mrmr9494

Chandler: first up, attached the "DSP adjusted" X3000 rhythm for E FUNK POP which you requested. As for the remaining rhythms: - R&B BALLAD: This is similar to "E FUNK POP" and "DISCO SHUFFLE" - some hidden aspect of the Preset rhythm just disappear as soon as it's converted to AC7. Unfortunately in the case of R&B BALLAD the effect is dynamic (it's the "Manual" input of the Wah DSP which is being modulated as if it's a Midi CC controller). I don't know how to incorporate dynamically modulated DSP effects into an AC7 rhythm. Maybe it's possible on a X5000 (as it's related to how the MODULATION/ASSIGNABLE button works), but I don't have access to a X5000 so for now this is .... not possible to reproduce exactly in AC7. - All Others: In every other case you've brought up, the Present X3000 rhythm matches most closely with your "X700 Swap" recording. If I load your two AC7 files into the X3000, they sound identical to each other and also identical to the preset X3000 rhythm. So it's the X700 preset which is odd-man-out here. The most interesting is "HIP HOP", where it sounds like a completely different drum sound is being used! (Snare vs. Hand clap). It's not anything to do with filtering or DSP, it's just different music! If it helps, I could make a recording of each comparison. There's one attached here for "COUNTRY POP". You can tell a difference in the timbre of the sound between the recorded audio and direct playing of rhythms, but that's unavoidable when recording and playing back audio on different audio system. Ignoring that timbre difference, the guitar sound of the X3000 preset is definitely closest to the "X700 Swap" audio: 0:00:00 - 0:00:05 Recorded audio "Factory Preset" 0:00:06 - 0:00:10 X3000 preset rhythm 0:00:11 - 0:00:15 Recorded audio "X700 Swap" 0:00:15 - 0:00:19 X3000 preset rhythm (again) Long story short, there's no easy way to improve what you already have for any of these rhythms. X3000 EFunkPop with Added DSP.AC7 comparing_Country_Pop.mp3

This is odd. I don't hear any difference 🤔. And reading with Sysex doesn't show up anything other than differences on unused tracks and other unimportant stuff. So for example, comparing "Dandiya" between your X700 file and the original ("Preset") X3000 rhythm, they sound the same except V2<>V3 are swapped (I guess that's intentional?). Similarly, "Rock Anthem" also sounds the same -- guitar has same DSP to my ears -- except that again parts are changed around. I tried a few others from your list as well and nothing was obvious. Is it possible X700 actually sounds different? There's unfortunately hundreds of parameters to read which might be different. There's a fast automated system for reading them all at ac7maker but it only works on Linux and it doesn't interface with MIDI-Ox or anything like that. In the case of "EFunk Pop", the DSP setting and parameters for the specific "problem" part are read with the following two: F0 44 19 01 7F 00 03 03 0E 00 00 00 00 00 00 00 00 00 55 00 00 00 00 00 F7 F0 44 19 01 7F 00 03 03 0E 00 00 00 00 00 00 00 00 00 57 00 00 00 0D 00 F7

Very interesting. This isn't anything wrong with ReStyle or even the CT-X700, it's actually the X3000 file which you're starting from which is missing the effect. If I load that file into my CT-X3000 the guitar has the "plinky" (dry) sound. If I try to save the preset rhythm "002 ELECTRO FUNK POP" in my CT-X3000 as a user rhythm it also has the plinky sound. I think what you've stumbled on is that preset rhythms in the CT-X have data that isn't saved to User rhythms. Maybe the presets are not even in AC7 format -- they're in some other (secret) format that has more flexibility than AC7. For another example, the Intro of "DISCO SHUFFLE" has a sweeping high-pass filter on the guitar part that goes from sounding "tinny" at the start of the intro to sounding full-bodied just before it goes into the variations - that's an effect that's not possible with AC7. If it helps, I'll try putting the original effect into your "X700 13-24" file. Will upload it here. EDIT: now added a file "X700 13-24 with Added DSP.AC7" which I think should sound like the original. To get it I used SysEx messages to read mixer state and then adjusted the file to get the same settings. The same method won't work on "DISCO SHUFFLE", because there the filter state is changing dynamically. X700 13-24 with Added DSP.AC7

Good news. The GUI is clearly progressing nicely. 1) There were originally some "get...." functions to complement the "set...." ones but I removed them because thought they wouldn't be useful. That was a weird decision - I can't now remember the reasoning behind it. Will add them back in. (UPDATE: now done: link). 2) Mclandy's file is from the CT-X5000 so it should cover pretty much all the variants. Some of regional instruments from the "Indian" keyboards are missing, so I could have a go at adding those in. But yes, it's fine for CT-X7/800 and even CDP-350 (maybe even PX-S3000?). UPDATE: Regarding credit: No credit in the GUI please. Where the library is used in the code, I'd appreciate a link to the github. Also if your code includes a copy of mine then please make sure the license files and "setup.py" are there too. Beyond that, nothing needed.

I'd reckon "Fingered 2" is a good one to start with. It gives you a lot of control without overloading with advanced features (e.g. "Finger On Bass").

I've put a library up on Github which I hope will do the trick. Everybody welcome to use it and give feedback. @Chandler Holloway, please pass on to your friend and maybe we can get things going on this. Here are a couple of pointers that you can pass on: 1. The "Quick Start" section of the ReadMe should have all the information that's needed for this project. Everything else in the documentation can probably be ignored. 2. If possible try to use the package locally without installing. (In practise, that means placing it in some xyz directory, and importing with "from xyz.casio_rbk.casio_rbk import ....". If needed can install it with "python setup.py install", but that's hopefully unnecessary).

Mclandy has very kindly provided a patch list (not derived from Casio's patch files) which I'll put up on Github. Cheers @Mclandy!

Chandler, here are the final two banks that you suggested above, using Volume=0 instead of turning parts off since that appears to be how it must be done for CT-X700. That seems to conclude the experimental portion of this project. We were able to get Mixer levels and pan controlled through registrations. Effects don't seem to work the way they do on CT-X3000/5000 for whatever reason. We didn't try Portamento, but from looking at the files it seems highly unlikely that functionality is implemented in the CT-X700 in a way that could be used. Is there anything else you wish to try out? Thanks for noting about Duet mode. I'll note that in my documentation, I think no need for detailed testing. I now have some work to do to get a Python library put together and expanding the documentation of what we've found. Will aim for a couple of weeks. RBK1.RBK RBK2.RBK

Also: 006.RBK - similar to 004.RBK, but with the reverb sends re-ordered to the end of the structure. Worth trying every possibility. 006.RBK

Good news indeed that you have someone to assist with the software. I will do everything I can to help out, only hope I can keep up with your pace 😃. I really don't know what's going on with the Reverb sends. The values are all in the registrations you export so there's no reason why the same information wouldn't be imported back in the opposite direction. So strange. As you guessed, the re-exported 003 is indeed removing the extra bytes, so in fact the re-exported 003 is an identical file to 002. That's expected, the CT-X3000 does the same thing when it's given a CT-X5000 registration bank it will simply remove the stuff it doesn't understand. Attached 2 more test files: 004.RBK - this is identical to 002.RBK except re-ordered so that the reverb sends now come before the tone specifications. I don't expect it to make any difference since ordering hasn't made a difference before. Worth a try though. 005.RBK - uses your original (piano/string/bass) set-up but testing out pan and on/off: Slot 1: All parts panned center Slot 2: U1 hard left, U2 + L center Slot 3: U1 hard right, U2 + L center Slot 4: U1 off, U2 + L on. I think it's worth trying out this simple test before your full suggestion. Very likely the X700 doesn't recognise On/Off so we will need to go for your Volume = 0 idea. The bank here works on the X3000 so let's try it on the X700. 004.RBK 005.RBK

@Chandler Holloway would you do more one test? If you load 002.RBK and press the "Registration 2" button to load the 2nd registration. Then go into the menu to where Reverb is defined. What is it set to? We're expecting "Delay 2".

Interesting observation. I've never noticed it before. For me it happens sometimes and not others, the difference being whether the tone being switched to has DSP or not. If yes, then the effect immediately changes, if not then it does as you say. Interesting do get this window into what the keyboard might be doing internally. You're absolutely correct that SysEx can "see" the DSP settings and control them (although Casio doesn't document how to do this). It's just that the .RBK files can't. Also, rhythms certainly use DSP but be aware that there are multiple effects engines. Keyboard tones use engine "DSP 1" while rhythm tones use "DSP 2 & 3". If your goal is to affect the DSP on keyboard tones then rhythms are probably not going to help you. AC7 files. Possibly also SysEx message.

I was also thinking of @shiihs ReStyle application, it's very similar to what we are trying to achieve here. Although it seems that you already have the graphical design element sorted 🙂. There's a lot to go through in your post, so let's start with Delay. Firstly the encoding: X700 Reverb "Delay 1" = X3000 Reverb "Delay". Not surprising. The attached file 002.RBK is as you suggest - starting with "Delay2.RBK" and adjusting the Reverb send and Delay send values for each slot exactly as you specified. When I loaded this into my CT-X3000 the Delay type became "Tone" but there is still a distinct delay effect on Slot 3 (quite a short one, similar but not the same as the "Short 2" setting). So your hypothesis that Seq. Saw would default to nothing doesn't seem correct, at least on the CTX3000. Let's see on the CTX700. The file 003.RBK is a bit of an experiment. It is the same as 002.RBK but with the 0x5E atom (that's the one that controls Reverb/Chorus/Delay types) set to a CT-X3000 format. On the X700 the atom is only 2 bytes long (since it only has reverb & chorus), but on X3000 it's 3 bytes long. There are several things that might happen here: - The X700 will reject the file - It will accept it but do nothing - It will actually change the Delay to the specified type (it is "Long Pan 1" - very obvious if it's applied or not!) 002.RBK 003.RBK

@vbdx66, too bad to hear about your brush with Covid, how long has it taken you to get through it so far? Good that things are already improving. I can answer some of these questions. The idea is you would be able to control Volume levels and Reverb/Chorus effects for keyboard tones using RBK files. Whether Delay effects can be changed remains to be seen (Chandler found last week they can be controlled for rhythms, so that's at least a promising sign as far as it goes). It's nearly certain that DSP won't be able to be controlled using this method - DSP is entirely defined by the selected Tone, and so not having a method to load .TON files is a limitation on what can be done there. I completely agree about the computer-based editor. While X3000/5000 have a pretty good built-in tone editor, there are lots of under-the-hood parameters that could be made available, and also the ability to tweak effects while "in the mix" (e.g., while editing a rhythm). Such an editor is also something I dream of. Unlikely that Casio will provide it though.

Also I've had another look at the interchangeability of the X3000 vs. X700 registration .RBK files, and it turns out there are very few differences - mostly just the number of registrations (4x for X700, 8x for X3000) and then some superficial stuff in the "header" portions of the files. The actual data content (the Registrations themselves) are very interchangeable between the models. That might open up some possibilities for ways we could use the files. Things I've checked: - Rhythm numbers seem to be encoded the same between X700 & X3000. - Reverb & Chorus are very similar: X700 Reverb "Off" = X3000 Reverb "Off" X700 Reverb "Hall 3" = X3000 Reverb "Hall 3" X700 Reverb "Plate 2" = X3000 Reverb "Plate 2" X700 Reverb "Delay 2" = X3000 Reverb "Long Delay 1" X700 Chorus "Tone" = X3000 Chorus "Tone" X700 Chorus "Flanger 4" = X3000 Chorus "Flanger 4" X700 Chorus "Light" = X3000 Chorus "Chorus 1" X700 Chorus "Chorus 6" = X3000 Chorus "Deep" It might be that the effects aren't actually identical between the two, just that they are encoded the same. Someone in Casio might have decided that "Chorus 6" was close enough to "Deep" that they could use the same encoding.

That's really good that it's worked out. I really enjoy doing this stuff, and when it ends up being useful that's even better. We can continue with the CDP-S350 next week. It could be internally quite different from the CT-X so interested to see how far we get. The other thread I guess to come out of this is a question of how to make our finding useful to others, for example how a CT-X700 owner could create an .RBK file to set things up as they wish. It really needs a user-friendly software solution which is something I'm not at all skilled at. I'm more than happy making Python scripts -- here is the one for creating 001.RBK. Do you have ideas for approaches that would be better than Python?

@Chandler Holloway, in this project are you changing voice assignments to ones that are specific to the CT-X700, or are you keeping the X3000/5000 voices and letting the CT-X700 choose a similar one to "fall back" to? There's a very small possibility the "falling back" process is causing the DSP issues. If a non-DSP voice falls back to a DSP one, could it be pulling DSP resources away from other voices? It seems unlikely, but the DSP issue is otherwise quite a strange one.

@Chandler Holloway, that's quite a comprehensive set of changes! I've made a start updating the Github document based on what you've sent. Lots more to do. The file attached as 001.RBK should be the initial changes that you asked for above. My CT-X3000 rejects the file with the message "ERR WRONG DATA" which is a little bit worrying. Let me know how you get on with the CT-X700. I thought the files should be somewhat interoperable, but it appears maybe not. In the .RBK you sent one thing that really jumped out was that the "Delay Type" setting was missing. (Only "Reverb Type" and "Chorus Type" are there). It suggests that Casio has deliberately nerfed delay in the X700, and so we might not be able to use RBK to do that. I'm wondering now if the CT-X700 even supports delay effects in rhythms? To test that out, would you try TAPTAP.AC7 on your CT-X700? It's a simple rhythm that alternates between a note with delay and then one without. Is there a difference? The difference should be super obvious if it's there. 001.RBK TAPTAP.AC7

Chandler, yes very happy to help out where possible. I've only just started looking at .RBK files so can't really say what can be achieved. That said, both your ideas here seem very likely to work out. Can Chordana change mixer parameters (e.g. per-part volume, or turn parts on/off)? Lots of settings in the keyboard can be changed with SysEx messages, but not mixer parameters as far as I've seen. Tone parameters certainly can be changed in real time.

I made an attempt to make a rhythm with a "Versatile" guitar here: https://github.com/michgz/ac7maker/tree/master/examples (Example #1). I did it a while ago and the rhythm sounds a bit naff now, but it illustrates the point. The point was to turn off "Chord syncing" on the strum / fret noise sounds so they're not transposed. Unfortunately that's not something that can be done with the on-keyboard rhythm editor. The in-built Casio rhythm 024 "GUITAR 8 BEAT" also uses a "Versatile" guitar and they don't bother turning off chord syncing, they just have "Inversion" set to "On" so it's not transposed too much (the strum note still remains in the range C7-G#7 so still sound like strums). That seems to work fine too. For the note sounds I recorded to MIDI using an ordinary (not Versatile) guitar sound and then scaled the velocities in the MIDI file (using a program, but maybe some DAWs can do it?). I found it sounded good with down-stroke velocities scaled to the range 0-60 and up-strokes to the range 76-90.

@MclandyNice chart, that presentation does clarify it quite a bit. @Chandler Holloway, please do include in the guide.

@mclandy, Yeah it would make sense that it's for jazz-style chords. In terms of the table, yes the first line is what happens to Note 60 (C4), next line Note 61 (C#4) ... up to C5. That's assuming the chord being played has a root of C of course.

The raw data is available here, but it's all just a bunch of numbers. Your ears as a musician are likely to be the better guide in this case. Chandler, that's unless you have a way of presenting this information? As best I can see, the main difference between "Basic" and "Var 4" is that "Basic" tries harder to create 7th chords. If the recording is C-E-G and a C7 chord is fingered: - if "Basic": keyboard will play Bb-E-G - if "Var 4": keyboard will play C-E-G Even with the data, I'm not sure what is meant by a "tension" chord. Many of the differences are in the 2nd and 4th degrees of the scale, so maybe it's something to do with how sus2 and sus4 chords in the recording are handled? Again, your ears are going to be the best guide unless someone else can make sense of this. Just for anyone who wants to delve into the github data, the numbering starts at zero and "f-root" variations aren't included. So lines starting "0,..." are for "Bass", then "1,..." for "Bass 7th", "2,..." for "Basic", "3,...." for "Var 2" etc.

From trial and error. It was from looking to see if the CT-X could be used as a synthesiser (i.e. to see if changing the timing/filter data in the TON file could be used to create arbitrary sounds). I wasn't successful -- most of the data in the TON file do nothing, or just similar effects to the main parameters. I think on the CT-X most of the sound control happens at a more hidden level. Only interesting parameters in the TON file were portamento (1B4h), monophonic (value of 10h at 1BAh), and a fixed filter defined at 1BCh-1C2h.

It's possible by binary-editing the TON file as long as you know the secret byte offset of the portamento definition. For CT-X3000 it's at offset 1B4H, and I've attached a file which is X-SYNTH 1 with that value set to 0. The CRC value also needs to be re-calculated. I think TON files are not compatible between CT-X3000 & 5000; is that right? - if so this is CT-X3000 only. NoPorta.TON