# I always thought imbre held a multitude of sins, is it even anything than a waste basket category... ## Skills Used - deep-writing-mode - epistemic-discipline - file-management ## Retrieval Notes - Session id: `547fa450-aca5-4c76-a2cd-1ea453027611` - Last activity: `2026-07-03T23:20:54.076Z` - Files touched: `2` ## Artifacts **Created:** - [[Writing/research/auditory images/Against Hearing Sounds (short paper draft)]] **Modified:** - [[Sessions/Auditory Images]] ## My Notes <!-- Add your notes here. This section is preserved across syncs. --> ## Conversation ### User I always thought imbre held a multitude of sins, is it even anything than a waste basket category, is there such a propery or can't we articulate source proiperties we hear. don't let me get away with anything, but stress test this analytci philosophiucal claim. i want to therow a wild card into the mix, here is an app i just downloaded. please assume i understand exactly what this guy ist alking about, no need to dumbdown any of your ideas or thinking. i know how to talk aboutaudition and acoustics, harmoicsa etc. sources... ![] basically, i never thought sounds as properties or invidiauals exist/ are heard, the claims about what is represented make no sense you can look at plent y of the drafts in the vault or drive i think about not hearing sounds a loit although it was years ago. can't we just think of audiory phenomnology in terms of hearing sources and their properties in the same way that we see objects (xosometimes moving in space yes i know it is different for audition, but look up my diasertation, i think we hear sources, not sounds (neither peroperites nor waves nor indiviuals) recordings are auditory images, assume i vcan meet your objections and you ewill dfind that i havr in the many drafts you can find in vsulty or drive. What do you think f all this, it aklwats msade me sad i could never nurn my phd main claim in to a publoishable paper. do you think we could do that note that i also have the who papers about auditoruy images and also all that argument i am very proud of about the lack of certain stuifff in audtiory perception, this is a big one, if you can't dinfa aout what i meant then you wll never understand , it involves fgarkas, chalmers, you need to look in drrive if you have no luck in the vauklt. (https://www.youtube.com/watch?v=-LvBBycM91o) ## Transcript ### Intro and sound examples **0:00** · Hi, meet coastline. ### What is Coastline? **0:43** · So what is coastline? Coastline is a complimentary texture generator. But what does that mean? It means that coastline generates sound that is related to its input. It might be tonal. **0:55** · It might be noisy. But all of that sound originates from the input source and cosine generates something to complement it. It can introduce new sound alongside your input similar to a reverb or a delay or it can affect the input sound as well similar to a filter or a distortion. The main mechanism that Coastline uses here are called sustainers. What is a sustainer? ### What is a sustainer? **1:22** · Basically, it's a reynthesized version of a recording which can run infinitely. **1:29** · It's basically a uniform version of the input that maintains the familiar characteristics and qualities of the input while being stretched forever. A sustainer is kind of analogous to a loop in a looper. It's a short snapshot of incoming audio. But a sustainer takes this one step further and reynthesizes that loop. So how do we create a sustainer? We do this by recording audio which automatically becomes a sustainer. ### Recording into Coastline **2:00** · The primary way for doing this is with the threshold control. When this volume threshold is met, a recording starts and a sustainer is made. Here's my input signal. **2:16** · You can see the volume of it is shown on the threshold knob. If we turn the threshold down, you'll see that once the volume meets the threshold, a recording is made. And now we have one sustainer running as shown by this mountain that appeared. **2:38** · You can also click this recording button to make a recording at any time. And if you want to disable the threshold in order to keep the sustainers that you currently have, you can turn it off right here. In the sustainers menu, we find some options for managing them. You can change how many there are total. ### Managing Sustainers **3:00** · You can change how long of a recording they're based on. And you can change how long they take to fade in and fade out. **3:12** · If you want to have a different fade in and fade out time, you can just enable this second knob down here. In addition, we have the hold time control, which controls how long a sustainer will run before automatically fading out. By default, this is infinite. So, if you make a recording, it'll run forever. **3:33** · But if you want to create something more momentary, you can turn this down and the sustainer will only run for that amount of time. **3:54** · Sustainers can also be played and paused. If we record a sustainer, Right now it's playing. If we want to pause playback, we can just hit pause here. To resume playback, hit play. **4:15** · The speed and the difference between the two states is defined by the fade in and the fade out time. **4:26** · And again, if you actually want to clear the sustainers, you use the X. If all of the sustainers are currently in use and a new one is recorded, the oldest one will get faded out and replaced. **4:42** · If you want to clear all of the sustainers immediately, you can just press this X button up here. **4:50** · And you'll see we also have a little mixer down here. where you can balance the volumes and the pan position of all of the sustainers. **5:01** · And you can also lock them or solo them. **5:05** · Locking is really helpful if there's a recording that you don't want to be replaced. If a recording is locked, that means that a new recording will never replace it and it will never fade out due to the hold time. This can be really nice if you want to create one drone that kind of runs underneath everything. **5:31** · And you can also use the solo option to figure out which of those sustainers is which. Okay, so we know what a sustainer is now, but how does it actually reynthesize the audio? There are two channels that accomplish this. They have different roles and they produce different results, but they work together very well. They're called Tide and Glade. Let's start with Tide in this menu. ### Tide (Ebb mode) **6:00** · Tide analyzes the strongest frequencies in the signal and creates an additive synthesizer based on those frequencies. **6:10** · By default, it's going to reynthesize using sine waves for very pure recomposition of the sound. Let's hear what that sounds like soloed. **6:29** · So you can hear that tide has taken the input signal and it's assigned a bunch of sine waves to match the primary frequencies in it. We have three main controls here. Partials is going to change how many oscillators are played in recomposing the signal. These oscillators are ordered by significance or volume. **6:55** · So, the first one is going to be the loudest signal that was detected in the original recording. And as you add more, we're going to bring in the fainter signals that were in that recording. **7:16** · Wave shaping changes the shape of the oscillator. Like I said, by default, they're sine waves at 0%. **7:26** · If we turn this down, we're going to start morphing towards a saw wave. **7:34** · And you can see kind of a mini oscilloscope on our volume knob here. **7:42** · If we turn this up, we're going to wavefold those sine waves. **7:53** · I think these are two really nice options because with the saw wave we can kind of add new harmonics into the signal and with the wavefold you can start to kind of imply a different octave of the output. **8:15** · And finally, we also have this chorus setting, which is a stereo chorus that can add a little bit of lushness to the sound. **8:27** · By default, there's a little bit of that. And I should note that over here on the main panel, there is a pitch control for Tide as well. **8:41** · In addition, Tide has two modes. The first one is called EB. This one attempts to reynthesize the input by ordering the partials by their significance and replaying them. This is kind of a pure mode that's the most transparent and representative of the input signal. **9:02** · There are two main controls here. Offset is going to change kind of what the loudest partial is. It's kind of going to take the set of partials that we're hearing and it's going to shift it such that we're removing the most significant partials. So right now we're hearing 15 partials. You can think of that as hearing 1 through 15. **9:28** · If we change the offset we might hear 2 through 16 or 3 through 17. So, we're kind of shifting what part of the set we're hearing. **9:52** · This is a really fun effect because it kind of lets you bring out the weirder parts of the signal. And I think it's also really fun once you kind of remove the most significant frequencies. It really changes the context of what you hear. **10:15** · So this can be a really fun control to play around with. In addition, we have this harmonics control. This is basically going to change the proportion of the frequencies that are identified to be harmonic or inharmonic. **10:31** · At higher values, we're going to hear those harmonic frequencies which are going to be more pleasant and kind of more consonant with each other. **10:46** · If we reduce this, we start to bring out the inharmonic frequencies, which are going to be the ones that kind of stand out as unique. I think it's really cool because it brings out these frequencies that we might not have really noticed in the input signal. **11:19** · I think with these two controls combined, they're kind of where you can go to change how consonant or how weird the output of coastline is. I think it's really fun to just kind of dial these in and think about exactly how strange you want coastline to sound. ### Tide (Flow mode) **11:51** · The second mode in tide is called flow. **11:56** · This one is going to be a bit more characterful and it's going to introduce more unique tones and harmonics compared to the original signal. In this mode, Tide is going to identify the notes played in a recording, and it's going to create a harmonic series of partials for each one. As a result, it's going to add new harmonics that might not have been present in the original signal. **12:32** · Partials controls how many harmonics are played for each fundamental. **12:42** · It's almost like opening or closing a filter. **12:55** · There are two main controls here called chroma and shift. Chroma is going to change the balance of the partials. Sort of like a formant filter. To the left, the output tends to be a little darker and to the right, it tends to be a little brighter. **13:25** · Shift changes how chroma behaves. **13:28** · At 100% chroma operates totally relative to the pitch of the fundamental frequency. **13:37** · Meaning at a certain setting for chroma the third harmonic is always boosted. **13:43** · The fifth harmonic might always be cut etc. **13:53** · When shift is at 0%. **13:56** · It operates on absolute frequencies. **13:59** · So that might mean that at a certain setting chroma is always boosting 500 htz and is always cutting 800 htz. **14:25** · together. These can be really fun just for kind of dialing in that proportion. **14:30** · And they're also fun just to move around and kind of create these constantly unique textures. ### Glade **14:37** · Now, let's take a look at Coastline's second engine that it uses in the sustainers. This is called Glade and it's a granular synthesis algorithm that sustains the input with a semi-random cloud of grains. The signal is analyzed to reduce any harsh transients or clicks in the signal and the grains randomly wander around the sample to explore different parts of the signal. **15:19** · This channel really maintains the tambber of the signal, which makes it a nice counterpart to tide, which kind of distills the signal down into its purest form. Between the two channels, you can decide how much you want to mirror the original signal or how much you want to create something that's a bit more abstract and new. **15:44** · The main controls for Glade are size, which changes the length of grains. **15:54** · You can make them really short and percussive or you can use longer grains to really get them blurring together a bit more. **16:06** · Spread changes the random stereo spread of the grains. **16:19** · Blur softens the volume contour of each grain for a more consistent volume without dips or spikes. At 0%, no volume compensation is applied. **16:43** · Churn changes how quickly grains wander around the recording. **16:50** · At smaller values, the grains will move very slowly and really emphasize different parts of the sample for long periods of time. At higher values, they'll move very quickly, and you'll start to kind of hear the entire recording at once. **17:14** · Finally, we have erode which reduces the sample rate of playback. **17:26** · This is a smooth control, so changing it or modulating it can produce nice little tapeike warbles. **17:42** · In addition, Glades pitch is over here on the main panel as well. **17:54** · I also want to note that both Tide and Glade will always run even if you mute them. **18:04** · This is so they're available for some of the audio effects that we'll talk about in a bit. But if you want to fully disable a channel because you're not using it, you just click this enabled button up here. This might be if you want to save CPU or you're just not using it. Next, let's talk about the filter. The filter is a multiode filter that all of the uh audio in Coastline can run through. It's got three modes, low pass, band pass, and high pass. ### Filter **18:33** · And it probably has controls that are pretty familiar. It's got a cut off frequency and resonance for adding a bit of resonance at the cutff frequency. **19:02** · In addition, this visual over here is interactive. **19:09** · Next, we have a feedback control. If you're familiar with my plug-in, Outgrowth, you might know this already, but feedback adds a bit of audio rate modulation at the cutff frequency. It basically feeds the input signal into the cutff frequency of the filter to add a little bit of drive and a little bit of hair. **19:43** · This is the first of many ways in coastline to add a little bit of distortion to the signal. **19:52** · And furthermore, we have this stereo spread option which changes the cutff of the filter in the left and right channel. **20:24** · I really like using the spread just to get a little bit of motion and movement in the signal. In addition, there are independent filter settings for each channel. By default, the filter only runs on tide and glade, but you can enable it on the input signal as well as the reverb if you want. Furthermore, we can change the offset of the filter frequency for each of the channels. **20:52** · For example, if you have one of the channels pitched up an octave, you might want the frequency of that channel to be reduced a little bit. **21:25** · I also like using the filter on the input because you can totally just use coastline as a nice little filter if you want. ### Audio Matrix **21:56** · Next, let's talk about the audio matrix. **21:59** · The audio matrix is where cosine becomes a completely modular and configurable effect. This is where we can start having different channels affect each other and talk to each other to create something really unique. The audio matrix allows one channel to amplitude modulate another or itself producing something similar to West Coast synthesis where we have two oscillators working together to create a really unique sound. **22:31** · This can be used very subtly such as here. Tide is modulating itself just a bit to create a little bit of grit on the signal. **22:46** · Or you can use it in the extreme to create a full-on distortion. **22:52** · Let's hear what it sounds like to turn up the amount on tide. **23:24** · We can also dial in the tambber of that modulation using these two controls that affect a wide band pass filter. **23:36** · We can change the center of that band pass filter and the width of it. **24:05** · Critically, everything that happens in the audio matrix happens prefader and before any volume adjustments. As a result, you can mute the output of a channel from the main output, but you can still use that channel as a source or destination in the audio matrix. **24:28** · One of my favorite things to do with Coastline is to kind of use Tide as a utility. Often times I will mute tide, but I'll use it as a modulation source in the audio matrix. Here I'll use tide to modulate glade while tide is muted. **25:05** · And as a proof of concept, we can go into tide settings and change them and hear the result. **25:28** · This can be really subtle or it can just be kind of total audio destruction. **25:37** · I think this is where Coastline extends itself beyond a traditional looper or atmosphere generator and lets you create something completely new with this as well. You can also use this concept to modulate the input signal and we can create really unique distortion. **26:13** · You can turn tide down to just one partial, maybe just a couple of sustainers. **26:37** · In addition, the audio matrix supports sample loading as well. There are also a number of builtin field samples provided as part of coastline. **26:50** · Let's listen to one of those modulate tide. **27:16** · Any sample loaded into coastline kind of gets turned into this infinite random loop so it can be used as a noise source. **27:28** · This can be a really nice way just to mix in a little bit of unique texture into your sound. ### Reverb **27:42** · And finally, let's take a look at Coastline's reverb. **27:47** · The reverb is applied to all of the channels equally by default, including the input. **27:58** · We can change the reverb decay and the damping to suppress some of the higher frequencies. **28:05** · In addition, we have this lowfi control which kind of downsamples the tail of the reverb and makes it grittier. **28:24** · We also have three controls for how much modulation is applied to the reverb tail. **28:33** · We have a chorus for changing how much modulation there is overall. **28:38** · We have a rate for changing a slow LFO that modulates the reverb. **28:47** · And we also have this chaos control which changes how much random fluctuations there are in the reverb tail. If you just want a traditional LFO chorus sound, you can turn that all the way down. **29:07** · But if you want to get more of that chorus sound without it being completely synchronized, you can turn this up. **29:20** · Finally, we have two options down here which change the balance of what's running through the reverb. By default, all three of the channels, input, tide, and glade, are going to run through the reverb with the same volume that you've set in the mixer. Or you can override this and treat it more like a send effect. And you can configure exactly how much of each channel is going to the reverb. **29:49** · This is also pre-fader, so you can mute the actual output and still send it to the reverb. **29:59** · This is also how you can remove signal from the reverb if you prefer. ### Modulation **30:07** · Now, let's talk about modulation in Coastline. We have a modulation menu off to the right where we can modulate pretty much anything in Coastline. If you want to make an assignment, you can click the add destination button under any of the modulation sources, or you can simply drag this button to whatever you want to modulate. **30:31** · When something is being modulated, you'll see this kind of ghostly second value on it indicating what it's modulated value is on top of the base value. When we have a modulation destination set up, we can change the intensity of that modulation here. And we can also change whether or not the effect just adds to the value or adds and subtracts for a bipolar response. **31:04** · Now, let's take a look at all of the different modulation sources we have available to us. ### Modulation - LFOs **31:10** · First up, we have four independent LFOs. **31:14** · Each LFO can have a different speed, which can be synced to the BPM of your DAW, and it has a different shape. We can see all of the shapes here. By default, it's a triangle, but we could use a sign. Ramp down, ramp up, a square, a stepped random source, and binary, which is randomly low or high. **31:44** · In addition, each shape has a morph option which changes the shape even further. For triangle, we can kind of go from a completely symmetrical triangle to one that spends more time ramping down or more time ramping up. **32:10** · For the sine wave, we can change whether the LFO tends to stay higher or lower. **32:20** · For the ramps, we can change if they have more of an exponential decay or more of a logarithmic decay. **32:30** · For the square, we can change the phase. **32:36** · And for stepped random and binary, we can change the slooh. And like I said, we have four independent LFOs as well. ### Modulation - Jitter **32:48** · Next, let's take a look at jitter. We have two jitter sources, which are kind of a smooth continuous random as opposed to the stepped random LFO, which kind of goes from value to value totally independently. **33:06** · jitter, as you can see, kind of smoothly wanders around up and down. We can change the speed of that to get something really slow and really gradual, or we can crank this up to get something a bit more frenetic, but it's always smooth and it always kind of stays continuous. ### Modulation - Sequencers **33:28** · Next, we have two sequencers. **33:31** · You can see that these are the stages in our sequencers and we can click them to turn them on and off or we can drag them to change the amplitude of that stage. **33:42** · To demonstrate the different kinds of sequencers, I'm going to assign it to this filter knob here. **33:51** · First, we have a stepped option which is basically just going to go from one value to the next. **34:00** · We can add some slow if we want that to be a bit more gradual. **34:06** · But we also have a rhythmic option which is basically going to kind of pulse. **34:14** · It's going to go up to the value of that stage and then it's going to fall back down to zero. **34:23** · You can have a configurable attack and decay to control kind of how plucky that is or how gradual it is. **34:33** · You can see that in this option bumps up to the full volume and then kind of falls back down to zero. So depending on whether you want to get this uh target kind of jumping up and down or if you just want it to kind of immediately go from value to value, you can choose from stepped or rhythmic sequencers. We have two of those and each one can have an individual speed and it can have an independent length as well. **35:04** · Next we have a performance slider. When this gets assigned, we get this new slider here in the UI. ### Modulation - Performance slider **35:16** · This basically gives you kind of one big control and it's especially designed with performance in mind. You might want to assign this to a number of things and then you can kind of use it as um kind of a way to create macro transitions or um just kind of change a bunch of things at the same time. For example, we can increase the volume of tide while decreasing the volume of the input. But you can really assign this to anything. **35:51** · I should also note that there are a couple of destinations available in the UI as a dedicated knob. So this performance slider can be really useful for those. For example, we have a global dry wet amount which is just going to kind of uh create a proportion between the dry input running into coastline and all of the affected output that comes out of it. **36:20** · So if you want to, for example, just kind of completely mute the output of coastline, you could assign the dry wet to the performance slider, then really dial it up and kind of use it performatively, and that'll kind of enable or disable the entire output of Coastline at once. Next, we have an envelope follower. This is going to create a modulation source from the input signal. ### Modulation - Envelope follower **36:54** · You can see as I play a note, we see the history on the envelope follower start moving around. **37:03** · We can trim this and that's basically going to change what volume we consider to be the max. **37:16** · In addition, we have an attack and release time. So we can kind of slo that value and make sure that it's not very jagged. ### Modulation - Per-Sustainer sources **37:29** · Next we have random per sustainer. This is basically just going to be a static random value each time a sustainer is created. So that can be nice for variations. You could have slightly different settings every time a recording is made. Um, and that can just be a nice way to kind of just create a variety of outputs really. Next, we have sustainer H. **37:56** · This is a modulation source that's basically going to ramp up as a sustainer runs. **38:08** · It basically is going to be zero at this minimum time and it's going to be the maximum output this maximum time. This might make the most sense just to show you. So we're going to have the sustainer H modulate the filter frequency. **38:32** · As the sustainer runs, it is going to accumulate this value. And then once it's been running for 10 seconds, it's going to stop accumulating. **38:58** · So, I think this is kind of nice because you can use it to create a sound that kind of grows over time or you can use it to create a sound that kind of decays over time. So, for example, you could map the age to the erode and you can reduce the erode over time. **39:38** · And finally, we have a sustainer pitch modulation source, which is basically just a value that's based on the pitch of uh kind of the central pitch that's identified in Tide. Um, this could be nice. Maybe you want some sort of key tracking, like a filter that kind of opens and moves around based on the pitch of the sustainer. Um, maybe you want there to be less reverb. **40:07** · Um, you know, if a pitch is lower or maybe in tide you want to have less chorus if a pitch is lower. Uh, so that's available to you as well. Additionally, I want to point out that you can modulate the amount of other modulation assignments. **40:31** · So, for example, I've created a modulation target here. We have the LFO changing the uh input volume. **40:45** · I'm going to come down to the performance slider and I'm going to assign that to change modulation depth. **40:53** · of LFO1 modulating the input. **40:58** · So right now the performance slider is at zero. So it is basically completely attenuating that depth. But as we turn it up, we start to reveal it. **41:16** · By default, we're going to multiply these two values together. So, we're basically attenuating or boosting the signal. But we could also add to it if we want to increase the intensity of that modulation depth. So, for example, rather than attenuating that, now we're going to boost the effects of it and make it hang out higher or lower. ### Additional settings **41:50** · Next, let's take a look at Coastline's settings menu. This is where we can find the more advanced settings that you might need a little less frequently than all of the rest of them. **42:02** · At the top, we have a simple threeband equalizer where we can change the proportion of lows, mids, and highs. We can set how much we're boosting or cutting each of them and their crossover points. In addition, we can turn bands on or off in order to create kind of a broad lowass or a broad band pass or even a scoop. Uh you can use this for a fun creative effect as well if you want. **42:33** · Next we have some settings for massaging the input and output of coastline. **42:39** · First we have an input highpass and lowass filter which you can use to cut frequencies from the input signal. **42:48** · Coastline is very sensitive to the input signal. So this can give you some basic ability to knock out any problem frequencies that you might be experiencing. **42:58** · For example, in my studio, sometimes I'll experience a hum or a whine with certain instruments. So, this can be nice to just knock out a signal, especially before it ends up in tide. **43:13** · Next, we have input and output volume adjustment. You can boost or cut the signal before it reaches coastline, and then you can boost or cut all of the output to coastline. Next, we have some per project settings that you might find helpful. **43:32** · The first is called instant playback, and it basically controls whether or not Coastline will start playing and sustaining a sound as soon as possible, or whether it's going to wait until a recording has fully completed in order to start playing. **43:54** · By default, with instant playback, we're going to start playing as soon as possible. **44:04** · What this means, however, is that the uh analysis for tide and for glade is going to improve over time as the recording is made. **44:18** · If we turn this off, we are going to fully complete a recording and then we are going to sustain that sound with a one-time analysis. **44:35** · So, the trade-off is whether you want to hear the sound as quickly as possible or if you want the analysis to be as stable as possible. Next, we can control whether or not audio is stored in the project. By default, if you save a project that has Coastline in it, it's going to save and reopen the sustainers as well. If you want to start fresh each session, kind of like a delay or reverb, then you can turn this off. **45:07** · And that means that you're not going to save any of the audio in Coastline. Next, we have some options for how the sustainers and modulators react to your DAW's transport, meaning whether your doll is playing or paused. **45:27** · By default, our sustainers are going to pause when your DAW pauses. This can be nice just to kind of make sure that there's no sound when your DAW is stopped. **45:42** · If you want the more extreme version of this, you can set your sustainers to actually stop, meaning that when your DAW is paused, the sustainers are going to clear and be deleted. **45:55** · If you want the sustainers to just completely ignore your DAW setting, you can set them to ignore and they will just keep running forever. If you're doing a lot of bounces and kind of renders and stuff like that, um, you might want to set it to stop just to ensure that, you know, if you render one thing, the sustainers are going to be stopped when you're done. And then if you were to make another render, you can start fresh. **46:26** · Uh, the modulators react similarly. This is going to control how the LFOs and the sequencers react to the transport. By default, they're going to reset when you start playback in your DAW just to kind of keep everything synchronized. **46:44** · You can also set them to stop and reset. **46:48** · So, they're going to stop running when your uh transport is stopped or they can just ignore that. Next, we have continuous recording. **46:58** · This basically changes how the threshold is going to behave. **47:04** · By default, your sound has to go over the threshold to create a recording, but then it has to come back down under the threshold to reenable another recording. **47:18** · This is so you know one held note or held chord won't continue to create more recordings. **47:29** · So typically I would recommend that you set your threshold kind of right at the peak volume of your sound so it eclipses that and then it comes back below and then recording will reenable once it comes below. But if you want coastline to function more like a delay or a reverb and just kind of run more continuously, you can turn continuous recording on. **47:53** · In this mode, Coastline will just continue to record whenever you're over the threshold. **48:04** · So, you don't have to come back down below the threshold to reenable recording. It's just if there's input that's over the threshold, we're going to record. Next, we have an option for just intonation. This is going to change what ratios Coastline uses when repitching tide and glade. Personally, I think just inonation sounds a little bit better. So, that's the default. Um, but if you want to go into equal temperament, you can just turn that off. **48:33** · And finally, we have pitch quantization. **48:39** · This is going to restrict the pitch of tide and glade to um certain musical intervals if you want. So by default they can be repitched to anything but we can enable pitch quantization for example to restrict them to just octaves or octaves and fifths. **49:22** · This can be nice for something like uh performance if you want to move that around but kind of have confidence that it's going to stay at uh a nice pitch. **49:33** · And finally, we have some global settings that get persisted across different sessions and different instances of coastline. The first is uh we have some motion settings. If you want to disable the background animation, you can do that. Or if you want to have a completely optimized graphical experience, uh you can also turn that on. **49:58** · That's going to remove all of the kind of superfluous animations and it's going to make all of the menus non-transparent just to help out your GPU a little bit. **50:11** · And finally, we have tool tips. If you want some helper text when you mouse over certain controls, you can turn on tool tips. But if you want to get rid of them, they're right here. And then at the bottom, we have links to Coastline's manual and this video manual. I want to take a brief moment to talk about the sidechain input for Coastline as well. **50:35** · Coastline accepts a side chain input in a couple places, which can be nice to just kind of expand the functionality a little bit. In the sustainers menu, we can set whether or not the main input or the side chain input is used for recording. You should be aware that the actual input channel that runs through Coastline is always the main input to the plugin. **51:04** · However, you can set the side chain input to be what's recorded via the threshold. So the side chain is never actually heard in the output, but it can be what is recorded and persisted with sustainers. So you can get a fun kind of interplay between some dry signal and a kind of unrelated side chain input if you want. In addition, you can use the side chain as a source in the audio matrix. **51:36** · This could be interesting because you can use a totally different sound, totally unrelated sound as a modulation source. **51:48** · You could have the side chain modulate the input or you could have the side chain modulate the tide or glade or whatever you want to hear. Um, this can be nice just to get a little bit of interaction and to get these two unrelated sources kind of playing off of one another. And finally, you can use the side chain for the envelope follower in the modulation matrix as well. **52:09** · That can be fun if you would like to, for example, use the volume of some different channel kind of as a a modulation source that changes something within Coastline. Now, let's take a look at the presets menu, which can be found up top here. Coastline ships with a bunch of factory presets, and we have some tags for the different uh types of presets here. We can sort by atmosphere, texture, distortion, or reverb. ### Presets **52:43** · And furthermore, we have a favorites option as well. If you want to set up a favorites option, so if you want to load one of these presets, you can just load right there. And furthermore, those settings will be applied without getting rid of the current sustainers that are running. **53:20** · So you can just kind of flip through these sustainers and hear different options. We have these buttons up top as well just for moving through them quickly if you want. **53:29** · There's an init patch which will just restore all of the settings back to the default. **53:35** · If you want to save a preset, you can go to any folder besides factory and then save a preset. You can give it a name, give it a category, and then when you save it, it'll be in that folder. If you want to create a new folder, just hit add folder, and then you can create that. **53:57** · Furthermore, at the bottom of your presets list, you can either open them in your file browser or you can export them as a zip. **54:10** · You might want to send these to a friend or if you uh create presets and sell them, you can export as a zip right here and another user will be able to install them with this button here. So you just export as a zip and then as another user you just click install presets and you can install that zip and that will give you that folder right here in the list. ### Sound example - Instant atmosphere **54:37** · All right, let's dive into some audio examples and hear how coastline really sounds. **54:43** · I want to start by using Coastline as an atmosphere generator which creates a kind of pleasant synth and kind of surrounding ambience for any sound that you send it. **55:05** · To me, this is kind of the default mode of Coastline. It sustains sounds infinitely, and you can just kind of float here for as long as you want. **55:18** · There's a couple of sound design tricks that I want to point out here when you're using a patch like this. **55:24** · The first is stereo spread. You can see here that we already have an LFO assigned to it. If you position this kind of around the top of your content, you get this nice motion and get that stereo image kind of swimming around. **55:42** · You can make this a bit more pronounced with resonance and feedback. **55:50** · Another thing I like to play around with is the different pitches for the two channels. And to explore this, I'm just going to turn on pitch quantization. **56:01** · You'll see right now tide is at the original pitch while glade is an octave up. This can be nice for just a little bit of a shimmer effect. Maybe you want them at the same pitch, though. **56:16** · Or maybe for a little bit of complexity, you want one of them a fifth away. **56:25** · I also like to put tide down an octave a lot of the time for just a really rich and deeper tone. **56:38** · And the last thing I want to recommend is you can mute the input of Coastline and really just kind of hear this morphing atmosphere. ### Sound example - Call and response **57:30** · The next sound design thing with Coastline that I want to point out is whether or not you're using an infinite hold time or a temporary hold time. You can kind of change the behavior of coastline from being this infinite wash of sound or something momentary that kind of chimes in as a response to anything you play. **58:14** · And one thing you might want to think about here is whether you want a sound to kind of grow or whether you want it to kind of start strong and fade away. **58:24** · One of the things I like using for this is the sustainer age modulator. **58:30** · For example, we could turn our filter cut off down and use the sustainer age to open it up and kind of make our sound grow over time. **59:04** · Or we could do the opposite. We could kind of start with the filter open and clamp it down throughout the lifetime of a note. **59:25** · So, it kind of depends on what effect you're looking for, whether you want to get these echoes that kind of die out and fade off into the distance, or if you want something that grows and really swells and responds to what you're playing. This next example uses coastline as a kind of harmonic distortion. ### Sound example - Harmonic intermodulation **59:46** · With the audio matrix, I'm routing tide to modulate the input signal. But you'll notice down here that tide is muted. So we're not actually going to hear the output of tide except for its effect on the input signal. The result of this is going to be a kind of harmonic distortion. **1:00:09** · You can see that tide settings are very simple. Just one partial, a little bit of chorus, and a basic wave shape. **1:00:19** · To start, I'm going to make one recording. **1:00:25** · And now I'm going to turn the threshold off. And as I play more, we're going to hear this one partial in tide start to modulate everything else. **1:00:54** · Now we can reenable the threshold and get this harmonic distortion that kind of keeps up with what we're playing and follows along. **1:01:20** · You can also add more sustainers in if you want this effect to be a little bit more complex. **1:01:38** · You can also play around with the wave shape and tide to add even more complexity in harmonics. **1:01:57** · I also like to add chorus which kind of creates a stereo image to the modulator. ### Sound example - Flavors of distortion **1:02:27** · The last thing I want to show are all the different ways we can create distortion in coastline. **1:02:35** · This is kind of a really interesting effect to stack on top of the atmospheric washes that cosign likes to create. If you're familiar with genres like post rock or shoe gaze, a lot of times you'll stack a reverb with a distortion and create a really interesting complex result. This patch out of rope uses a very low erode setting on Glade. **1:03:03** · So, this is one of the ways we can add some grit to our sound. **1:03:16** · And not only that, but we have this jitter affecting erode. If we were to speed this up, we'd even get some kind of warbles. **1:03:37** · So, this is one way we can introduce some grit to our sound. **1:03:43** · Another way that's a favorite of mine is the filter feedback. **1:03:49** · This adds a bit of drive that is specifically centered around the cutff frequency. **1:03:59** · So you can kind of dial in exactly where you want to hear that drive. **1:04:08** · It can be brighter or it can be really murky. **1:04:16** · You can compound this effect with resonance to really pronounce it. **1:04:29** · You can also enable the filter on the reverb, which is going to create even more drive because we're already filtering our sound. And then we're running it into the reverb to be filtered again. So, I'm going to start with no feedback, no resonance to show this. **1:05:11** · And finally, we can use this lowfi setting in the reverb to make that tail even grittier. **1:05:32** · You can push this to total destruction if you want. **1:05:52** · Okay, that's it from me. This has been a demo of Coastline, and I hope that you enjoy it. I hope that it becomes a new musical partner for you and helps both inspire great new compositions as well as complement the ones that you already have. Take care. ### Assistant I'll load the epistemic-discipline skill and your research context first, then hunt down your dissertation drafts and the Farkas/Chalmers material before engaging the timbre claim. ### Assistant I've read everything: the dissertation, both current drafts, the 2022 Drive doc where the Farkas/Chalmers argument lives, your published paper, and the timbre PDFs you've already filed in Learning. Here's the full picture, then the stress test you asked for. ## What I found and read Your dissertation is [Not Hearing Sounds](Library/CloudStorage/[email protected]/My Drive/Sync/Learning/Dissertation Final Real.pdf) (Antwerp, 2016, Nanay supervising). The abstract states the claim exactly as you remember it: > I argue that although we hear material objects, their activities, and spaces, we do not hear sounds. ... This framework is used to identify three general approaches to auditory perception: on a sonicist account we represent only sounds, on a mediate account we represent sounds and the activities of material objects, such as collisions, scrapings, and rollings, and on an immediate account we represent only these activities. The two live drafts are [Against Hearing Sounds - Draft.md](My Obsidian Vault/Notes/Against Hearing Sounds - Draft.md) (~75% complete) and the Audible Presence / auditory images text at [whole undifferentiated text.md](My Obsidian Vault/Writing/research/auditory images/whole undifferentiated text.md), with the project mapped in [Sessions/Auditory Images.md](My Obsidian Vault/Sessions/Auditory Images.md). And the argument you're proud of — the one you said I'd never understand you without — is in a Drive doc called "Against Hearing Sounds February 2022", internally titled "Sounds, Eliminated Entirely". It's the only document anywhere in your Drive containing both Farkas and Chalmers, and it runs like this. First, Budek & Farkas's phenomenal presence criterion sets the bar for what counts as an object of hearing: > the reason why we say that neither the retina, nor the neighbor is the object of the perception—that we don't *see* the retina or the neighbor—is that they are not phenomenally present in the experience. (Budek & Farkas 2014, p. 360, as quoted in your doc) Sources pass this test; waves, medium disturbances, and vibratory dispositions plainly fail it. The defender of sounds retreats to "but sounds are *phenomenally* obvious even if metaphysically mysterious, like colours" (you quote Kulvicki making exactly this parallel). Then comes the move: Chalmers's Edenic description of colour phenomenology commands near-universal assent — > Phenomenologically, it seems to us as if visual experience presents simple intrinsic qualities of objects in the world, spread out over the surface of the object (Chalmers 2010, p. 398, as quoted in your doc) — whereas there is no analogous consensus about how sounds seem. Nudds says sounds "appear to be individuals that instantiate properties... independent of their sources"; Leddington says they are "heard as bound to, or fused with, their sources". Your inference: > Why think we seem to hear sounds, when there is so little agreement on how sounds seem? And then Section 4 gives the error theory: modelled on Miller et al. on temporal passage, "sound friendly language" (driven by the transitoriness of auditory sources and the fact that audition presents objects only *doing* things) generates the belief that we hear sounds, via either misdescription or inferentialism. That's the lack-argument: not just that sounds aren't heard, but that there is no *seeming*-to-hear-sounds at all — the phenomenal datum the whole literature starts from is a cognitive artifact. One housekeeping note before the stress test. Your published "Sounds as Properties" (Thought 2021) positively proposes "a new account on which sounds are temporally extended properties of objects." Given what you've said today, I read that paper as an intra-orthodoxy move — if you're going to have sounds, put them on objects — while the dissertation and the 2022 doc carry your actual view. A referee for any new eliminativist paper will notice the author of that paper now saying sounds don't exist, so the strongest fallback position you'll have to defeat is your own 2021 self. Worth deciding early whether the new paper treats Young 2021 as the last, best property view that still fails. ## Stress test one: the timbre claim needs disambiguating, and the versions come apart You asked whether timbre is "even anything more than a wastebasket category, is there such a property or can't we articulate source properties we hear." That's three claims wearing one coat, and they have different evidential situations. The wastebasket premise itself is solid and citable — you already own the sources. Bregman, quoted in Siedenburg & McAdams (the paper is in your Learning folder): > "This is, of course, no definition at all. [...] The problem with timbre is that it is the name for an ill-defined wastebasket category. [...] I think the definition [...] should be this: 'We do not know how to define timbre, but it is not loudness and it is not pitch.' [...] What we need is a better vocabulary concerning timbre." (Bregman, pp. 92–93, quoted in Siedenburg & McAdams 2017) But here's the first thing I won't let you get away with: *negatively defined* does not entail *nonexistent*. The ANSI-style definition is a fact about our taxonomy, not about the world. "That which distinguishes two sounds of equal pitch and loudness" could name (a) nothing, (b) one multidimensional property, or (c) many perfectly real properties lumped under one lazy word. The psychoacoustics actually pressures option (a): multidimensional scaling of timbre similarity judgments recovers *stable low-dimensional structure* — dimensions tracking spectral centroid, attack time, spectral flux (McAdams's work; the timbre-spaces history by Reuter & Siddiq; both in your folder). Pure wastebaskets don't produce orderly similarity spaces. If subjects' "how similar do these sound?" judgments were tracking nothing, MDS would return noise. It doesn't. So eliminativism about timbre (your option "is there such a property → no") has to explain the orderliness away, and "we hear source properties instead" is exactly the right shape of explanation — spectral centroid orderliness is orderliness in heard material and excitation. Siedenburg & McAdams themselves end up at pluralism rather than elimination: "the wastebasket turns out to be a colorful umbrella(-term) upside down", against Isaac's timbre physicalism ("In direct opposition to physical realists such as Isaac (2017), we propose to locate timbre on the perceptual side of the 'psychophysical divide'"). Note that Isaac's paper is also in your folder — that's the opponent who takes the wastebasket and tries to make it a respectable physical property. Any paper you write here has to beat him, not just Bregman's throwaway line. Second thing: your own dissertation already helps itself to timbre-talk in ways your official view has to pay for. The transistor radio aficionado passage: > Imagine a transistor radio aficionado, who has listened to so many of these things that he can tell just from the particular character of added tinniness whether he is listening to a Sony TR–55 model with a 1 inch speaker or a Sony TR–63 with a 1.2 inch speaker. One way we could describe this case is by saying that he is making an inference about the size of the speaker from the type of timbre it is adding to playback. (Not Hearing Sounds, ~p. 72) Tinniness *added to playback* — added to what, exactly, if there are no sounds to be the bearers of the added quality? On your view the answer has to be: the aficionado hears the speaker (a material source) and hears its size, directly, the way your Against Hearing Sounds draft says we hear the door's woodenness. Fine. But then "timbre" in that passage was always eliminable in favour of a heard source property, and the dissertation should have said so. The fact that even you, mid-dissertation, reached for the sound-friendly idiom is either evidence for your Section 4 error theory (the language is that sticky) or evidence that something resists the paraphrase. You need to decide which, case by case, because the paraphrase does *not* obviously go through everywhere — which brings me to the plugin. ## Stress test two: Coastline is a machine for manufacturing your hard cases You wanted the wild card played properly, so here it is. Four cases, in increasing order of how much they hurt. The sustainer versus the attack. A sustainer is "a resynthesized version of a recording which can run infinitely... maintains the familiar characteristics and qualities of the input while being stretched forever." Which familiar qualities? Not the attack transient — there is no onset in an infinite drone — and log attack time is one of the most robust dimensions in every timbre space since Grey (your folder has "Isolating the dynamic attributes of musical timbre" making exactly this point). So the sustainer preserves "the timbre" in the spectral-envelope sense while annihilating it in the temporal-envelope sense, and listeners still say it "sounds like" the input. Notice what the plugin's own architecture concedes: two engines, because one word covers two different things. Tide "distills the signal down into its purest form" (additive resynthesis of the strongest partials); Glade, the granular engine, "really maintains the timbre of the signal" per the designer. A single folk category, implemented as two orthogonal processes you can mix independently. That's not a philosophical thought experiment — it's a shipping product whose interface is a decomposition of the wastebasket. If "timbre" named one property, Tide and Glade would be redundant. They aren't. This is grist for the articulation reading of your claim (many real, separately manipulable properties) and against the flat elimination reading. The harmonics knob is aimed at your own foundations. Your draft's case for source-hearing leans on auditory scene analysis: "Objects tend to vibrate at multiple, harmonically related frequencies simultaneously, and so the auditory system will tend to group together frequency components which are related in this way." Ebb mode's harmonics control "change[s] the proportion of the frequencies that are identified to be harmonic or inharmonic" — continuously. It is a sourcehood dial. So: sweep it to 50% and report what you hear. One source? Two? A smeared, unstable something? If auditory experience is the presentation of material objects, the view owes an account of intermediate cases — vision has fog and camouflage as models of degraded object-presentation, and maybe that transfers, but you should have to say so explicitly, because the phenomenology of half-harmonicized partial sets is arguably not like mist over one thing; it's like objecthood itself flickering. The offset control makes the same point at the level of parts: shift from partials 1–15 to 3–17 and, in the designer's words, "it really changes the context of what you hear." Occluding a third of a seen object does not change *which object* you seem to see. Deleting the three most significant partials of a heard source frequently does. If sources were heard the way objects are seen, partial-set membership shouldn't behave like identity. Why does it? Sidebands and the audio matrix. When Tide amplitude-modulates the input pre-fader while muted, the output contains sum and difference frequencies present in neither signal — the West Coast trick the designer name-checks. There is no event in the room whose vibrational structure those sidebands report. On your view what's heard is the speaker cone doing exactly what it's doing, directly, which is actually a clean answer — but notice the cost: it makes the "familiar characteristics of the input" that survive into the output perceptually *irrelevant* to what is heard (you hear the speaker; the guitar the sustainer came from is nowhere in the perceptual story unless the image account kicks in). So everything now rides on when the image account kicks in. And it doesn't kick in here — this is the case that hurts. Your recordings argument in both drafts is built on twofoldness: hearing the absent event *in* the present speaker, with the speaker as the auditory analog of the picture surface. A twenty-minute Tide drone — offset 5, wavefolded, chroma jittering, just intonation, pitch-quantized a fifth up — presents no absent event. Not because the event is unidentifiable (your fox-in-the-gloom move covers that) but because there is no event of that kind for it to be an appearance of: an excitation with no onset, sustained forever, in a tuning no physical object was in. A sustainer isn't a recording; it's a generated auditory image at best, and at the abstract end it's arguably not an image at all. So Coastline exposes a continuum your account currently treats as a binary: recognizable sustained strum (twofold image), degraded resynthesis (image giving out), abstract drone (plain single-fold hearing of a speaker doing something exotic). I think this is a gift rather than a wound — pictures run the same continuum from photograph through Richter blur to full abstraction, and an account that *predicts where twofoldness gives out* (my speculation: where source-recognizability of the first fold fails) is stronger than one that doesn't — but as the drafts stand, nothing in them decides what a Coastline drone is. And note your Sessions file lists the missing third objection for Section 3 as "music/pure tones... tinnitus?" — resynthesis is that objection, in its strongest modern form. One Tide partial through headphones is a manufactured tinnitus case. Headphones make it worse, via Farkas — and this is the deepest problem I found. The Farkas paper you own but did *not* use in the 2022 doc, "Constructing a World for the Senses" (2013), says: > when these sensory features are received by the subject in a highly organised and predictable structure, one that responds to actions and further inquiry in a systematic way, the experience may become suggestive of the presence of something beyond this experience, namely, an experience-independent object. Your own Audible Presence draft establishes that headphone listening fails exactly this condition: "The principle reason for this [in-head localization] is the invariability of stimulation: sound waves do not vary as you move your head or walk around." Put Farkas 2013 and your headphone section together and you get an argument *against* you: audition-through-headphones fails Farkas's structural conditions for presenting experience-independent objects, so in the drone-in-headphones case nothing material is presented — not the absent event (there isn't one), and not the driver either (it isn't presented as located in the world; the experience sits between your ears). Phenomenally present *something*, materially present *nothing*. That is precisely the pure audibilia Soteriou wants, generated by the theory's own machinery. Budek & Farkas 2014 giveth (waves aren't phenomenally present) and Farkas 2013 taketh away (in-head experiences aren't object-presenting). A paper that uses the first Farkas needs a section defusing the second. I don't think it's undefusable — you can bite the bullet that headphone listening is a degenerate, non-world-presenting mode of audition, a kind of perceptual imagining-adjacent state, which is independently interesting and consistent with your aspatialism sympathies — but it must be faced, not skirted. ## Stress test three: the Farkas/Chalmers argument itself Three pressure points on the argument you're proud of, since you told me not to let you get away with anything. The colour-consensus premise is doing all the work, and it's softer than the doc admits. Chalmers's Edenic description is contested — relationalists deny colours look intrinsic; and colour phenomenology itself has modes where colour detaches from surfaces (Katz's surface colour versus film colour versus volume colour: the blue of the sky is nobody's surface quality). If expert disagreement about how sounds seem shows there's no sound-seeming, parallel disagreement about whether colours look intrinsic or relational threatens the contrast case. The reply I'd build: the colour disagreement is about the *metaphysical gloss* on an agreed seeming (everyone agrees colours seem qualitative and out there on things); the sound disagreement is about the *grammatical category* of the seeming itself (property versus particular) — and category-level disagreement is the kind that indicates no determinate seeming. That's defensible, but it needs arguing, because as it stands a referee will just say "philosophers disagree about everything; that shows nothing." Second: there's a pluralist escape route the doc doesn't close. Maybe Nudds and Leddington disagree because they're introspecting different cases — Nudds's individuals-intuitions come from music and streaming, Leddington's boundness from impact sounds. Sounds might seem particular-ish in sustained/musical cases and property-ish in impact cases. That would still wreck the orthodoxy (no unified phenomenon answering to "sound") without vindicating full elimination. You should either co-opt this (pluralism as a stepping stone: once the seeming fragments by case, the error theory explains why one word covers the fragments) or refute it. And notice — this is the same dialectical position as timbre. I interpret your two suspicions as one thought at two grammatical categories: "sound" is the wastebasket noun, "timbre" is the wastebasket predicate, and both dissolve into heard sources, their properties, and perspectival/medium contributions, with language papering over the join. If that interpretation is right, the timbre material isn't a digression from the PhD claim — it's the property-side half of the same argument, and the strongest version of the paper runs them together: the best remaining candidates for sound-properties (pitch, loudness, timbre) either redistribute onto sources, forces, and perspectives, or turn out to be no one thing at all. Third: the error theory needs its empirical legs. Miller, Latham and colleagues didn't just posit passage-friendly language — they ran the folk-phenomenology studies. Your sound-friendly-language story generates testable predictions (do naive subjects, asked to describe what they hear without sound-vocabulary available, reach for source descriptions or quality descriptions? does sounds-as-particulars talk vary cross-linguistically with count/mass grammar?). A paper can flag these as predictions rather than run the studies, but the doc currently ends exactly where a referee will ask "and why believe the language story rather than the phenomenology story?" — the transitory-sources and objects-only-doing-things observations you give are good, and they're also, notice, an articulation claim: audition presents *happenings*, language reifies them. Your dissertation's things-and-forces parsing (with the Handel source–filter quote — "the acoustic properties underlying timbre arise from the physical processes (e.g. banging, blowing, scraping) that excite the source and the physical material and shape... of the objects that create the filter") is the positive account that plugs the Section 2.1 gap the Sessions file flags: hearing directly = auditory parsing into vibrating things and stimulating forces, with timbre-talk as the folk's compressed report of filter-properties and excitation-properties. Handel's model is literally your articulation thesis stated as acoustics. Where the articulation claim genuinely runs out — and you should concede this in print rather than get caught — is acousmatic and synthetic listening. Sweep Coastline's chroma and what changed? No material, no force, no room. "Brighter" is not source vocabulary; it's spectral-centroid vocabulary wearing a metaphor. Your options, as I see them: treat brightness-talk as perception of perspectival properties (your Kulvicki move — but perspectival properties are relational, and nothing relational changed; the generated spectrum itself changed); treat it as hearing the speaker's activity under a fineness of grain our source-vocabulary can't match (articulable in principle, not in practice — the aficionado again, scaled up); or admit a residual class of audible qualities that are bearers-optional — which is just sounds sneaking back in through the timbre door. The second option is, I think, your best line, and the chimeric-sounds paper in your folder (Siedenburg, Jones-Mollerup & McAdams: timbre dissimilarity has separable acoustic and categorical components) supports it — a chunk of "timbral" phenomenology is trained categorization riding on top of perception, which is exactly what your cognitive-error machinery is built to absorb. ## Can this become the paper you never wrote? Honest assessment: yes, and the situation is more favourable than in 2016. What's changed: Casati, Di Bona & Dokic's Ockhamization made source-identity respectable enough to cite as a halfway house (your AHS draft already positions you as going one step further — they keep the audible event as "the sound", you deny even that residue is needed); the Miller-et-al. cognitive-error program gave eliminativism about a *seeming* a reputable template; and the timbre literature (Siedenburg & McAdams 2017, the Oxford Handbook of Timbre 2021 — in your folder) handed you the wastebasket premise with full psychoacoustic credentials. The 2022 doc is a genuine skeleton of a self-standing paper: presence criterion → sources pass, candidates fail → no-consensus argument against the seeming → error theory. Its gaps are the ones I've listed: the colour-consensus premise needs shoring, the pluralist escape needs closing or co-opting, Farkas 2013 needs defusing, and the acousmatic residue needs an explicit treatment — for which Coastline-style resynthesis is honestly a better case-generator than anything in the 2016-era literature, because it lets you dial source-hearing's own grouping cues up and down and watch where the phenomenology of objecthood, imagehood, and "timbre" each give out. The auditory images material is a second, separable paper (the Sessions file already treats them as siblings), and its missing third objection is the one the plugin manufactures. I'd say the one decision that shapes everything is whether the new paper leads with the negative claim (no sounds — the Farkas/Chalmers/error-theory line) or with the articulation claim (timbre and the rest redistribute onto things, forces, and perspectives — the Handel/psychomechanics line), with the other as support. They need each other: elimination without articulation leaves the aficionado's tinniness dangling; articulation without elimination is just one more property view. But that's your ranking to make, not mine. Want me to capture any of this — the Coastline test cases, the Farkas-2013-versus-Budek-&-Farkas tension, the two-wastebaskets reading — into the Auditory Images session file or a standalone note, so it's there when you come back to the drafts? *Chi ha smontato il timbro in mille proprietà udibili non ha più bisogno del cestino in cui gettarle.* ### User it's late, i would like tomorrow to be delighted by a perfect paer wtrtitten by be , no more than 4000 words. about the fact that we defintily don't hear sounds i have so much notes on this stuff, the question is can you extract from thedrive stuff, i am very happpy about the phenpmeolgy doesn't tell on whether sounds are individuals vs properties, so much the worse for auditory perceptiopnk. but you know thims, there is an obvious elegant argument. rmake sure that you explain your analysis, justification, reasons etc etc. for your answer BEFORE giving me your final answer. ### Assistant I'll build this properly: first load the deep-writing-mode workflow (the canonical path for drafting in your voice), then pull the remaining Drive drafts so the paper is extracted from your own material rather than reinvented, then give you my full analysis and plan before the draft itself. ### Assistant 4,331 words — over the limit. Trimming ~350 words without losing argumentative content. ### Assistant Still 190 over — final trim pass. ### Assistant Done — the paper is written and saved. As you asked, my analysis and reasoning first, then the result. ## Analysis and justification What I extracted, and from where. The paper is assembled almost entirely from four of your own documents: the September 2020 "Against Hearing Sounds" Drive doc (the most developed prose — it supplied §1's four reasons, §2's taxonomy, the Campbell/fox material, and the recordings section), the February 2022 "Sounds, Eliminated Entirely" doc (the phenomenal presence criterion, the no-consensus argument, and the whole error-theory section, which existed only as bullets), the vault [Against Hearing Sounds - Draft](My Obsidian Vault/Notes/Against Hearing Sounds - Draft.md), and the dissertation (the pitch/timbre/loudness challenge on its p. 18, which I answer in §5 using the transistor-radio material from its Chapter 5). Wherever you had finished sentences, I kept them; my own writing is confined to bridging, to completing your bulleted skeletons in your register, and to one substantive addition described below. The architecture, and why. You said the elegant argument is the one you're happy with — phenomenology doesn't settle whether sounds are individuals or properties, so much the worse for sounds — so I built the paper so that everything funnels into it. §1 gives the four reasons sources are heard (immediate awareness, causation, grouping, function). §3 then runs those same four tests on sounds and shows that three fail outright, using Budek & Farkas's phenomenal-presence criterion to establish what a candidate for hearing must satisfy. That structure is the justification for the elegant argument's weight: by the time §4 arrives, seeming-to-be-heard is the only ground sounds have left, so destroying the seeming destroys the view — the argument isn't one consideration among several, it's the last one standing. §5 (the Miller-style error theory) then discharges the residual explanatory debt: if there's no seeming, why the universal belief? Your sound-friendly-language story answers it. §6 handles the two objections your drafts treat as most pressing — Campbell's grinding noise and Soteriou's pure audibilia — the latter compressed to two paragraphs that gesture at the auditory-images account as a companion paper rather than developing it, which is what the 4,000-word budget demands and what keeps the two projects cleanly separated. The one substantive addition. In §4 I built in the defence I pressed on you last time: a critic will say philosophers disagree about colour phenomenology too. The paper now answers this with the gloss-versus-category distinction — colour theorists dispute the metaphysical gloss on an agreed seeming (Chalmers's Edenic description, Strawson's window-gazer), while sound theorists dispute the grammatical category of the seeming itself (Nudds's independent individuals against Leddington's bound properties against O'Callaghan's unfolding event-like individuals against Kulvicki's equivocality). If sounds were present the way colours are, the debate would open with 'sounds seem to be Fs; what are the Fs?' — and it doesn't, because there is no agreed F. Without this paragraph the argument is vulnerable to a one-line referee objection; with it, the disagreement's shape becomes evidence rather than noise. What I left out, and why. The Farkas 2013 structural-conditions problem, the headphone/in-head-localisation crunch, and the Coastline-style resynthesis cases all belong to the auditory-images companion, and raising them here would hand the referee your own best counterexamples inside the wrong paper. The timbre material appears only in one compressed paragraph answering the bearers-of-audible-qualities objection; the full wastebasket argument deserves its own treatment. And I kept your fallback 2021 view out of the text entirely — the paper's dialectic never needs to name "Sounds as Properties", though you should expect a referee to. Voice and grounding. Per the deep-writing-mode protocol I grounded in "Sounds as Properties" (Thought 2021, read whole — the model for interlocutor-entry through exact quotation and for the §3 dilemma construction) and "Hearing Objects and Events" (Phil Studies 2018 — the model for question-as-puzzle openings and the restrained roadmap). British English throughout, no metacommentary, long connected sentences, examples developed rather than listed. Honesty flags. All quotations are copy-pasted from your own documents, never reconstructed — but several lacked page references there, so the text carries [REF] markers (Kulvicki's veil and no-consensus quotes, O'Callaghan's "audible individuals appear to occur", Strawson's window-gazer, the Miller et al. citation) and the bibliography carries %%check%% comments (Budek & Farkas venue, Nudds 2013 source, Schuster 1986). These want verifying against the PDFs before this goes anywhere. Word count: 4,015 counting the title, headings, and all block quotations; your own prose comes to roughly 3,580. ## The paper It's at [Against Hearing Sounds (short paper draft).md](My Obsidian Vault/Writing/research/auditory images/Against Hearing Sounds (short paper draft).md), in the auditory images research folder, with the session file's last-worked updated. Structure: introduction → §1 Hearing Sources → §2 Hearing Sounds → §3 What Would Justify Sounds? → §4 Do We Seem to Hear Sounds? (the elegant argument) → §5 Sounds as Cognitive Error → §6 Objections → conclusion, ending on: audition never tempted us to insert a veil between perceiver and world until we started talking about it. Sleep well — it'll be there in the morning. When you've read it, the natural next passes are a source-check on the quotations and a depth audit on §4 and §5, since those carry the paper. *Chi non sente mai i suoni dorme comunque benissimo: sono le fonti, non i suoni, a svegliarci nella notte.*