VinsCool

@Pat Brady wow I just took a look at your tables in the google docs... holy fuck this is awesome stuff, it's all exactly what I was looking for and what I was attempting to do with my own set of frequencies as well! Thank you so much, this will totally save me a lot of time trying to calculate everything by hand and by ear In fact, from what I could tell, I got really darn close to a mix of your Major 64khz tables, which looks like I had kind of "combined" based on the notes I was able to find resonating nicely together, so I really was on the right track! Seriously, thank you so much!

Thanks a lot! That could be really helpful for what I'm trying to achieve.

Ah I see now, unison and octave resulting sound, that's indeed very pleasant, and a bit too easy to get weird because of the tuning.

That would be pretty damn cool! especially considering how the POKEY tuning is limited, it would help a lot on some tunes to achieve the best possible harmonisation The values I theorised here were based on as many notes as possible I could get in-tune to each others, and I believe the same principle also applies to distortions since I got my Distortion C bass to suddenly sound "in-tune" as well, as a side effect, which was unexpected but really welcome. There's still a lot of improvements and most likely more alternative tables that could come into living from this concept, and I do plan to experiment with these for sure, hahah. hopefully I can get a finalised table for $A in both 64khz and 15khz modes that are compatible, so far I was at least able to get a preliminary 15khz table that has the 2 first octaves in-tune to A-4=440, but it will be complicated to use since (at least, RMT) can only use the same table regardless of the mode used. And doing the fine tuning manually in the tracker itself is technically possible, but it's painfully slow and very easy to mess up lol Edit: by the way, I am so thankful for POKEY Explorer, it has helped me so much figuring out tuning and combinations of sounds as well as many more tricks I am starting to use more and more

There you go Corridors of Time Alternate Tuning 50hz.xex It sounds a bit different since I only really adjusted the tempo, nothing else was changed. Tuning may be a bit off as well, but I plan to look into PAL later as well Oh yeah might as well post the modules. They rely on the tuning of my custom notes table in the RMT version I posted earlier in the thread, but otherwise they might still work fine in Vanilla RMT 1.28 (1.30) Corridors of Time Alternate Tuning Final Loop Adjusted.rmt Corridors of Time Alternate Tuning Final Loop Adjusted 50hz.rmt

15kh mode, or 16bit tuning combined to high pass filter can work for creating convincing SID bass on POKEY

Doing some custom tuning is slowly paying off, but there's still a lot of things I need to figure out. Maybe I'm crazy or maybe I'm onto something but I hope I can fully get most notes to sound better between each others. For now I only changed the Distortion A in 64khz mode tuning table, where the actual pattern of notes being "in-tune" appears to be around 444hz in NTSC, running the same frequencies in PAL is a near 440hz pattern 😮 For the moment I had to make my own patched RMT version to actually be able to test what I have figured out without having to make an instrument for literally each note, but that did work well I believe. A few notes may still sound a bit wrong but otherwise most do sound pretty good together, there is a nice resonance going with the distortion bass now :3c Corridors of Time Alternate Tuning.mp4 To make that RMT2LZSS conversion I had to be a little creative but thankfully that did work well, haha In short I exported a .xex from the patched RMT version I made currently, recorded as SAPR in Altirra and then threw it into RMT2LZSS. The song looping twice then fading out was actually done manually, couldn't be bothered to manually make the "loop" out of a SAP dump The RMT version itself is actually nothing too special, it's just the Distortion A table changed to my own table, which is still not perfect so I wouldn't recommend using this one seriously, in fact I do not like having to rely on a hacked version at all, but it made my tests a lot easier to do The .txt file is mostly a dump of ramble, ideas, and guesswork, so again I do not recommend taking anything from it too seriously, but maybe it could be useful, it's definitely something I want to improve later. Corridors of Time Alternate Tuning.xex Rmt Vin Tuning Patch V1.exe POKEY Table 64khz (A=443hz) WIP14.txt

If my ears aren't deceiving me, this is the tune.

That really depends on how much changes would be done in the first place too. If for example it's mostly a tuning improvement, that would still work but most likely sound weird in RMT itself. If it's a listenable state that would at least give an idea of what happens before conversion. I've gone to the point where I rely on export/conversion behaviour in Altirra (and now real hardware!) in the first place anyway, so having to convert a couple times between versions isn't truly annoying me I personally would totally love this instead of manually tune notes or make hacked up instruments just to get something to work, the only difference would be that it would then become RMT2LZSS exclusive features, haha It's so clunky... but doable. Exactly this!!! Yes! I've been doing quite a bit of tuning and combination of frequency/AUDCTL stuff and they absolutely require very specific set of stuff going, making them work manually in RMT isn't impossible but it's becoming overwhelming when you get 10 extra instruments for a handful of notes, haha Even worse is that RMT by default does not even have the most optimal tables of notes for most distortions and modes, so a lot of the stuff, if not all of it, sounds out of tune as a result, which probably gave the chip such a reputation for the last few years --for most people, or even myself until I tried to dig a little deeper with that stuff, it starts sounding weird very quickly. Back to a year ago when I attempted to poorly cover the Prince of Persia tunes (I'm still a bit sad about deceiving you about it...) I didn't actually understand why things just sounded bad (or good), and that was directly related to this. There was a lot of stuff that was just... coded as is and then it was up to the user to figure it out. Clearly a more experienced person can get away with it, but for those who don't know the chip very well, it made absolutely no sense, and then stuff just sounded that way because of how the note tables were implemented, which clearly contradicts more recent, and better tunes even an idiot like myself was able to do with some (a LOT of) dedication and manual tuning being fed into the projects In fact now I totally understand the motivation behind patched versions of RMT, it's pretty much a necessity if you aren't patient to get certain things done manually, and it's also far from being the most user friendly kind of stuff as well. Having a feature as such would make certain projects INFINITELY easier to manage, seriously 😮 As much as I love using RMT to make Atari chiptunes, there's so many things that could be done better, and it makes me sad to know there aren't many alternatives that don't require a musician to also learn programming or have a workflow made mostly of hacked up instrument and tracking design (the latter applies to myself currently, haha). Its creator is also no longer with us , and all we can get for now is either a handful of hacked versions that aren't even cross compatible between each others (a module created in one will sound wrong in the other and vice versa), or a lot of patience and a very hacky workflow, which was absolutely necessary for me to achieve some better sounding tunes. All that bit of rambling to say how much I am thankful for the dedicated people who have helped me a lot to learn new tricks and improve my own skills, and having tools like RMT2LZSS is really awesome to push the elaborate sound design further. Who knows, maybe we'll get a new tracker someday, and maybe there is still a lot of surprise from the 8 Bit machines awaiting to be found! Thanks! I also agree for the disappointing part lol I knew it was possible to get 50hz design to sound good, but in reality it's only just the 100hz tune with some tweaks that was played back with every other frames skipped, which is surprisingly not that bad after the tweaks! I'm pretty sure 100hz could produce even better results, but I am not the most skilled to make use of it at its fullest yet I believe the .rmt could indeed be trimmed down considerably too, having a lot of instruments (and I mean, a LOT because it was absolutely necessary!), volume/speed commands and unique patterns that cannot be optimised further eat up a lot of memory very quickly. Speaking of optimisations, how does LZSS work at its core anyway? Is it entire song compression? Or are there some special tricks involved to reduce file size? I have noticed that just playing a couple identical patterns for a few times can drastically increase the file size, which makes me wonder if there couldn't be a way to literally remove any redundancy by making certain parts (like, patterns specifically) only compressed once, and then having them called into memory each time they are needed instead of adding them up infinitely? A bit like how the loop point was implemented, except it would then become part of the whole data output, being made up of "subtunes" that can then be loaded each time they're actually required, which would then avoid having any duplicates in the compressed data, similar to the way trackers are made, if that makes any sense Speaking of loop point, I was wondering if there could be a way to actually chose how many times a tune would loop, until maybe fading out using a very crude volume change based on the volume offsets? Actually there would be many things that would be nice to have but I don't wanna start to sound like a nagger so I'll leave that for a better time and thread

Well speaking of experiments I finally gave a proper attempt at converting my Battle Squadron cover from 100hz to 50hz and got very positive results! I'd consider this one about 90% identical to the 100hz version, which is really impressive in my opinion Battle Squadron Title 50hz Reduced WIP.xex So how did I achieve it actually? I didn't even try to rearrange my module from 100hz design to 50hz, I actually let RMT2LZSS do all the work for me, and only adjusted the module itself to find out what frames were "skippable" and what shouldn't be, so I was able to make up a direct conversion with barely any losses! Pattern data remained intact, I only had to adjust the instruments themselves, and sacrifice what had to be sacrificed to fit half the amount of frames used to forge the sounds. It's not perfect yet but I think that speaks for itself just by the sound, I think I'll try to polish it up a bit more later, but that may be as good as I could make it I think.

More or less, yes that's exactly that! How to use the said table of notes is an entirely different story, however In my case, I did all my notes by hand using several instruments to match my frequencies, so that in return remains 100% compatible with RMT 1.28. Most people seem to prefer hacking RMT directly, which also works a lot better on the long run, but then incompatibilities between modules and RMT versions bothered me so I tried to be a little bit creative with my tunes instead. haha That honestly could be really awesome, actually... To have a way to apply our own set of patches and frequency tables directly into RMT2LZSS, that would make certain experiments infinitely easier! Something like overriding the original RMT tables with custom ones right in the conversion process 😮 No idea how that could even work however. Having a new tracker would be even more for sure, but that might take a really long time, I imagine 😆

A few days ago I started doing an experiment related to the tuning itself. I have noticed RMT uses the same standard Distortion A table for both 64khz and 15khz modes. While by itself that isn't a big deal, I eventually noticed that almost everything in 15khz mode was out of tune due to this implementation. So I ran my own attempt at tuning notes directly using POKEY Explorer and a tuner as a reference, then wrote down my own table of notes. Even if not perfect, it's considerably better compared to the original implementation, making octaves 1 and 2 (bass range) a lot less off I think. It's not very easy to make use of it manually, but not impossible either, just takes a bit more time to get it to work without having to hack RMT itself. So here's a quick test I did using one of my own tunes in progress, which was the ideal candidate for that test since it uses mode changes very heavily. Note that I only applied the changes on the bass itself running in 15khz mode, everything else was left unchanged, so it might still sound a bit off as a result. Just compare to the recording in the post just above, I hope this could be a useful improvement over the original out of tune bass (Edit after reloading the page several times: sorry for posting links to an external place, looks like the site is very slow right now and I'm unable to attach files >.<) 15khz tuning table for Distortion A, NTSC: POKEY Table 15khz (A=440hz) V1.txt -> I also learned a bit later synthpopalooza made one before me, which is also almost identical, haha Test tune using the improved 15khz bass tuning (without actually changing the normal 64khz mode): Temple of Questions Tuning Test 6.xex The same tune in progress that was posted in the previous post above, with the lovely 15khz bass being out of tune Sketch 24 v36.xex

I really don't know how to describe either, honestly... It just seems like LZSS data fed into the register is a lot more stable and likely to produce predictable results, which is a really good thing, in my opinion!

Temple of Questions New WIP2.mp4 Been revisiting an old tune of mine, been a lot of fun to design so far

Tried to be a little more creative with the video format for this one! The Tower of Turmoil - The Atari PoKEY Cover Soundtrack (Recorded from Real Hardware) The audio itself is also available in the GDRIVE folder I posted in the other page.

That would make a lot of sense. I noticed real hardware was also quieter, and I had to manually amplify the sound when I had it recorded, but the quality doesn't suffer from it, thankfully.

Altirra has been surprisingly really good from my own tests, everything expected in Altirra happened on hardware almost exactly the same, which is really good news to me. Hardware has that hum and hiss that doesn't exist on emulator, however, and seems to not clip as much as higher volume, so a few things that sounded distorted on emulator sounded perfect on hardware, which is once again really good news I think the real deal was how my cover of Battle Squadron was behaving on hardware. I used a PAL 800xl, the executable ran at 100hz, and everything I had finely put together using Altirra as a reference was accurately reproduced, and sounded even better than ever. The only downside was that when I recorded individual audio channels, few things were impossible to fully predict because of the nature of the chip, such as buzzy Distortion C bass timbres or noise. My oscilloscope and the sound recorded from the full tune didn't 100% match visually because of it, since I basically had to record that tune 5 times, once for each channel, but it was good enough to for me. The pulses manipulation using high pass filter were perfect however, because I had absolute control on them, they were predicted and very consistent. As show in the channel 2 of that video, the pulses were exactly how calculated throughout the entire song, and leaving it to loop for several times still produced identical results, hahaha. I have more stuff planned to be recorded that way too. Not everything will be recorded 5 times to make fancy visuals however, because that takes time to actually record and synchronise everything One thing I can say for sure, however, all my future work will (and should!) be intended for real hardware playback accuracy, and now I know Altirra is a really good emulator to get the overall idea of what it should sound like The RMT2LZSS conversions also make things a lot easier for me, for example using the volume offset to mute channels, that saves a lot of time over editing the .rmt modules several times Anyway speaking of hardware recording, I have got this edit you sent me a while ago that I really liked. I had changed it a little bit, but the overall idea is still there. The end section sounds so organic, compared to what I originally did. This is actually where I began to understand how the chip really worked, and has since been the design I have used and improved for my later songs. [PAL] VinsCool - Amberstar - Ode to Schnism Emkay Edit (Hardware).ogg I might try to touch it up a little more someday and make a new upload to replace the version I have made first, which I can tell isn't very good compared to what I can do now. This edit however, sounded great on hardware I recorded it off the PAL 800xl as well.

Tried something a little different with the video format. An old soundtrack, featuring real hardware recording. Haydenwoffle - Atari PoKEY Soundtrack (Recorded from Real Hardware) The audio itself is available in the GDRIVE folder I shared in the previous page of this thread too. I'll make a couple more uploads like this, with a bunch of stuff I never released as well, soon

I'm disappointed, I expected my module would have been used for this version The high pitched sounds in the intro running at 1.79mhz are a little too high pitched to my taste, too

should be this one I think! This song was composed by Spring, not me, btw I also planned to record it from hardware, this one runs at NTSC speed! judas_betrayal_of_christ.xex

In the meantime here's some more stuff by Spring that deserve more attention! There's always a way to get some impressive results from RMT itself

Alright, here we gooooo! Welcome to Battle Squadron Battle Squadron - Intro & Title (Atari PoKEY Cover, Recorded from Real Hardware) Module and executable converted using RMT2LZSS attached to the post. This one is a PAL 100hz tune. I'll squeeze out a proper 50hz version out of it later, it's more than possible to make it still sound good Battle Squadron Title 100hz FULL.xex Battle Squadron Title 100hz Final.rmt

I'm sorry, I simply don't understand what the problem is, really 😅 I personally liked the way I did my version, but I imagine there's always room for improvements I'll post the finished version in this thread soon!

Honestly I have no idea, I wasn't even able to tell what the speech was saying. I just know it used a few channels at once on the Amiga to produce it, so I tried to do something sort of similar Also thanks! Couldn't be identical but hopefully not too bad