2016 21:12:23 +0200 Robin Gareus <robin(a)gareus.org> wrote: What would then be an alternative ? To switch DAWs because of the processing they offer ? Something akin to Artificial Intelligence. A small 'decisional' (don't have the right word here) network, each node providing its own view on the subject, with a decision taken as a common accord.

On Thu, 13 Oct 2016 20:55:53 +0200 fred <f.rech(a)yahoo.fr> wrote: Well, it's all about users of the software. The idea with this eventually, is to impact the mixing engineer work in such a way as to have decisions being taken using (distributed) artificial intelligence and networked plugin nodes. Why not talk about it at the user level ?

On Fri, 14 Oct 2016 01:06:44 +0200 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote: Now where did you get the notion that this is about creating music ? It is about plugins. I might be rustic, but so far I do not use plugins in creating music and if I did, I would treat it as a device of some sort, artificial intelligence or not. I do use plugins FX to add characteristics to sounds and that kind of use would still be with Flintstone-era (!) plugins, and not the fancy AI-like ones which are much more for mixing purposes in which case distortion would be more like using tape saturation than a Big Muff. Hey, don't give me ideas for a next tune. These are not the same. True, there is art in audio engineering. And mastering. Only recently have I started to hear the subtle difference of a Fairchild 670 barely compressing on a drum/bass stem. The sum of subtle impressions, their interrelations, and how to drive them in a direction or the other is part of the art. Same would be with those AI-like networked plugins. Maybe a bit like a grammatical/spelling corrector. You still write the text. And then there would be cloud mixing. There must be a way to introduce the concept of 'clouds' in there. Plugins that subscribe to a cloud in which they can download parameter data based on audio/genre characteristics. The mixing engineer throw a set of tracks to them, they find out the instruments,they find out the style, the target reference, then they go hunt for 'experienced parameters' in the cloud, always according to the engineer wishes. There wouldn't be any 'style police' that would prevent an engineer to slam-compress a ballad.

On Fri, 14 Oct 2016 09:19:27 +0200 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote: [snip Track Assistant text] Since this is about mixing, processing audio, then it should apply to every audio material. It does not do that. This is extrapolation. This said, in mass consumer musical products there are products who will assist the player in rendering melodies, in various ways. I think even Roland has addressed this market. An overdrive component is not included in mixing. This is out of boudary, at least AFAICS. Tape saturation, yes. Overdrive, no. Overdrive, flanger, etc, are FX. Whereas a mixing engineer can decide to add some to a track on his own will (or rather after consulting with the client), they are clearly out of the mixing realm. Mixing happens after these FX are added, so to speak. So it helps to clarify the role of interconnected, decision-taking mixing plugins by keeping out squarely what is not about mixing. No, the machines are not taking over. And from interaction of parameters, to assisted decision taking. A template will obviously not recognize the audio it is being applied on since it has no intelligence, since it is static. This is a large difference with the approach sketched here, for the enjoyment of a discussion.

On Fri, 14 Oct 2016 13:07:38 +0200 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote: It's a balance of the two, mostly. The end result has to be in conjunction with the remaining of the tracks but then, it can very well be that a single track by itself does need so adjustments no matter what. This said I am certainly not going into the details of how Track Assistant works as the subject as I defined it, is not about that. I think it steers away from the subject. So I will not address it. If this is what Track Assistant does, then that's their choice. To continue on this 'gimmick' would have to be defined. Eg. is this out of TA or is it an example you bring forward on your own ? Because, as with, if you wish, Pianoteq sounds, the selection does not exist before the action of creating it is being made. It is created on the spot. The selection does not exist. Because the 'selected templates of settings' does not exist.

On Fri, 14 Oct 2016 14:27:08 +0200 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote: Pianoteq is only a parallel to demonstrate a principle. The finger pointing at the moon, or the moon ? I brought it up as a parallel that it creates sounds, it does not trigger existing samples (at least as far as I understand it works). The sounds do not exist until they are created. Likewise, the 'selection of effects' do not exist before it is created. With all the propension to put forth 'art' as an opposition to machines controlling expression of musical ideas, one would have thought that the difference between something that already exists and something that is about to be created would have been clear. And from this clarity, move on with the discussion. It is not. It is the finger. And now the quality of code. It would be so easy to start shooting around on this, so easy that I wonder if this is not tongue-in-cheek. And so on so forth... GIMP: It's really up to the individual, really. Some, it seems, can get easily influenced to become lazy. Others do not. The finger, or the moon. To go back on track, for people who might be happening to only read this or a couple of messages before, the subject I brought up for discussion, was the concept of audio plugins, for mixing purposes, being able to network together to take decisions concerning a track, a stem, a master buss. As a related recent example in that direction, and although not implementing network communications between plugins, iZotope's Track Assistant/Neutron.

On Sat, 15 Oct 2016 11:38:23 +0000 Fons Adriaensen <fons(a)linuxaudio.org> wrote: What is amusing is how this can be distorted. Speak about processing. But this is interesting, this reaction. And so on for any other machine that can detect something automatically in every day life: the user is a moron. An automatic transmission would be one example. So if there is any kind of audio analysis going on, then it means that the "audio engineer" is actually really caught between apostrophes. Any type of analysis, whether or not this analysis includes identifying an instrument. What's the point ? Just like a human. How many mixes can you do of a set of track. Unless that human thinks he's perfect, of course. In wither case it is amusing. (Ralf) A discussion maybe ? But we got it after all, isn't it.

On Tue, 18 Oct 2016 21:01:44 +0200 Emanuele Rusconi <emarsk(a)gmail.com> wrote: Hello, A discussion. Well, if the whole sentence was taken, there could perhaps be more sense. The whole sentence was: "Plugins that network themselves together, sharing audio analysis, sharing data, taking decisions together on how to apply themselves as a group to an audio track, to stems, to a whole session." The quality, or feature, would be applied in those three contexts. The feature, or functionality, itself consists of the sharing of audio analysis. I disagree on that, based on so many protocols existing out there that are not proprietary, that can be used openly. There is no need to do mind analysis and it is actually very surprising that this is brought up, as some kind of B-series SF flick. I do not know if, when the first automatic transmissions were proposed, that some people have exclaimed "oh ! we will loose control, we need to stick (pun ?) to manually shifting the speeds as the auto transmission does not know". As we know today automatic transmissions do respond to what the driver wants, in their own way. Floor it, and it will downshift. Likewise with anything else. This is no 'Robot Mixer Invaders From Loudness Hell' type of thing. To assume that engineers will de facto loose total control is, totally, absurd. Again. Slight variation here, is that this is about the mixing stage. When people go and hire a mix engineer, most of them have figured out that the guitar part will be electric and have a Big Muff and will not be an acoustic 12-string. This is already done. The colours and tones and all that is part of the composition and happens at an earlier stage, in very large parts. So this is about the mixing stage, where the composition of overall tone is already made. The mixing engineer can modify some things, but he is basically a stranger to the client's project. Yes there are actually tons of these, and tons of discussion. It's even possible to buy as a kit, the same equipment that Flea uses, bass, amp, speakers, pedals, and all that. Some people are totally dedicated to take apart the sound configuration of some of the famous musicians out there. Your example above stems from Guitarix. This happens way before the mixing engineer is on the payroll. Hence the comparison is not valid at all. Keeping it simple, let's put forward some examples. As you might have guessed by now, it is based on principles. So let's bring one: Separation. Clean separation of instruments in the audio field is a must. You cannot do without it, unless the mix is meant to be leaning towards mud, which can be a style on its own. Do you see how plugins communicating between themselves could set up a basic separation framework to assist the mixing engineer ? Another one: Three dimensional placement of objects. Likewise here, plugins that can share data, and based of what they did about separation, can establish, *based on configuration styles or otherwise*, a basic placement of objects for the mixing engineer to refine. These two tasks have large technical parts to them and yet, they have to be done for every single mix. In Mixbus now there is the analysis of phases between tracks, analysis that will actually optimize the phase settings. This is a simple, uncomplicated (in its own way) step towards having more of the same. No-one has said "well, I do want the left room mic for the crash to be out of phase by that much so this thing in Ardour is total crap." Nope. And from the phase feature in Ardour/Mixbus we can see why people would think you need a DAW. No, it would be about extending and sharing analysis amongst plugins, based on an open protocol that allows for decision-making. Well, the way it goes, a company will do it and then 3 years later, or 5, there will be a parallel open source project. :) What I'm trying to achieve ? Just a discussion. Cheers.

On Tue, 18 Oct 2016 16:45:10 -0400 Paul Davis <paul(a)linuxaudiosystems.com> wrote: Well, if one uses OSPF to try to do VRRP...

On Tue, 18 Oct 2016 17:16:09 -0400 Robert Edge <thumbknucklerocks(a)gmail.com> wrote: You mean, to prototype a discussion ? Obviously. The guy who first thought about a wagon thought oh, it'd be so nice to have something that would be able to carry all these things around, it could carry so many, so many things, very big things, over such great distances. And then the guy besides snipped, we only have donkeys, this will never work, you see, this will never work. So it depends.

A stab in the dark, and at the risk of grossly oversimplifying the task of audio engineering: Let's say you have N tracks, each with mystery TheIntelliNetExcitingAutoReacta(TM) plugin (TINEAR). TINEAR analyzes the spectra of each track, and, on the fly, based on the spectra of all other tracks, adjusts appropriate frequency bands so each track occupies a predefined spectral niche. Similarly, MadUberDimensionializer(TM) plugin MUD does the same but for spatial separation. On Wed, Oct 19, 2016 at 11:16 AM, Robert Edge <thumbknucklerocks(a)gmail.com> wrote:

On Wed, 19 Oct 2016 08:14:36 -0400, jonetsu wrote: 3D location a basic mixing task? At best I would call panning a basic mixing task.

On Tue, 18 Oct 2016 23:27:53 +0200 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote: Bloatware. When you only have donkeys, do not dream of horses. Because horses are just bloated donkeys. First of all, you would not start necessarily with EQ, but that's another matter. In any case, using EQ and using presets would be not very smart isn't it ? Not smart at all. EQ presets. Wow. As a swag, after the first object location pass is done, based on the style wished as a basic framework for 3D locations in the sonic field, the knowledge about what was done is kept, which includes many aspects such as what are the instruments and characteristics of each track. This knowledge would then be used to help is establishing separation. Now, a basic question here would be, where is that knowledge to be stored ? That knowledge that was created by the joint analysis of several plugins and modules. Making it a distributed knowledge would not be realistic (TM) at all.

On Wed, 19 Oct 2016 00:50:17 +0200 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote: Ah, the ultimate side tracking starts to happen. It is not possible to state 'EQ' at large, as we understand EQ to be. It is subtractive EQ to remove undesired frequencies on which there is no further work to be done.

On Wed, 19 Oct 2016 01:02:11 +0200 Robin Gareus <robin(a)gareus.org> wrote: Again, it is not about creating personalized tones so much. It is at the mixing stage, where hopefully, when the mixing engineer is on the payroll at $300 an hour (highly hypothetic), the decisions about the tones of the tune have been taken. By humans. At the mixing stage, to assist the mixing engineer. Of course. Naturally. Cool, this DeepMind thing. Yes, this is very likely how it will start to be known. At least it will not be part of a specific DAW, but as a plugin, will be OK for all DAWs out there.

On Wed, 19 Oct 2016 02:12:07 +0200, Ralf Mardorf wrote: I don't expect that in the year 3016 Nissin instant noodle soup further development will make it tasting as good as noodle soup freshly made by a top chef.

Le 19/10/2016 01:25, jonetsu a écrit : Is Picard that frozen food brand Robin? ;)) What you need if you are paid $300/h is what is called a B.Boy in a studio: the guy who goes to the "hardware plugins wall" and turns the buttons for ya! Optionnally, the same best boy will have done the coffee, and maybe he have rolled those hilarous cigarettes! Every plugin, AFAIK, can save presets. Are you trying to build "meta-presets" for a bunch of plugins presets Jonetsu? And are you willing those meta-presets to dynamically evolve, depending on audio tracks analysis, and having an infinite option box where you can check "type of music produced/ kind of sound desired/ loudness to achieve/ signal - noise ratio/ track 17 is solo1 at 1:32/ track 18 is solo2 at 2:45/ end of song is a 20" fade out/ not too much reverb except on female choir/ etc..? Am not sure it's a cool thing (especially for the $300/h guy)) that machines could replace humans for all. More, as heard in some medias, it's kind of danger... Have fun, trying stuffs to give "life" to tracks! -- Fred,

On Wed, 19 Oct 2016 08:28:29 -0400, jonetsu wrote: What is a 3D basic placement? The analyser does notice a female voice and a male voice, so the analyser assumes a difference in height? The man is 15 cm higher grown, than the women and some 3D plugin will place a virtual man's mouth 15cm above the virtual woman's mouth? The 3D placing plugin is network connected with a reverb and a reverb algorithm optimises the reverb in a way that it plays nicely in regards to the size difference? Others and ... or at least I, don't understand your vague insinuations. Assuming there should be suggestions done by an analysis tool, it safes some minutes or hours of work? Regards, Ralf

On Wed, 19 Oct 2016 09:41:16 -0400 jonetsu <jonetsu(a)teksavvy.com> wrote: Frequency density has to do with up/down placement. Not the amount of high frequencies, for instance, but their density. As with any other thing in audio: in relation with the other audio events.

On Wed, 19 Oct 2016 16:56:53 +0200 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote: As long as castration is not ever considered. :)

On Wed, 19 Oct 2016 09:41:16 -0400, jonetsu wrote: Then simply try to answer this question: On Tue, 18 Oct 2016 22:16:38 -0400, Robert Edge wrote: You replied: On Wed, 19 Oct 2016 08:14:36 -0400, jonetsu wrote: But it does not explain what the network protocol is used for. IIUC you want an analysis tool, that does part of the work an audio engineer does. IIUC you think an audio engineer needs more time to do it, than a tool does. The analysis tool should load tools and talk to the tools, to provide good settings. Based on what? How should this work? If an engineer listens to sound and should describe another engineer what to do, who doesn't listen to the sound, it already is unlikely to get a good result. How should the analysing tool find the right words, to explain the tools what they need to do? I got a lot of jobs, because film students, who aren't audio engineers, try to do the audio engineering for a film maker and after that there was the need to hire another audio engineer or me to fix their mistakes. Fixing mistakes takes much longer, than doing the audio engineering correct in the first place. It's my conviction, that analysing tools gain you nothing, but only introduce bad decisions, at the cost of work/time an audio engineer needs to fix those wrong decisions. Please, try to give a serious answer to Robert's question. What problem do you try to solve? IMO your vague idea would cause issues. Regards, Ralf

On Wed, 19 Oct 2016 17:25:34 +0200 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote: See the reply by Michael Noble then. I'm afraid examples were given and there is no need to repeat ad vitam since after all, the text is there and can be re-read. You may call then 'settings' although IMHO it is a bit more than that if only because when using the term 'settings' there is a heavy connection being dragged along which consists of 100% static definitions that exists before the task has started. And this was also mentioned previously.

On 10/19/2016 01:25 AM, jonetsu wrote: For deep learning, the analysis tool itself would be the DAW in this case which has context: All tracks, random data access, and it can control the plugins. There was an endeavor to standardize a very basic part: Allow a plugin to analyze a complete track and write its own automation data: http://lists.lv2plug.in/pipermail/devel-lv2plug.in/2016-February/001571.html Something like this would be a very basic building block: the plugin knows everything about itself, can do analysis and then "control" itself. Even without fancy inter-plugin communication this could go a long way: gates, adaptive de-esser, pitch-correction, noise cancellation, heck even a channelstip plugin. There are working prototype implementations (host & example plugin) but so far it came to nothing. Go figure. The actual use-cases were eventually implemented as VAMP-plugin (Polarity optimizer) and plugin-internally (XT-TG Tom Gate). ciao, robin

I think a robot that controls a DAW is an alternative and simplest solution to avoid a complex DAW with specific AI, shared memory, etc for a network of plugins. For example, the robot knows how to read the session file: it reads that file to discover the network of the used plugins. The robot knows how to control the DAW (i.e. MIDI and/or OSC messages for transport control, [un]bypass the plugins, move faders, panpot, etc), therefore it changes some parameters and gets feedback with audio data (i.e. from master output). It analyzes the output and sends messages to mute tracks, bypass plugins, alter parameters and automation). Play, stop, rewind, continue and jump. The CPU limits the speed. In practice, it is necessary at least: 1) A comprehensible session file (simplification: external conversion to a standard session file). 2) A controllable DAW (MIDI, OSC, etc). 3) Algorithms for artificial listening: culture, aesthetics and psychoacoustic are important elements and the complexity depends on the requirements. The "artistic" sensibility is difficult to simulate (also for humans), however I think that robot is probably a friendly (*) assistant for a person without experience. A simple case: the human requires a mix without to use equalizers or dynamics processing. The robot knows the style of the music a priori or it is able to discover the musical genre. If the robot works fine, the human get a decent automatic mix and, possibly, some suggestions about applicable EQ, dynamics processing, re-record some tracks, the alternative mic to use, etc. The human asks to apply some DSP, the robot provides a new mix, etc.. (*) "ha un cuore di bambino che non si rompe ma a a i" https://www.youtube.com/watch?v=zDzieioM05g

On Wed, 19 Oct 2016 18:54:33 +0200, Tito Latini wrote: It could be useful, for somebody who might be a good musician and spending the time with making music, but not with mixing the music. I can't play drums, but can play guitar, a drum computer could be helpful. We could have friends, but don't need to have a friend able to play drums. There's nothing wrong with this. However, the main goal of audio software should provide tools to make and mix music, but keeps artistic experience in the hands of the users. Replacing an orchestra by a computer allows those who can't pay an orchestra and the room for the orchestra, to hear their compositions for orchestra, more or less close to a real orchestra. Others might have the money to pay for the orchestra and the location, but can't compose for an orchestra. On the one hand nothing is speaking against a tool that helps a musician able to compose for a band, to rewrite the score for an orchestra ... _but_ ... IMO fist of all we should focus on tools that help us to compose and accomplish our own ideas, without assistance in artistic experience. The purpose of analysing recordings to assist in doing the mix, seems not that much needed as other helpers. Regards, Ralf

On Wed, 19 Oct 2016 18:54:33 +0200 Tito Latini <tito.01beta(a)gmail.com> wrote: Another thought about this concerns the work done by the mixing engineer assistant. When a client has tracks for a mixing engineer with a Name, then the assistant, a human, will set up the mix. The engineer will not do that. The assistant, according to the engineer wishes, will organize the tracks in a preferred way, set up stems, do basic clean up of tracks, even perhaps some multing. Spending those hours, even more, to set up the mix up so that when the engineer comes in, everything is prepared to do the actual mix work. I do not know if the assistant will also do basic sound placement and separation, but the question about artistic input can still be valid: to what extent the assistant work can be considered as such ?

On Wed, 19 Oct 2016 14:27:24 -0400, jonetsu wrote: The audio engineer with the promotionally effective name demands another audio engineer for this kind of assistance. IOW a human does the mix and another human does the fine tuning. This is quasi the counterpart of what Fons already mentioned. Let iZotope's Track Assitant/Neutron do the mix, then use iZotope's Track Assitant/Neutron to do the fine tuning of the mix it produced before. While you assume that a tool could do the raw mix and an audio engineer finishes the mix, my claim is, that this only would be possible, if it could be done the other way around, too. A human audio engineer does the raw mix and the tool must be able to do the fine tuning. If the tool shouldn't be able to do the fine tuning, after a human already did the raw mix, then it's impossible for the tool to do a usable raw mix. Regards, Ralf

Of course audio engineers would think something which could potentially put them out of work was stupid (or something they *think* could never put them out of work). What does the AI community think of it though? On Wed, Oct 19, 2016 at 4:40 PM, Fons Adriaensen <fons(a)linuxaudio.org> wrote:

On Wed, 19 Oct 2016 20:51:52 +0000 Fons Adriaensen <fons(a)linuxaudio.org> wrote: Apart from believing, any substance ? Which would be, of course, aside from anything having to do with AI applied to EQ in the context of this thread here. As a general question, OK. But that's not the context here. Some people might have large laughs when seeing the first rice cookers. I can cook perfect rice, no need for a machine !

On Wed, 19 Oct 2016 21:33:19 +0000 Fons Adriaensen <fons(a)linuxaudio.org> wrote: Sorry, the rice cookers with fuzzy logic are doing an excellent job at cooking many types of rice. Although the human must know how to use them.

On Wed, 19 Oct 2016 21:52:36 +0100 Chris Cannam <cannam(a)all-day-breakfast.com> wrote: Not a recording engineer myself (and never will be) but I do know a couple. When I tried to describe this as best I could to one today he jumped in with "Automation. Great. Provided I'm the one who set it up and know exactly what is happening at all times". -- Will J Godfrey http://www.musically.me.uk Say you have a poem and I have a tune. Exchange them and we can both have a poem, a tune, and a song.

On Wed, 19 Oct 2016 16:43:18 -0400, Jeremy Carter wrote: If this kind of artificial intelligence would have all the required skills, it would work towards an imagined final result. Human audio engineers have different tastes, too, so I don't expect that AI needs to satisfy my taste, but I expect that such a tool is able to finish a mix in an educated manner. If the tool just suggests something and claims this to be useful starting points, without the ability to finish the mix on its own, then the useful stating points are fraud. If you don't have the skills to finish the mix, how would you know what would be good starting points? It's ridiculous, because we are that far away from the required knowledge about audio and artistic taste, we need to have to program such a tool, with what ever kind of perfect artificial intelligence, that already might be available. It's like preparing a manned spaceflight to Alpha Centauri, while we are still far away from having the skills to do manned spaceflights to Mars. This has nothing to do with being afraid that technology could put audio engineers out of work. Regards, Ralf

On Wed, 19 Oct 2016 17:30:12 -0400, jonetsu wrote: And then? If you provide the analysing tool what you call the "placement", then you only guess you provide what you have got in mind. To make the sound transparent will change the impression of the song. There is no such an order of "placing" and then adjust levels, add compression, without readjusting what was done before, to get what you have in mind. Regards, Ralf

I think that may or may not be useful, depending on how well it works. Declaring such a thing 'THE FUTURE OF AUDIO PLUGINS' and worrying about a standardized protocol is goofy as hell. The OP seems very resistant to the idea of actually doing something, but insists that this is THE WAY FORWARD. That's why people are not taking the discussion seriously. On Wed, Oct 19, 2016 at 4:40 PM, Fons Adriaensen <fons(a)linuxaudio.org> wrote:

On Wed, 19 Oct 2016 20:40:43 +0000 Fons Adriaensen <fons(a)linuxaudio.org> wrote: I would laugh too. EQ, in the general sense, is not used for such purposes as sketched here. Subtractive, yes, perhaps. Since it might be fairly reasonable to identify what to ditch out at the beginning of any mix job. But then that's a perhaps. When transposing the dynamics of this discussion a bit further, then no wonder it's such a mess at the United Nations. Take what this guy said, interpret it w/o fully understanding it, then serve it to comparses for a good laugh.

On Wed, 19 Oct 2016 23:35:58 +0200 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote: Since the beginning the context is not about a complete mix. I wonder...

On Wed, 19 Oct 2016 21:45:30 +0000, Fons Adriaensen wrote: Full ACK. But now I'm unsure what Jonetsu means by the word "placement" and I still don't understand what starting point settings Jonetsu is talking about. However ... On Wed, 19 Oct 2016 17:40:58 -0400, jonetsu wrote: ... my concerns are related to exactly this attitude. If the tool starts mixing, then why doesn't it finish the mix? Assuming it doesn't finish the mix, because it has got no idea how to do it, then I wonder how it could provide an unfinished mix, that should be useful to finish the mix. If EQs are not used at this stage of the mix, then I wonder what tools should be used. If e.g. only faders should get commands from an analysing tool, I wonder how it decides which track should get what level for the raw mix and what it should gain to make the work easier. Regards, Ralf

On Wed, 19 Oct 2016 18:50:02 -0400, jonetsu wrote: No, I expect my Boss Turbo OverDrive OD-2, just to overdrive the guitar sound. You compare apple and oranges. The overdrive does not play the guitar, but it provides controllers to adjust the overdrive. The kind of automation we are talking about would probably analyse the music and adjust the controller settings. But seemingly the overdrive is used at another stage of the music production, not at the "framework" mixing stage. What are you talking about? On Wed, 19 Oct 2016 18:43:31 -0400, jonetsu wrote: Are you just trolling? What kind of "framework" do you want to get? Trimmer adjustment? 3D positioning, perhaps based on microphones with GPS devices? Please be clear! Could you describe a complete production of a two men project, 1 acoustic guitar and 2 vocals. Which part of the recording and mix are you doing? What is the ominous framework the networked tools provide? And what are you doing to finish the mix? Describe such a session and point out possible pitfalls, or if you prefer a session without any pitfalls, but describe the whole production session with all stages and what is done at what stage by a human and or networked tools. I doubt what that I'm the only one, who completely doesn't understand what you are talking about. Please stop joking. It's completely unclear, if you a serious about anything of this thread. Regards, Ralf

On Thu, 20 Oct 2016 10:20:27 +0200 Thorsten Wilms <self(a)thorstenwilms.com> wrote: The proverb goes like this: “When the wise man points at the Moon, the idiot looks at the finger.”

On Thu, 20 Oct 2016 12:26:56 -0400 Jeremy Carter <jeremy(a)jeremycarter.ca> wrote: There was a study done some time ago (sorry can't remember where) into the relationship between humans and animals. If you point, dogs will *uniquely* look where you are pointing. -- Will J Godfrey http://www.musically.me.uk Say you have a poem and I have a tune. Exchange them and we can both have a poem, a tune, and a song.

On 10/20/2016 05:42 PM, jonetsu wrote: my short-list of "raw" animal intelligence: 1. humans 2. apes 3. house cats 4. all mammals don't mentioned above and i'm not really sure whether nr.3 is there ranked so low :) cheers -- rncbc aka. Rui Nuno Capela

On Thu, 20 Oct 2016 10:26:53 -0700 (PDT) Len Ovens <len(a)ovenwerks.net> wrote: Yes, but this has nothing to do for instance, in the use of compressors since just about radio broadcasting started. And then the use of compressors in order to prevent cutting through the lathe when making LPs. There is a market for 'plugins' (general sense - like saying the 'internet' as in : "I read that on the in-ter-net") as with anything else. But there is also basic 'plugin' use in all main recordings that goes to the market. At this point one can object, nah, I can record things without adding any reverb or anything else, not even EQ and certainly not compression. Sure. But what would be the point in doing so ? I guess we could have a contest consisting of a plugin-free song. If we want to stretch the concept it would have to be recorded in a complete acoustically dead room because after all, exploiting the room acoustics would be akin to add ER or reverb. Which would be forbidden by concept. And absolutely no use of Ardour-Mixbus ! This is to say that 'plugins' are used in recordings and this has at its root nothing to do with marketing and their target, the aspirations of customers.

On Thu, 20 Oct 2016 12:22:19 -0400 jonetsu <jonetsu(a)teksavvy.com> wrote: Or if you wish, on video: https://www.youtube.com/watch?v=4O9o4CKTGzQ (although it is from Confucius originally)

On Thu, 20 Oct 2016 12:36:53 -0400, jonetsu wrote: This is complete nonsense. Which DAW requires to trim the input of the channels for the mastering? To ensure that the inputs are set to line level, is required for the analog domain only. Apart from this, setting all input channels automatically to line level, doesn't require a network connected trim pot. Why exactly did you mention "iZotope's Track Assitant/Neutron" as an example, if it isn't an example? You even won't describe a real world scenario, so it's completely open, if anybody did understand you. Regards, Ralf

On Thu, 20 Oct 2016 14:44:39 -0400, jonetsu wrote: Pff, DAWs in the 70s? You only need to trim for the analog domain. I don't know what and why mixbus provides, perhaps they add a feature to attenuate or boost a signal, but adjustment isn't required. There is nothing like line level required, because it is not analog. You seem not to understand what a trim pot is good for, when using analog mixers. A trim pot for DAW mixers could make sense, if a signal should be too silent, while the fader is at maximum, but a DAW's mixer doesn't require to adjust input levels. This is ridiculous.

On Thu, 20 Oct 2016 20:58:03 +0200 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote: No, consoles, as written above. Thinner and thinner. Is there any shame ? And maybe more while at it. That's it. Enough of your circus.

Have not yet tried but what Ive read atleast nebula needs proper gainstaging because of non linear (in regards to level) processing of signal. Cannot swear on it but would guess atleast some other modeling plugings share similar approach. Anybody know if convolution reverbs treat signal in a non linear manner, and if its strictly internal processing or if multiple impulse responses could be utilized? Would be fun to experiment with morphing verbs. Also, please excuse a nubes ignorance, /n

On Thu, 20 Oct 2016 19:38:30 +0000 Fons Adriaensen <fons(a)linuxaudio.org> wrote: Certainly. They have several uses.

On Thu, 20 Oct 2016 22:49:51 +0200 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote: You are freewheeling wildly. Me: "Anyone doing a mix will start by proper gain staging." To which you first mention that trims are not useful: "Which DAW requires to trim the input of the channels for the mastering?" To which I replied: "There are trim pots on every channel strip in Mixbus" Then: "I don't know what and why mixbus provides, perhaps they add a feature to attenuate or boost a signal, but adjustment isn't required." And then, all by yourself: "Trim pots per se are seldom useful for mixing and especially not in context of this thread. Trim pots linked by networks with AI analysers is absurd." It looks like some kind of over-activity of some sort. Some freewheeling.

On 20.10.2016 23:25, jonetsu wrote: The funny or sad thing is this fits for both you, jonetsu, and Ralf. Taken together, you are not that far from troll-bots playing Ping Pong.

On Thu, Oct 20, 2016 at 12:36:53PM -0400, jonetsu wrote: Which uses the term 'density' 7 times, in all cases as a measure of how many samples are clipped. So what do you mean by 'frequency density' ? And yes, compression can (but doesn't have to) affect perception of transients. What does that have to do with just using EQ and faders ? As should be clear from the context, 'a mix' means using recorded signals. If the recording was done by someone knowing his trade all tracks and hence all inputs to the mixer should already have normal levels. Then there is no 'gain staging' involved at all. Not even on analog mixers. Of course you can continue to reply to any remark with something vaguely related but completely irrelevant, like the two examples above. It will only serve to confirm the rather poor impression you have made so far. Ciao, -- FA A world of exhaustive, reliable metadata would be an utopia. It's also a pipe-dream, founded on self-delusion, nerd hubris and hysterically inflated market opportunities. (Cory Doctorow)

At some point the proper conclusion is not that everyone on LAU is either mentally stunted, or maliciously taunting you, but rather that the original description was not actually clearly stated. I think a more concrete example would be helpful. I don't consider your vaguely worded description of automating 3D placement sufficiently detailed. In the limit that would be something like an automated panner placing each track in a slightly different panned location, which does not sound like something useful or desirable to me. Obviously you had in mind something which you would find useful. Yes, I did look at the Neutron video, it looked like something that would be useful for two months while you first learn how to use a DAW, and that would get turned off after you saw what "choices" it was making and why. If you are going to have to adjust everything it does to fit your taste anyway, might as well make the beginning settings yourself as well. I assume it did not make you think that, so what did you see differently than I did? -- Chris Caudle

On Thu, 20 Oct 2016 15:27:34 -0500 "Chris Caudle" <chris(a)chriscaudle.org> wrote: There is a limit to that. There is a lot of noise that was brought on and after a few times one can wonder, in good conscience, if trolling is not in order. The two examples given (1) were only that. There might be other uses to the principle. This is not about mixing techniques. If there's a need to discuss mixing techniques, then a thread can be started. Does this prospect hinder the trigger-happy reaction of some ? (1) There's also another example, the separation of instruments. That one must be crystal clear, though. as well as the idea of sharing information between plugins to achieve it. Again this goes on mixing techniques which is not the main subject. For instance, a certain framework setup style would have the drums sitting behind the bass. It is possible to spend the time to do that, and people do. But then it could also be useful to have this placement done by computer analysis. As a setup for further work. That was already mentioned. Another framework would contain electric guitars behind the acoustic. This is not, done by EQ or volume which would affect a lot more, especially if used at that stage. The actual physical ground must be set first by shaping the transients according to how sounds are perceived by the brain. An example was asked so this is the closest one that is related. I'm not fond of this neutron. What the principle of sharing audio analysis data will be in 10 years, is not known at the moment, obviously. Now we have seen that related work is done on the polarity analysis in Mixbus as well as the XT-TG plugin inasmuch that they push forward the analysis part. Although these two examples might be too much on the developer side while the scope is on the suer side here.

Jonetsu, when this thread started you gave this example: On Thu, 13 Oct 2016 16:51:25 -0400, jonetsu wrote: Now you claim this is absolutely not what you are talking about and instead of explaining what you are talking about, you start talking about 3D, electricity and guitar playing. Would it be possible to give a 2D example or much better just a stereo or mono example and to avoid absurd statements. If we mention the most important tools for mixing, then don't claim that electricity is a tool, too. It isn't a tool, neither is playing the guitar a tool. Regards, Ralf

On Wed, 19 Oct 2016 21:45:30 +0000 Fons Adriaensen <fons(a)linuxaudio.org> wrote: You also forgot electricity. Electricity plays an important role in mixing and audio. That the faders are the only thing you'd need, sure. As a hobbyist, perhaps. Otherwise anyone else would do proper gain staging. I wonder...

On Sat, 22 Oct 2016 13:51:17 +0200 Emanuele Rusconi <emarsk(a)gmail.com> wrote: Good. I know what I like and I like what I know. Well, in most, if not 98%, cases the client goes to the mixing engineer after he has a set of tracks, or rather, a song. In most cases. And when the mixing is done with the mixing engineer, then the result goes to the mastering engineer. This is how mixing engineers usually earn their living. Ah ? You are free to take a listen to any, and more if you are so inclined. https://soundcloud.com/nominal6/jam12 https://soundcloud.com/nominal6/jam25 https://soundcloud.com/nominal6/jam22 https://soundcloud.com/nominal6/jam15 https://soundcloud.com/nominal6/jam6 Sorry, no progressive metal. Now, the way it went with this so far, is that now it might very well veer off with a critic of those songs. So I would like to say that despite the trend, I would - strangely - welcome critics on the songs and the way the sounds are put together. I always ask anyhow when I post them. Which will get totally off the subject here, but it's OK. Some are possibly itching to talk about actual mixing.

let me offer you a hint. what the plugins need to share are not messages but computer (analysis) data. normally (though not universally), when entities run inside a single process and need to share information, they do so by sharing access to memory. On Tue, Oct 18, 2016 at 4:58 PM, jonetsu <jonetsu(a)teksavvy.com> wrote:

On Tue, Oct 18, 2016 at 5:50 PM, jonetsu <jonetsu(a)teksavvy.com> wrote: I'm not saying nothing. I'm trying to tell you that if you want a set of plugins that behave as an integrated whole, sharing data about the tracks they are processing and potentially using data from other tracks to adjust their own behaviour, then you need them to share *memory*, not exchange messages. There's effectively no chance that different plugin manufacturers will ever agree to a single standard for such a thing, so there's unlikely to be any "protocol" or "specification" for this. It is something that a single plugin company could do on their own, to notable effect.

Sorry, the plugin (and DAW and audio technology worlds) simply haven't worked that way in a long time, if ever. We never even managed to agree on a way to share a tempo map in JACK ... you think it is somehow easier to agree on how to share "this is what i think the current spectral energy of my track looks like?" On Tue, Oct 18, 2016 at 5:58 PM, Jeremy Carter <jeremy(a)jeremycarter.ca> wrote:

The argument against this seems to be basically, it can't happen because people have differing financial interests in having their own proprietary solution which they want to keep private. But this is not the proprietary world here, this is the world of open source. So what's the problem with doing it anyway? Maybe it takes off and maybe even gets used eventually in proprietary stuff (if that's what you want) On Tue, Oct 18, 2016 at 6:16 PM, jonetsu <jonetsu(a)teksavvy.com> wrote:

On Tue, 18 Oct 2016 17:55:13 -0400 Paul Davis <paul(a)linuxaudiosystems.com> wrote: I see it as protocol. Whether it ends up using ZeroQ, network connections, laser beams, whatever, does not matter inside the context of this discussion. Hence, the size and type of bolts when the vehicle is not even talked about. This is what I wrote earlier today. Until there is an open source project that deems making a parallel and/or until there is willingness to open the protocol. After all, many protocols being used daily started like this. But other do not, they start immediately in the open. For instance the upcoming protocol to configure devices for in-field validation of updates to US Federal FIPS modules. There are many factors that will set the context whereas a protocol will be made privately first or not. In the audio world, specifically for the plugins, it might very well be private to start with. That people in the open source world would come with such a proposal first hand... naaaah.

On Fri, 14 Oct 2016 07:11:02 -0700 (PDT) Len Ovens <len(a)ovenwerks.net> wrote: Shhh! Don't give them ideas :( -- Will J Godfrey http://www.musically.me.uk Say you have a poem and I have a tune. Exchange them and we can both have a poem, a tune, and a song.

On Fri, 14 Oct 2016 07:52:41 -0700 (PDT) Len Ovens <len(a)ovenwerks.net> wrote: Again, it is at the mixing engineer level. Not at the stage of personalizing the characteristics of individual sounds. The mixing stage sees a set of tracks that hopefully embodies what the client wants to do, wants to express. There might be some characteristic tweaks here and there, re-amping, widening with chorus, early reflections, etc, but the core of 'I want to sound like David Glimour' or such is already being done. NI has a guitar processing plugin based on Rammstein's guitarist sound. That would be more in the vein maybe. There might be others out there like that. OTOH, I don't see what would be wrong with having a set of presets that supposedly replicates a certain guitarist setup. That can provide a starting point to explore further. I haven't checked for a (very long) while but there was always a constant artist/gear relationship in the press and in products. Guitars made according to Brian May, what kind of pedals used for achieving this tone, etc...

On Fri, 14 Oct 2016 19:27:07 +0200, Joël Krähemann wrote: Hi Joël, this is in reply to what of my statements? The question is to what degree and for what purpose plugins should be able to share data. A given example mentions an audio track, an audio analysis recognises that it is an vocal track and not a guitar track, in case the audio engineer should be unable to distinguish a vocal track from a guitar track and after the analysis is done, useful tools are automatically added to the effect chain, with useful settings, I guess the analysis tool will send data for useful settings to the plugins, that is what the network is required for. For a vocal track e.g. the very useful exciter gets added with some useful settings that fit to the track and an EQ with some useful settings and two compressors with useful settings, too. This is absurd! Let alone that nobody often would use an exciter, useful settings are related to so many things, that an analysis of a track does lead to nothing useful at all. Regards, Ralf

On Fri, 14 Oct 2016 20:39:04 +0200 Joël Krähemann <weedlight(a)gmail.com> wrote: Two very solid arguments. Really ? Maybe if anyone goes forward with having plugins that can do what was sketched here, it should be brought up to the Human Rights. Just a thought, going along with the solidity being expressed.