Photosounder Spiral development updates

[A_SN]

It also works in SAVIHost, but seems a bit stuttery, and has a message at the bottom that says "Low frame rate. Set your hosts buffer to <1700 samples (40mS)"

Well yeah, the message is telling you just how to fix this! SAVIHost defaults to 10 buffers/second for some reason, but you can safely lower it (I use 100 buffers/second (441 samples) without any cracking problems).

I didn't notice any difference between the CPU load of the SpiralCM and Spiral (mono): both pretty low CPU.

Why don't I want gain? Because my gain is set up right If it isn't I'll adjust the controls in the DAW.

Actually gain in the plugin could be useful, but IMO, SpiralCM is a thing of simple and functional beauty, but Spiral risks looking too big, confusing, and complicated. Features vs Complexity trade off.

Err.. what? The gain is pretty essential, and has nothing to do with the volume gain, so you might want to change the visualisation gain even if the sound volume is just fine.

Anywho I just don't get why you care if there are controls you don't use, whereas I care if I have to go through some menu or popup just to access something that could be right there. I'm against the Google kind of minimalism, you know, removing as many features as possible and hiding the rest into some button's menu's submenu's dialog's tab. If it can all fit all at once on the screen then that's the way I want it .

[busywait]

OK, thanks for the tip regarding gain. My version of Spiral would auto-gain, but I'm not planning to have time to write that atm

I often map plugins to the control surface that I have, and that only has 8 rotary controls. Pretty common?

Anyway, my preference regarding the number of controls is "don't make me think" (if you can avoid it). But you're not designing nuclear power plant control panels, so I'll leave you to it

[A_SN]

Good news everyone, a new beta of Spiral is out (since yesterday, I waited until the new KVR to announce it) and it finally has the most important feature remaining, which is the time-stretched playback. It took me a while but the functionality is pretty polished and I've already tested it a lot (mostly to learn guitar solos).

Here's how it works: after you've recorded a chunk of sound it starts playing back at the rate defined by the Speed knob, which is between 0x and 1x. If the rate is exactly 1x then you hear the original sound unchanged in full quality. If the rate is less than 1x then you hear the sound played back at that rate, for instance 0.05 means the sound is playing 20 times more slowly, which makes a world of difference when trying to transcribe a tune for instance, and also it sounds pretty cool.

Adjusting the Visualisation Resolution affects the time-stretching resynthesis resolution, so what you see really is what you hear as what you hear is resynthesised from exactly the same data you see and nothing else. You can change playback speed and resolution pretty smoothly, even between the unchanged 1x rate and the other speeds.

The next features to come are a small visualisation area in the Record/Replay area so that you can scrub through time which would make it much easier to go back in time than with using the Position knob which isn't precise enough, the visualisation thing in question would allow to be a lot more precise as you'll both see the notes and whatever else is going on and you'd be able to drag it around to go back or forward. Also I'm considering making the Pan Window settings affect the sound, at least for the time-stretched sound, probably also for the unstretched playback and maybe I'll make an option to affect the input live.

Oh also don't bother with the AAX version even if it's still in there, I just found out that it cannot work given that I don't have Avid's PACE signatures, which I'd be supposed to buy... However if anyone wants I can do RTAS builds on Windows.

[A_SN]

Did anyone get to try the time-stretching?

[Naenyn]

Just wanted to stop by and say that I recently discovered Spiral and it is beautiful! Is it still under active development? I'm not sure how useful I'll find it to be, but I'm definitely interested to see where you take it.

[A_SN]

Just wanted to stop by and say that I recently discovered Spiral and it is beautiful! Is it still under active development? I'm not sure how useful I'll find it to be, but I'm definitely interested to see where you take it.

Yep, working on it right now! I'm reorganising most of the code (because it's a big mess with copies of the same code across functions and projects, with some versions more advanced than others) and putting all the stuff that all my recent projects use into a common library (which I will then open source), when it's done I will make SplineEQ use that library and release a new and improved version of SplineEQ. Then after that I'll complete the extra controls and visualisation for Spiral's playback feature, and then Spiral will be finally complete!

[Naenyn]

Awesome, that's great news. I like SplineEQ quite a bit. Spiral is definitely cool (especially visually!) but I'm not sure what I'd use it for. What do you see being the primary purpose of the plugin? How do you envision producers using it?

[A_SN]

Awesome, that's great news. I like SplineEQ quite a bit. Spiral is definitely cool (especially visually!) but I'm not sure what I'd use it for. What do you see being the primary purpose of the plugin? How do you envision producers using it?

For me its primary purpose is as a learning tool, I let it run while listening to music and using this I can analyse and learn any music very efficiently, it really takes the guesswork or the tedium of other approaches out of it. But it's not just about the notes, it can also teach you a great deal about proper stereo mixing and give you a distaste for the nowadays all too common "put everything in the centre and apply widening effects to some parts" approach. So it's meant more as a learning tool than a making tool, and it might be more relevant to someone who plays instruments than to your average dubstep producer, but it's also meant as a great way to see what's going on in a mix, mostly if you're used to seeing other people's music with it. Analysers that show you a jumpy frequency curve don't tell you much except maybe whether boost a certain range or not, and oscilloscopes don't show you much besides the dynamics, Spiral just shows you everything as it happens, of course you see the dynamics and the relative strengths of different ranges, but you can also recognise instruments, see what they're made of, see how everything overlaps and interacts, and you can compare with what you've seen through it before.

A good example is if you're producing something with a very distorted electric guitar sound, you'll know from experience looking at other similar tracks through Spiral before how it looks, how strong that "ring" that loops quite neatly around the 7th octave looks like so that you can make it sound right, not too high or too low or too wide or too narrow, how much distortion noise in that ring there should be between the bright lines that represent the harmonics, how much of the harmonics outside of that ring you should be able to see, how much the attack should light things up, how the sustain looks like over time and so on. People like to think it's all better and more natural to do everything by ear, but the power of analysis of your eye helps a lot so you don't have to beat around the bush, insisting on doing everything by ear is kind of like insisting that x-rays are not needed because you can just feel what's inside your body. In other words it's a useful help for most aspects of production, but for its potential to be fully realised you need to learn, simply by making a habit of looking at music through it, then you'll know what looking through it can tell you.

[chonicle]

Is spiral still in development?
For the harmonics overlay, one thing to consider is intensity gradients for 1st/2nd/3rd... (1st is the strongest) rather than the current all white dots. Also for a music structure revealing visualization(like spiral), a chord overlay would also be very useful. It's acturally already pretty useful, just looking at the inner two or three circles for not-so-loud songs will tell you what chord it is . You can even add some colour design to these chords for more beautiful representation(This is the original purpose, right? The spiral representation is not that good for single-pitch note compared to a linear/exp representation, but for chord it's a much better representation compared to a linear/exp.)

Human voice is distracting and overlapping with chord area. Well it can be solved somehow using the spectral pan by avoiding the central area, but still not so good(no way to get around this I guess). Also for loud/noisy songs, it's just too much stuff going on there and can't get much information out of it(Well sometimes you can see the spiral shape when you hear a synthesizer sweeping the frequency, that's cool).

This spiral representation idea is denifitely brilliant! Thanks you for making such a plugin

[A_SN]

Yes Spiral still is in development, although right now I'm actually working on both SplineEQ and Photosounder at the same time!

Why would you want the other dots to become fainter? Seems like it would only make them harder to see. I made every one of them easy to see with a concentric black-white-black pattern because there might be stuff going on below the overlay that would otherwise easily drown them out visually.

What do you mean by a chord overlay? Can you give me an example of how it would work? How would you colour the chords? I didn't really specifically consider chords (I guess you mean in a guitar strumming kind of way, right?) when thinking up Spiral, just a practical way to see the notes in music, more specifically the single notes of each instrument at the same time.

The spiral representation is not that good for single-pitch note compared to a linear/exp representation

I couldn't disagree more, because you can tell what note something is without even looking at the note label just from seeing the angle, works like the hands of a clock really, whereas with a classical flat representation you have to see how it matches with a linear sequence of notes, plus you have to zoom in, viewing the 9 or 10 octaves that might be present isn't practical in that case. Plus since the harmonics line up in a familiar pattern in the spiral representation you can more easily avoid mistaking an harmonic for a separate note, you see what belongs together as a single note.

The way it works for me is that first I try to identify the key and scale used, then when I do I rotate the visualisation to have the key be on top, then I mentally visualise the scale, for instance if it's played in D minor I set the Key knob to "D top", then regardless of the key I know that the whole minor scale is found at 12 o'clock, 2, 3, 5, 7, 8 and 10 o'clock, so if I see that familiar pattern of harmonics pointing towards 5 o'clock it instantly means something to me (a perfect fourth from the root in this case).

Spiral is meant to show you everything as it is, so you might also say that other instruments distract from the instrument you're interested in, not just the voice . I myself rarely feel bothered by voices because they move in a very different way from anything else, if anything I find it too easy to tune it out and ignore it. Actually my biggest problem is when the lead instrument and the rhythm section drown out the more subtle accompaniment. As for the loud songs you're right, though one solution is to crank up the Gamma, it makes fundamental tones stand out usually, by making things more contrasted and dimming broad noises and other less dominant parts of the spectrum. Another problem is when the song is poorly mixed in stereo (like most of everything made in the last 20 years) and that everything is centered, "stereo width" being only an effect. Then it might as well be mono, and everything ends up looking about the same colour.

Do explain your colouring ideas though, I'm open to ideas on different visualisation colouring schemes, mostly to help with mixes that are too mono or centred, and also maybe ways to make different octaves clearer, although this one isn't obvious since knowing the octave of the overtones doesn't necessarily help you identify the octave that the note actually is in very much. Also I'm thinking of changing the rest of the interface to something darker, as the brightness of the grey background kind of works against you when trying to see dark shades.

[chonicle]

OK I see your primary concern is to get individual single note out.. I'm more interested in using this to figure out what chord(or key combination) it is. Here is a comparison between a guitar playing some note(on the left) and a song that has a guitar playing the chord(also other stuffs, not much. On the right).
Rather than focusing on the individual harmonic series, I would say it's a better tool to determine the chord... If there is a overlay which allows you to click on these notes and show you something like the chord name, it will be a very useful tool to probe the song structure. The diminishing intensity also applies for most natural instrument(like guitar/piano etc), high order harmonics are rather weak and more importantly, in a real song it is nearly impossible to match these higher order harmonics(see the spiral on the right in the picture, that's already a song basically all guitar. There are harmonics that can be figured out through comparison). The fundamental frequency is the most important feature that can actually been seen in a real song, which shows you the chord(The most important stuff when you transcribe songs). This picture shows a E2, B2, D#3, F#3, G#3 which is a EMaj9.

[A_SN]

OK I see your primary concern is to get individual single note out.. I'm more interested in using this to figure out what chord(or key combination) it is.

Well the problem is kind of the same, and it's solved by using the harmonics overlay tool for each of the notes that makes the chord. I think I probably should find a way to have several of those at the same time for each candidate note, because with just one at the same time it can be hard to remember if everything is accounted for or to which note something belongs. I guess if I do that then I can also display a chord name.

Not sure how I would handle managing different notes and their overlays though, maybe the same way as I handle control points in SplineEQ, that is double click to create, double click a point to remove it, and drag the points around?

Edit: Oh and maybe a way to move all the points together if for instance your Emaj9 then becomes a Amaj9, right?

[chonicle]

AUTO-ADMIN: Non-MP3, WAV, OGG, SoundCloud, YouTube, Vimeo, Twitter and Facebook links in this post have been protected automatically. Once the member reaches 5 posts the links will function as normal.

Well the problem is kind of the same, and it's solved by using the harmonics overlay tool for each of the notes that makes the chord. I think I probably should find a way to have several of those at the same time for each candidate note, because with just one at the same time it can be hard to remember if everything is accounted for or to which note something belongs. I guess if I do that then I can also display a chord name.

Not sure how I would handle managing different notes and their overlays though, maybe the same way as I handle control points in SplineEQ, that is double click to create, double click a point to remove it, and drag the points around?

Edit: Oh and maybe a way to move all the points together if for instance your Emaj9 then becomes a Amaj9, right?

Well if a chord name can be displayed that will be useful. And yeah move all points together EM9 will become AM9, that's where the rotation tool is useful. Make a overlay for each type of chord(m,M,m7..etc) and click on A -> display Am/A/Am7 overlay, should be very nice for chord detection
Here is a more realistic situation with a relatively loud song with different chords: http://1drv.ms/1qROGGh (http://1drv.ms/1qROGGh)
Left is the original song, right is the guitar chord I made according to spiral visualization(of course also need some listening). The chord part is pretty constant and bright in each measure and can be figured out Human voice is also obvious.

[A_SN]

Make a overlay for each type of chord(m,M,m7..etc) and click on A -> display Am/A/Am7 overlay, should be very nice for chord detection

That's a good idea, I think I'll make an "overlay mode" knob with first a simple single note mode as we have now, then a list of chords that would display all the notes and their overtones (so you would simply locate the correct note and everything else should line up) and then the aforementioned multiple note mode with the displayed detected chord for the times you can't just guess what chord it is.

[A_SN]

We're getting there! I just have no idea what colour I could make the Spiral logo be. Or maybe it just needs white outlines? Also I don't know if I should leave the frames (that frame groups of knobs together).


Also is it weird how there's a big black rectangle in there? It's for visualising/going through the recorded sound, so it would only show when you're going through a recorded bit of sound. Otherwise I guess I should leave it empty and black?

Also busywait I realised you were right and removed a bunch of controls. Partly because I couldn't find an alternative pan gradient that was any good, partly because when I saw in YouTube videos what people did I knew I had to remove that.