pawaga

hi, here is a bug report. i made a fresh session with 4 drum tracks (64/2*64) on group 1. Copied the 4 track from group 1 to groups 2 3 4. On group 1, everything is fine, but in groups 2-4, track 3 has a weird behaviour : active steps are skipped. I joined a zip containing the session. Best regard. mbseq_debug.zip

hi, here is what i've done with some pieces of wood, aluminium and acriglass i had around, very cheap and yes, it's really usefull :) that way, i can use my seq also flat when i wish.

@mongrol yes it's possible to shave a bit the panel, maybe 1cm on each side, if you tie the cs to bottom. it would be a cool evolution. i did it that way because i'm too lazy to redesign the whole thing ! here is the original acrylic layout : @EsotericLabs i don't think i will use an ethernet module as i don't have use of the OSC and probably won't..i did not knew that when i build it, so.. i put an ethernet plug to fill the hole, to avoid *bugs* to come in. You will says that there are some holes on the sides, but those are for bees.

about the usb panel mount screw holes, i send an updated version of the svg file in the original post, i'm not sure it's 100% perfect but it should be good now. Of course you should check with datasheets if possible.

yes it is ! cheapo panel mount usb and ethernet plugs, from the bay. *beware* of the screw holes for the usb adapter, they are a little bit too distant (~1mm each ; it works as it is but it should be better). Here is the regulator board layout, adapted from the ambika schematics :

@mongrol yes, the cs is mounted with 10mm standoff, the midi io and quad iic are mounted on 12 mm standoff. the core on 10mm standoff. there are also 6 extra standoff from cs to bottom, they are ~ 61mm long : 2x 25mm + 1x10mm + 1 washer to make it really tight here is a shoot of the inside :

yes, they are cheap, but there is a cost for that : they do not fit exactly with standard tact switch : i had to put some tiny piece of wood inside the caps to allow them to push correctly the switch.. not a big deal but it has to be said

the caps are chip one from ebay, like those : http://www.ebay.fr/itm/25-Pcs-Round-Tact-Tactile-Piano-Key-Switch-Caps-Pushbutton-Covers-Gray-/231735004579?hash=item35f47b8da3:g:zS0AAOSwA4dWMYFE the top diameter is 9mm.

hi, if you want to go the ponoko route, i made a design with rounded buttons.. It can be a bit better as the button holes are a bit loo large for mine (~0.5mm, but it's still good enough) : the laser beam size must be included, so beware of that.. Fortunately, here is the svg file so you can modify it if needed :) seq_enclosure_pvi_3.svg.zip seq_enclosure_pvi_3b.svg.zip

hi, some troubles with the USB Host support here : i use a k-board which works fine on a pc. if used with the midiseq, sometimes, the noteoff or noteon events are lost. It seems to happen randomly though only when the sequencer is running...

hi, i just wanted to share my view : some times ago, i experimented a custom firmware which use the delay layer to record that human touch ; that way we can use the benefit of the 384ppqn, for any step resolution.. it would also be much easier that way to view and modify the quantization, re-quantize the whole pattern, use different quantization methods... The way i did it is just a hack and it would be better written in a proper way (with a bipolar % of the the step length, from -50% to +50%..). Maybe an idea for the seqV4+ ? best regards, Patrice

hello Thorsten, thanks for your help, i understand now. yes, i tested it and it work now for different clock dividers ! i also added the quantization : } else { //handle the delay layer // if a delay layer is present, record to it u8 num_p_layers = SEQ_PAR_NumLayersGet(track); u8 *layer_type_ptr = (u8 *) & tcc->lay_const[0 * 16]; int par_layer; u8 instrument = 0; u16 delay; for (par_layer = 0; par_layer < num_p_layers; ++par_layer, ++layer_type_ptr) { if (*layer_type_ptr == SEQ_PAR_Type_Delay) { delay = ((SEQ_BPM_TickGet() - t->rec_timestamp ) * 96) / t->step_length; if( delay >= 96 ) delay -= 96; // because quantisation has put the event into the next step delay -= (delay * seq_record_quantize) / 100;//quantize it SEQ_PAR_Set(track, ui_selected_step, par_layer, instrument, delay ); break; } } } The next point i want to tackle is the pitch layer : discrete values are not very useful for that, it would be nice to glide values between two step...(for CCs as well) best regards Patrice

hi, the "delay" layer is not recorded in live rec mode, so i tried something : I added this in seq_record.c, around line 472 : } else { //handle the delay layer // if a delay layer is present, record to it u8 num_p_layers = SEQ_PAR_NumLayersGet(track); u8 *layer_type_ptr = (u8 *) & tcc->lay_const[0 * 16]; int par_layer; u8 instrument = 0; u8 delay; for (par_layer = 0; par_layer < num_p_layers; ++par_layer, ++layer_type_ptr) { if (*layer_type_ptr == SEQ_PAR_Type_Delay) { delay = (SEQ_BPM_TickGet() - t->rec_timestamp) / 4; // not sure about the "/4", but give weird results if the delay val is too large SEQ_PAR_Set(track, ui_selected_step, par_layer, instrument, delay ); break; } } } it works but not very accurate... Anyway , so far, i like it! best regards Patrice

it works great now. thanks a lot !

About the nth feature, I think there is a problem with drum tracks : if i set a nth param on a step, it will work until another drum play on the same step.

:) Many thanks Thorsten, I love the new feature, loads of fun ! There is a display bug on my side : the nth parameter is too long by one char, so the params are not aligned with their step.. regards, Patrice

oh, it will be really great that way. thanks !

hi, I would like to suggest a feature : it would be another kind of layer that i call "nth" : if the value > 0 then the step is only played every nth bar. if the value < 0 then the step is muted every nth bar. for example for a 4 step sequence "CGDF" with nth values "1,3,-3,4" would give bar 1 :C-D- bar 2 :C-D- bar 3 :CG-- bar 4 :C-DF bar 5 :C-D- bar 6 :CG-- bar 7 :C-D- bar 8 :C-DF ... To implement it, you can use the % operator : if barNum%nth == 0 then play or mute i experimented that on some max patches and find it very useful to easily make long (or very long) musical patterns.

hi Gaël, have you seen that pages ? http://svnmios.midibox.org/filedetails.php?repname=svn.mios32&path=%2Ftrunk%2Fapps%2Fsequencers%2Fmidibox_seq_v4%2Fhwcfg%2Fwilba%2FMBSEQ_HW.V4 and http://svnmios.midibox.org/filedetails.php?repname=svn.mios32&path=%2Ftrunk%2Fapps%2Fsequencers%2Fmidibox_seq_v4%2Fdoc%2Fmbseqv4_remote_functions.txt you should be able to configure your seq for your keyboard by modifying MIDI_REMOTE_KEY in MBSEQ_HW.V4

Hi Thorsten, yes, it is fixed on my side too. Many thanks !

hi, just wanted to report an issue with the glide notes (MIDIboxSEQ V4.088 revision 2135) : When I record a sequence live, everything is fine until i exit the Rec. page, then the glide notes recorded do not send noteOff anymore...

ok, I found the missing code in midibox_seq_v4/core/seq_file_hw.c in function SEQ_FILE_HW_Read, around line 771, i added else if( strcasecmp(word, "DETENTED4") == 0 ) { enc_type = DETENTED4; } else if( strcasecmp(word, "DETENTED5") == 0 ) { enc_type = DETENTED5; } my ALPS STEC12E do no more jitter using DETENTED4 mode ! I love my brand new seq V4 even more now ! Really great machine and platform. many thanks to TK and all the MIDIbox community for making this possible. best regards Patrice

hi, did you solve your jitter issues ? I also have jitter troubles with my alps encoder... I hope i haven't melt anything.. the jitter happen only on the increment side, i use DETENTED1. as Borfo said, the extra debounce modes doesn't seem to be known anymore in the current mios build.. Is it still possible to use it ? The custom encoder modes shown would be also great if working. Another solution is hardware debounce with 10-100nf caps..haven't tried it yet. thanks for any clue best regards Patrice