Hacking the 404SX

7th CITY · **Joined:** Wed Jun 18, 2008 2:22 pm **Posts:** 216

someone....anyone...make a bank JUST MONO. JUST ONE BANK MONO-TRIGGER thats all i am asking! THATS ALL I WANT. (FOR NOW) oh is that even possible? i read someone talking about auto chop....yeah someone hook that up too!

SP-Forums is the best.

cartesia · **Posted:** Thu Jan 14, 2010 4:03 am

PHeMoX wrote:

Jim Y wrote:

I'd really like to see an app that can export the 404 sequence patterns as midi. Otherwise they are stuck in the 404. Probably the same process applies to 404SX and the original 404.

Both SP-404s don't have midi out though (not over USB either), so it might require hacking and updating of the actual OS for that.

Hey man we're talking about reading/writing the pattern files from the SD card, so should be no need for a midi out port or any device hacking

ModuLR · **Posted:** Thu Jan 14, 2010 8:09 am

PHeMoX wrote:

I see, that in itself means trying to build one ourselves isn't a futile endeavor at all. :p

Could you post some screenshots of that editor just to get an impression?

That's the utility that comes with the SX. It works, but could be so much better.

ModuLR · **Posted:** Thu Jan 14, 2010 8:43 am

Jim Y wrote:

That's what I meant by "apply trimming", because if the sample will be effectively re-written to the new length, then the existing SX utility should do all you need - it ought to produce a trimmed sample when you pull it off the card (provided you've applied the truncate)?
There's no Sample Start followed by a loop region such as instrument type samplers understand on the 404, so knowing the loop points of a trimmed/truncated sample probably won't be necessary - That's IF the sample really does get truncated of course

Yups.. the sample doesn't get truncated unless you tell it to delete the excess on the 404. But as of now, you can't set the sample start/end to a custom value via the utility. You'ed either have to edit the sample first, then load it or do it on the 404.

Quote:

Thinking about the patterns, I realise I have no definite idea of how a sequencer actually plays midi. Knowing the sequencers PPQN (ticks) rate and the maximum pattern length, bar count and smallest quantise value might help.

Well I just broke out renoise and built a pattern from one pattern files on the SD card... and got it to match. Some parts I'm still trying to figure out, but at least I was able to extract the timing, note, velocity info. Which means I see no reason why a midi export wouldn't be possible.

More nerding... the SP is 96ppq... so that 384 ticks for every bar. This is the binary from the pattern file...

44 43 00 00 7F 60 00 40
1C 46 00 00 63 7E 00 0D
35 44 00 00 74 60 00 0F
35 44 00 00 74 60 00 0F
17 43 00 00 7F 7F 00 0A
46 43 00 00 7F 65 00 23
30 46 00 00 77 6C 00 14
63 44 00 00 77 7C 00 2F
46 43 00 00 7F 5E 00 3B
19 46 00 00 70 7F 00 0D
2F 44 00 00 7B 6F 00 1F
18 43 00 00 7F 7E 00 0E
47 43 00 00 7F 76 00 31
2E 46 00 00 74 58 00 14
60 44 00 00 74 7A 00 22
00 8C 00 00 00 00 00 00
00 02 00 00 00 00 00 00

starting from the top...
44 43 00 00 7F 60 00 40

44 = delta time to trigger next note
43 = note 43 (Bank B / Pad 9)
00 = flag??? doesn't seem to ever change (yet)
00 = flag??? doesn't seem to ever change (yet)
7F = Velocity
60 = ??? (no idea)
00 = flag??? doesn't seem to change (yet)
40 = ???

Obviously I have a ways to go... but the timing info allowed me to build the pattern... works like this

Play note 43 (Pad 9/Bank B),
move forward 44 (Hex) ticks
play note 46 (Pad 12/Bank B),
move forward 1C (Hex) ticks,
play note 44 (Pad 10, Bank B),
move forward 35 (Hex) ticks,
etc.

One part that's weird is this two lines..
35 44 00 00 74 60 00 0F
35 44 00 00 74 60 00 0F

They are identical, and the sp seems to only play one and ignore the other.. when I did this the file came out perfect timing wise. It was identical to the SP.

At the end of all patterns is something like...
00 8C 00 00 00 00 00 00
00 02 00 00 00 00 00 00

I think the 02 is the number of bars in the pattern, cos when I did one with 1 bar, it changed to 01.

I'm sure some of the unknown stuff has to do with quantize. I also did some real simple test files and it added what appeared to be empty notes so it could pad the timing out so it always adds up to number bars x 384 ticks.

Still figuring this one out... but progress none the less.

drewzle · **Posted:** Thu Jan 14, 2010 11:09 am

very impressive ModuLR!!

i'm keen to see where this is going

PHeMoX · **Posted:** Thu Jan 14, 2010 1:51 pm

cartesia wrote:

PHeMoX wrote:

Jim Y wrote:

I'd really like to see an app that can export the 404 sequence patterns as midi. Otherwise they are stuck in the 404. Probably the same process applies to 404SX and the original 404.

Both SP-404s don't have midi out though (not over USB either), so it might require hacking and updating of the actual OS for that.

Hey man we're talking about reading/writing the pattern files from the SD card, so should be no need for a midi out port or any device hacking

Yeah, I was hugely mistaken. I thought both SP-404s had USB connection as well, giving us access in the sense I described.

Seems the cards and what comes off of it, is all we got.

That's less than I had hoped for.

cartesia · **Posted:** Thu Jan 14, 2010 3:44 pm

Damn ModulR... You'll have done all the fun bits of figuring it out before I even get my 404SX!

One thing I'm thinking maybe you do to test out the patterns -

make a pattern which is just one sample pad held down for a beat or so then let go and analyze that.. or maybe hit it a few times, holding it for different lengths each time.

There may be some kind of note-off message as well as note-on. (as there is in midi)
OR
it could be that there is a value for how long the note is held in the note trigger message.

Possible has something to do with the 3rd last or last byte?. . heck maybe all 3bytes at the end even (could the 00 being the same in each be to do with quantise? or swing? thats probly guessing to far for now.. but yeah..just check the bytes for length to hold information or check for note off triggers.)

eg

17 43 00 00 7F 7F 00 0A
46 43 00 00 7F 65 00 23

Did you hold that note down for longer the first time than the second (or longer the second time than the first?) ? try switching the sample to gate mode then record a wav from the pattern playing - look at it to see the length - or measure the time or number of samples each hit is long, and see if the ratio matches either 7F:65 or 0A:23 - or even 7F000A: 650023.. but then I guess 7F:65 would be close enough to tell.

Les Braun · **Posted:** Thu Jan 14, 2010 6:06 pm

It'd be awesome if you could somehow improve the 12-voice polyphony.

sharpmotif · **Joined:** Wed Nov 25, 2009 11:14 pm **Posts:** 222

Focus people: hacking the binary files on the card = messing with whatever the SP-404SX records to organize the sample and pattern data.

We're (or rather They're) talking about making a software thing to edit samples and turn patterns into MIDI sequences.

Hacking the actual OS is a bit far from this.

PHeMoX · **Posted:** Thu Jan 14, 2010 7:25 pm

COMPUTERS wrote:

It'd be awesome if you could somehow improve the 12-voice polyphony.

Can't be done unfortunately, that's a hardware limitation.

ModuLR · **Posted:** Fri Jan 15, 2010 12:08 am

cartesia wrote:

Damn ModulR... You'll have done all the fun bits of figuring it out before I even get my 404SX!

hurry up and get one, lol! I built some more patterns, and you're right about the note off bits... plus I learned a few other things.

Example 1 bar pattern with 1 note.

FF 53 01 00 7F 40 00 1D
6D 80 00 00 00 00 00 00
14 80 00 00 00 00 00 FF
00 8C 00 00 00 00 00 00
00 01 00 00 00 00 00 00

(again, all of this is hex)

FF = # ticks to next note
53 = pad # and bank (next switch determines which bank)
01 = bank switch (00=banks ABCDE, 01=banks FGHIJ)
00 = another switch?
7F = velocity
40 = it changes value, but ???
00 1D = # ticks to turn current note off. It's 2 bytes long which means you can
provide a number large enough to cover the SP's 99bar limit on a pattern.

now notice in the pattern it's got this bit...

6D 80 00 00 00 00 00 00
14 80 00 00 00 00 00 FF

My guess is that it's padding out for timing reasons. Note 80 is not a pad on the SP, and notice almost everything is set to 00 except that last FF (which I don't understand).

However if you add all the "on" timing bytes...
FF+6D+14 = 180 (hex)
which is 384 (decimal), which is the exact number of ticks in a bar on the SP.
If you had a 2 bar pattern, the timing byes would add up to 768, etc.

again, that last block...
00 8C 00 00 00 00 00 00
00 01 00 00 00 00 00 00
seems to tell the sp this is the end of the pattern file, and I have confirmed that the 01 is reflective of the # bars in the pattern.

ma_headphone · **Joined:** Tue Jul 28, 2009 10:42 am **Posts:** 6

It would be nice to be able to make patterns on on the sp, export them to a computer, apply swing, and get them back on the sp404

PHeMoX · **Posted:** Sun Jan 17, 2010 2:52 pm

ma_headphone wrote:

It would be nice to be able to make patterns on on the sp, export them to a computer, apply swing, and get them back on the sp404

Yeah, cosign to that! Did anyone ever try to use the Midi in on the SP-404 to let an external sequencer kick in and do a similar thing?

I think I might be able to use the MV sequencer to get an easy swing going.

22tape · **Joined:** Sun Jan 25, 2009 4:32 am **Posts:** 62

great stuff, modulR. good to see a familiar face here!

i'm considering buying a 404sx. a couple of things are holding me back:

viewtopic.php?f=4&t=6865&p=60907#p60907

of course i'm not asking for feature request already, but i'm wondering would this sort of hack have the potential to address these two issues?

thanks!

*edit*

nevermind, i just re-read the thread. this wouldn't be possible. still, great ideas anyway!

ModuLR · **Posted:** Mon Jan 18, 2010 12:34 am

@22tape... yeah, the sp is kinda restricted in that sense, but honestly I kinda find the workaround more fun than a regular sampler. I think that's part of what makes the SPs addicting. lol.

update... I got a lot of the basics figured out for patterns code wise... and managed to write a test program that *successfully* exported from the SP! Muahaha! yyyEESsS!

(btw, that a factory pattern #1 from the sx)

Hacking the 404SX

Who is online