Page 1 of 1

System 100

PostPosted: Fri Jan 18, 2008 11:58 am
by rachelp
Gee I love playing with it! I have been wasting my evenings jamming on it, triggering the sequencer from Logic using a Gate Pulse
from my Midi2CV interface and then keying the changes on the 101, with a big fat note on the 102 over the top. Instant "New Order" and a whole
bunch of retro stuff. It sounds huge. And the patch is dead simple....

The thing is, I am not recording any of it yet! One reason is because although the gate pulses appear to be coming in on time,
when I try to play the System 100 over MIDI from my master (going through the convertor) it starts to sound laggy and loses the groove.

What I actually do is patch the arp into the 101 on the console and then you can play the keyboard directly and it sounds great.
All the key changes stay in tune and the arp plays in tune very nicely.

But if I break out the connections from my MIDI2CV/Gate interface into the "Computer Input" on the back of the 101, it lags. Also, I want to
be able to play it over MIDI and on the 101 directly at the same time if I want. The only way I have found to do this is really dodgy.

It involves getting the CV out of the MIDI2CV and just plugging the patch cord in part way to the back. This works surprising enough,
it sort of "denormals" the connection so the keyboard and "computer in" are available at the same time. But then the tuning goes out,
by a whole fifth and I either have to retune or play in a different key. Also, the "computer in" does not retransmit to the 102 module
so I can't trigger both the 101 & 102 from the one connection, the 102 then needs a separate CV/Gate channel which then breaks my
whole workflow.

When I play the System directly from its own keyboard, it sounds great, but from the CV/Gate interface it really doesn't sound quite the same.

I guess my whinge is - do I record the audio track of the 100 "old skool" style from it's keyboard and then just play the rest of my
gear over the top using MIDI or do I try to program it in Logic? What I guess is the ideal is to play my key changes in Logic and trig
it from there, but it really doesn't sound as funky. Perhaps I should try using MIDI delay in Logic or something?


PostPosted: Fri Jan 18, 2008 2:48 pm
by heathen
If your going through your converters and monitoring from software you'll get latency, try adjusting the buffer down as low as it will go, or you may need to monitor another way.

Have you tried to actually record it in? If it is latency on the monitoring signal it should still actually record in time, you just wont be able to tell until you play it back.

PostPosted: Fri Jan 18, 2008 3:02 pm
by rachelp
heathen wrote:If your going through your converters and monitoring from software you'll get latency, try adjusting the buffer down as low as it will go, or you may need to monitor another way.

Have you tried to actually record it in? If it is latency on the monitoring signal it should still actually record in time, you just wont be able to tell until you play it back.

I'll give it a wizz this weekend. It's just doing it from the computer is really killing the groove. I think I might just lay a guide track down
in audio and then play against that, but I was hoping to record my key changes as MIDI events partly so I can reproduce it and
so I can tweak the MIDI events a bit. But I think you are right, because I am monitoring via software.


PostPosted: Fri Jan 18, 2008 8:58 pm
by Thirteen
I had a big talk with Dave Lackey about just this subject this week, and the system 100 came up, as one of Dave's friends was playing his and he couldn't get him off it, it was so immediate and punchy. His website has fascinating test results and graphs showing just how much delay sequencers and synthesizers with microprocessors, MIDI, scanning keyboards etc. have. Dave is really passionate about timing. He has a device called a Syncshift, which allows you to to grab a knob and slide the slave sequencer back and forth, even AHEAD of the master sequencer. You can simply slip the sequencer timing around in real time until it sits right in the pocket. I can't wait to have a play with one with Nuendo, an MPC and a Roland CSQ600.

Have a look at the Clock Watch page especially, and the Litmus page that lists results for specific machines, and how to do your own tests. I will post my results when I have had a chance to experiment with one. I have noticed often that slaving computers and sequencers together often leaves a vague timing result, where the groove just subtly loses itself, seeing the graphs that dave made made me realize why.

PostPosted: Fri Jan 18, 2008 11:07 pm
by rachelp
IThat's a very interesting site. I've seen Dave's sync shift on ebay a couple of times and didn't know whether I should get one or not.
I am using a Frostwave Quad CV, which has pretty good timing, but I think Logic has a lot to answer for. I need something like that,
that I can send MIDI clock to, that converts to DIN sync, but also gate pulses and keep it all in sync. I knew I wasn't imagining the
groove disappearing, it really sounds great the less I drive from the computer and then is really quite sloppy over MIDI.
I think I will have to do some more study and see if a synch shift will help!


PostPosted: Fri Jan 18, 2008 11:23 pm
by rachelp
Mmmm, the SyncShift does DIN synch, but not gate pulses. I'm using a channel on the Frostwave to send gate pulses that I setup in Logic
as a simple loop of 16th notes. That way I can program rhythmic patterns in the loop. The timing of this is OK - the MIDI lag seems to be a CV
problem. I think I need a mythical box that does what SyncShift does with DIN or Sync24, except for CV/Gate..... ...something that I can slide pieces
of music around with, after it has been through the MIDI2CV....


PostPosted: Sat Jan 19, 2008 7:46 am
by Thirteen
I have been thinking of how to use a Syncshift with the sequencer in my modular, I think that I will build a clock divider, it is a really simple circuit using a couple of chips (flip flops) that divide clock streams into halves, quaters eights etc. (The same circuit can make a sub-oscillator that you can easily build into any analog synth with a square wave output, the flip-flop chip divides the frequency by 2 and 4, giving you an octave and 2 octaves below, it is a dirt simple 1 chip way to get a sub, Roland did it all the time in their machines, from memory the SH101 being one of them).

PostPosted: Sat Jan 19, 2008 9:16 am
by heathen
Hey Rachel,

The way I get around latency is by using the spare outputs on my apogee rosetta 800, the spare output is for dual wire 96k and above recording, it mirrors the 8 inputs at 44.1 to both the adat outputs. I only use 44.1 so I patch the mirrored output into my O2R console adat channels 17-24, these are my main monitoring channels while tracking. Gives me zero latency, any digital mixer with adat ins should do this.

My computer soundcard is an rme hdsp24/56

Even with severe latency on playback the artist I'm tracking will be playing in time with the latency, even though there is massive latency going to disk also, there is none on thier direct performance monitor path from the microphone or instrument.

So on playback everything is perfectly in time and lines up.

Converters don't really have a perceivable latency it is usually software and soundcard drivers giving latency.

I think apogee are claiming zero latency with thier "Ensemble" system and Logic. Nic (aearth) on this forum has this rig and I'll be playing with it tomorrow, I'll post back on the latency if your interested.

So if your using only hardware synths and no plugins you should be able to get near Zero latency, but yep I know even near zero latency sux.

What type of soundcard are you using?

This is all just in regard to audio and monitoring, midi always has its own quirks on different DAWs. Also the intricacies associated with analog synths I'm not really familiar with either. But I have done lots of recording with digital outboard synths.

I'm sure a work around should be reasonably cheap and easy if dropping the latency does'nt help.

Monitoring your computer outputs and a split of the synth signal through a small be#$^&ger mixer could work. That will remove the software monitoring and give you a direct instrument monitor path. You won't be able to confidence monitor though but thats not such a big deal, you'll know if its not right on playback anyway.

PostPosted: Sun Jan 20, 2008 12:03 am
by rachelp
Argh! so much to digest! I am using a Ta@#$% 1804 which is a fairly budget interface, but it is OK. The latency is as low as I can get
which according to the device is 3ms. I think some of my problems are to do with software monitoring, so I am experimenting with that.

I tend to write all my parts into Logic, but arpeggiators, drum machines and so on are all triggered "live" by MIDI clock
and I like to feed them with source notes (to trigger the arp) and then I track everything back into Logic. I am gradually working around the
problem - didn't really notice it until I started with the System 100 like this! Part of my problem too, is that I am working alone
and I believe I am becoming hyper-critical, whereas if you have someone else working with you, you can jam your way out of it
or are moving very quickly to complete the music before you get bogged into hearing every little tiny flaw.


PostPosted: Sun Jan 20, 2008 1:29 pm
by harry
Hey Rachel,

unfortunately it has been noted that external sync on Logic has become progressively wrose since it left the PC at ver5.5..

I would attribute the problems your having to this.
I know of many hip hop MPC 3000 users that cant sync their samplers very well since V7 of the software

I think that sequencer manufacturers tend to think that we'd rather use their often crap sounding plug ins instead of real hardware....

maybe u want to run NOTATOR with UNITOR on an Atati ST - now that would be tight.....

PostPosted: Wed Jan 23, 2008 12:31 am
by rachelp
Hey, I solved some of my problems. I wanted to be able to trigger the CV note from Logic, so I didn't have to play the System from the
keyboard while I fleshed out the other parts. I found that the longer the Note Event I was sending the laggier the rhythm becomes.

So, since I only need enough CV to change the pitch of the Note, I shortened it down to pretty much a "click" that seems to set off the Pulse
with much better timing, then I concentrate on Note Duration using the Gate TIme on the 104 sequencer and using the envelopes. It is still
restrictive because I can't modulate Note Length in Logic much, without losing timing accuracy, but at least I can get the bassline rhythmically
in sync with the drum machine. I will experiment a bit more with things as I progress, but I am getting closer....

BTW, I do not know why stupid computers can't send a Realtime sync with such a simple protocol as MIDI. Maybe one day someone will
build an FPGA type device that can handle 10,000 MIDI Streams at once and Mux them out the back via a special port that then converts
to MIDI at the last nano-second?! I used to have an Atari way back - it was my first music computer and it was extremely tight. Then someone
stole it and I switched to a PC and now a Mac and things are no better than back in the 80's in spite of biggerfaster/cheaper/better.....
(pick any 2)....