The No.1 Website for Pro Audio
Oberheim Matrix 1000 firmware hacks
Old 8th January 2018 | Show parent
  #601
Lives for gear
 
tux99's Avatar
 
🎧 5 years
Quote:
Originally Posted by EvilDragon View Post
If you're not using edit buffer, no, they're not.
You will struggle to find an editor that doesn't use the edit buffer.
Old 8th January 2018
  #602
Gear Guru
 
EvilDragon's Avatar
 
🎧 5 years
Seems like nobody even made Bob aware of that bug. I've emailed him and he told me that. He's going to check it out.
Old 8th January 2018 | Show parent
  #603
Lives for gear
Quote:
Originally Posted by EvilDragon View Post
Seems like nobody even made Bob aware of that bug. I've emailed him and he told me that. He's going to check it out.
I just mailed him too [I hadn't made him aware yet because I've not yet verified it's a 1.20 thing]. Didn't hear back yet, but the day is young.

So Edisyn's workaround for the corruption bug is as follows. Normally Edisyn can (1) send single parameters to the working memory of a synth (2) send all parameters to the working memory of a synth (3) write all parameters to a patch in permanent store. Since the M1000 can't do #2 reliably, Edisyn reserves patch slot 199 as a scratch space. When the user "sends" a patch, Edisyn just writes to the backing store on slot 199 and then does a PC to that spot. When the user writes a patch, Edisyn writes it to the correct slot and PCs to that. This works fairly invisibly except that 1.20 shows "199" on the LCD to (correctly) indicate the slot being used.

The alternative is to send every single parameter as a separate NRPN. But you can't do this very quickly -- you have to pause significantly between NRPN messages to ensure reliable transmission -- so it's not useful for doing things like patch mutation and hillclimbing.
Old 8th January 2018
  #604
Gear Guru
 
EvilDragon's Avatar
 
🎧 5 years
I got a reply from Bob. He confirmed that behavior in 1.11, 1.13, 1.16 and 1.20, so it's not a new bug.

In fact, he posted a reply in the thread about the bug: Undocumented Matrix 1000 Sysex Bug
Old 24th February 2018 | Show parent
  #605
Here for the gear
 
🎧 10 years
I've just got back into my matrix 1000 after a few years of not using it. I had done the CPU upgrade and installed 123R2 back then.

Is that now dead? As in it seems 1.16 and 1.2 are what everyone is using now.
Should I revert to the old CPU/8Mhz and try one of these firmwares instead?

If so, is there a clear winner between 1.16 and 1.2?
Old 17th May 2018 | Show parent
  #606
Deleted c1b33bf
Guest
Clicking Envelopes

Hey everyone, I've been using the matrix 6 editor, and suddenly had this issue with clicking envelopes.
It occurs if I softly, or quickly press a controller key, or the computer keyboard on any patch.

I've tried freeing up lfo's, envelopes, vca settings, velocity settings, global settings etc but nothing seems to help!

Anyone else run into this issue with the 1000/matrix6 ? I'm using an older firmware.

Matrix-6 – Ctrlr
Old 18th May 2018 | Show parent
  #607
Lives for gear
 
🎧 5 years
Have you tried 'self-calibration' function? RetroSound - Oberheim Matrix-1000 service information
It helps when my Matrix 1000 starts behaving weird like filters/envelopes going out of sync or some keys produce no sound at all

(also if you are using M1000 this one is a better choice matrix-1000 – Ctrlr )
Old 18th May 2018 | Show parent
  #608
Lives for gear
Quote:
Originally Posted by kraze View Post
(also if you are using M1000 this one is a better choice matrix-1000 – Ctrlr )
Try Edisyn.
Old 18th May 2018 | Show parent
  #609
Gear Head
 
🎧 5 years
Quote:
Originally Posted by EvilDragon View Post
I got a reply from Bob. He confirmed that behavior in 1.11, 1.13, 1.16 and 1.20, so it's not a new bug.

In fact, he posted a reply in the thread about the bug: Undocumented Matrix 1000 Sysex Bug
If I read the thread well there is no bug at all. It was an errorneous patch which caused all this whirl. But the thread is hard to read with all this sensless flame in there. If I interpret it well tux99 was quite upset because of feijai's "false alarm".
Old 18th May 2018
  #610
Lives for gear
 
tux99's Avatar
 
🎧 5 years
It was a patch containing an invalid parameter value that feijai knowingly created leading us all on a wild goose chase. It's only natural that I was annoyed when I found this as he basically abused our willingness to help each other (which is what forums are for), by knowingly wasting our time with a non-existent problem.

That thread should really be deleted as the problem doesn't exist and it probably keeps confusing people who happen to find it.

But this has nothing to do with the topic of this thread so shouldn't be discussed here.
Old 18th May 2018 | Show parent
  #611
Deleted c1b33bf
Guest
Quote:
Originally Posted by kraze View Post
Have you tried 'self-calibration' function? RetroSound - Oberheim Matrix-1000 service information
It helps when my Matrix 1000 starts behaving weird like filters/envelopes going out of sync or some keys produce no sound at all

(also if you are using M1000 this one is a better choice matrix-1000 – Ctrlr )
Thanks Kraze!

I'm using the matrix6r, and use workarounds to avoid freeze ups. For example, every time I use porta I get problems.

I ran the calibration a few times but it didn't seem to effect the envelope issue.
I did notice the sound was more focused compared to just running the tune function.

On closer inspection I realized the worst offenders were patches with low resonance.
This machine seems to be more clicky without it.

The envelope reset function helps a bit with that, but I think I'll have to resort to equalizing out any issues.

Last edited by Deleted c1b33bf; 18th May 2018 at 04:27 PM..
Old 18th May 2018 | Show parent
  #612
Lives for gear
Quote:
Originally Posted by Possemo View Post
If I read the thread well there is no bug at all. It was an errorneous patch which caused all this whirl. But the thread is hard to read with all this sensless flame in there. If I interpret it well tux99 was quite upset because of feijai's "false alarm".
The history goes back earlier than that. Despite what I think are its clear improvements over 1.16, tux99 doesn't like v1.20 of the firmware because, so far as I can tell, it displays value changes on-screen.

In development of the edisyn patch editor for the 1000, I came across what I thought was a bug: I could create certain patches which when *sent* to working memory were corrupted, but when *written* permanently worked just fine. The Matrix 1000 has incomplete (frankly incorrect) sysex documentation regarding sign-extension of 6-bit and 7-bit data, and it I had read it wrong. But my invalid patch was being handled properly by the 1000 in one situation but improperly in another. Frustrating. :-(

I posted asking about a possible 1000 bug. Tux99 immediately proposed that the problem was (!) the v1.20 firmware. This was testable, so I provided patches which exhibited the bug and asked for people to try them on various firmware. But rather than simply try the patch on 1.16, as would be normal tester behavior, Tux99 just kept insisting that he'd never seen a bug like that when he developed *his* editor on 1.16, (BTW I have not seen this editor. Which one is it?) and so therefore it "must" be 1.20. Finally I and EvilDragon got ahold of Bob Grieb, who made v1.20, and he kindly tested it not only on 1.20 but on 1.16 and various earlier versions, and verified that the bug in fact occurs everywhere.

It was at this point that tux99 then finally looked at the files, and noted two bytes with invalid encodings. This led me to nail down that I had misinterpreted the signed extension documentation which was tickling this m1000 behavior. After fixing that, things worked great.

I'm a software developer and an academic and I well understand the ego investiture which leads to ridiculous catfights like this: it's the nature of coders, and I am guilty of it as much as anyone else. So rather than gang up on tux99 again, I would like to note that ultimately it was tux99 who provided the clue which led to nailing the bug. And I did not thank him for this nor even attribute it to him. So I'm going to remedy that and thank tux99 now.
Old 18th May 2018 | Show parent
  #613
Gear Head
 
🎧 5 years
Quote:
Originally Posted by feijai View Post
The history goes back earlier than that.
---snip
And I did not thank him for this nor even attribute it to him. So I'm going to remedy that and thank tux99 now.
Thanks for this elaborate reply. I think it is a pity that all this had to end as a fight as you probably all like the Matrix-1000. So do I.

BTW. I am not a coder but I made an editor/librarian for the Matrix-1000 based on Ctrlr. Your post about signed values worried me a bit. I had a look into my code, I didn't remember how I handled signed values: If I am not completely wrong the method used to represent negative vaules is called "two's complement":

All 8 bits of the two 4-Bit nibbles are used. Positive values from 0 to 63, negative values start at 255 for -1 and end at 192 for -63. So you definitely need the 8th bit. I don't exactly understand what your problem was, but as stated before I don't know a lot about coding. When using Ctrlr most of the hard work has been already done.
Old 25th May 2018 | Show parent
  #614
Lives for gear
Quote:
Originally Posted by Possemo View Post
All 8 bits of the two 4-Bit nibbles are used. Positive values from 0 to 63, negative values start at 255 for -1 and end at 192 for -63. So you definitely need the 8th bit.
Actually that's not quite how signed two's complement works. You can do two's complement in any number of bits if it fits, and -63, indeed -64, fits just fine in 7 bits. What's interesting is that it's easy to convert from one number of bits to a longer string of bits (signed extension).

For example: +60 in 7-bit two's complement is 0111100. To get -60, you flip all the bits (to 1000011) and then add 1, which gets you 1000100.

And +60 in *8*-bit two's complement is 00111100. To get -60, you flip all the bits (to 11000011) and then add 1, which gets you 11000100.

Notice that the only difference is that 8-bit has an additional 1 in its leftmost bit (11000100 vs 1000100). Similarly, -60 in *12-bit* is 111111000100. And -60 in *16-bit* is 1111111111000100. You get the idea -- you can bump up to higher bit lengths by just repeatedly copying the leftmost bit. This is signed extension.

The documentation flub is as follows. Remote Parameter Edit (06H p. 43) says "Note: all values are sign extended from bit 6 into bit 7 except for parameter 121 (VCF frequency)". It says a similar thing for 0BH. Note it doesn't say "into bit 8". Just extending bit 6 to bit 7. This is of course because the MIDI byte is only 7 bits long and these parameters are only 7 bits long, so why bother.

But in the patch data format, it describes various parameters as "7-bit (signed)" and one as "6-bit (signed)", so they would fit in 7 bits just like remote parameters. It does not say they're supposed to be sign-extended to the 8th bit [though they are] and are thus represented differently from the remote parameters. That's what messed me up: I was assuming consistency. And in fact it works fine if you write the patch, presumably because the first thing that the Matrix does is truncate or sign-extend the high bit. *That* is bizarre.
Old 25th May 2018 | Show parent
  #615
Gear Head
 
🎧 5 years
That makes sense, thanks for the explanation. Of course I talked about the sysex dump format. They had to find a way to pack the 8-Bit two's complement into the MIDI 7-Bit format. This probably is because the Matrix-1000 has an enhanced firmware based on the the Matrix-6 which didn't allow tweaking the modulation parameters. On the Matrix-6 you have to send the whole patch sysex dump if you want to edit modulations. Bob Grieb Matrix-6 firmware allows editing the fixed modulations the same way as the Matrix-1000.

Quote:
Originally Posted by feijai View Post
... -63, indeed -64, ...snip
If I remember well -64 is the same as value 0.
Old 16th June 2018 | Show parent
  #616
Lives for gear
 
🎧 5 years
Well I have an 1.16 ROM and it allows this

A single track and a single, but heavily modulated, preset (except for a slight crossmod-style Eventide send towards the second half) and tweaking with Oxygen 49 controllers assigned to internal Matrix 1000 functions through Ctrlr panel

Old 8th September 2018 | Show parent
  #617
Lives for gear
 
Gnalvl's Avatar
 
🎧 5 years
Quote:
Originally Posted by EvilDragon View Post
I got a reply from Bob. He confirmed that behavior in 1.11, 1.13, 1.16 and 1.20, so it's not a new bug.

In fact, he posted a reply in the thread about the bug: Undocumented Matrix 1000 Sysex Bug
Just want to say thanks for clearing up this confusion. I had been waiting to pick up an M1000 for a long time, figuring updated firmware would sweeten the deal. Recently I finally spotted one with Tauntek v1.2 installed, but I'd never seen that name before (only Gligli) so I decided to do some research before buying.

When I first encountered threads saying v1.2 had bugs, I almost decided to pass. Luckily my instinct to keep reading into it and not jump to conclusions prevailed, as when I finally got to the end of this thread I discovered the truth that it was a false alarm. Thus I was able to jump on the deal and not let the opportunity pass me by.
Old 13th February 2020
  #618
Gear Addict
 
🎧 5 years
Hey - I've upgraded to 1.2 and want to update a few basic midi CCs directly from my Novation SL Mk III - dont want to use CTRLR app

Anyone got a template or a way of creating one?

Or... a CC list for stuff like cutoff?

Sorry for the noob questions!
πŸ“ Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 1387 views: 187838
Avatar for Squeegee 303
Squeegee 303 1 week ago
replies: 54 views: 9525
Avatar for Teddy Ray
Teddy Ray 12th November 2009
replies: 57 views: 37415
Avatar for Deleted 7a792f4
Deleted 7a792f4 12th April 2019
replies: 86 views: 8049
Avatar for bdk.nyc
bdk.nyc 31st July 2020
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
πŸ–¨οΈ Show Printable Version
βœ‰οΈ Email this Page
πŸ” Search thread
πŸŽ™οΈ View mentioned gear
Forum Jump
Forum Jump