Midi accuracy, midi jitter.

[jksuperstar]

Windows will have delays, but in my experience the jitter should be less. Adjust tracks to account for the delays (if you need to), and that's it. Also, since your clock and MIDI data are now both coming from an external source, that delay may be the same for both.

Recording MIDI straight into the QY-70 is definitely an option, and you should get very accurate 480ppqn there.

[lagoausente]

Windows will have delays, but in my experience the jitter should be less. Adjust tracks to account for the delays (if you need to), and that's it. Also, since your clock and MIDI data are now both coming from an external source, that delay may be the same for both.

Recording MIDI straight into the QY-70 is definitely an option, and you should get very accurate 480ppqn there.

mmm... but the midi driver will continue using the mobo clock?
In the other hand, with a tempo os 240 bpm will have double resolution than with 120 bpm isn't? If I undestood ok 480 ppq is relative to tempo.

[jksuperstar]

Yeah, ppq(n) = Pulse per Quarter Note, it is related directly to tempo, so higher tempo means more clocks per second. Hardware still transmits 24ppqn, always. No matter the resolution of the source or the destination. The jitter on that 24ppqn is far less on the QY vs native.

The midi captured will get a timestamp from the midi driver, but the clock and data will come from hardware. You can experiment with using the QY during recording, or only during playback. I can isolate and hear the improvement when I quantize a recording, and then play back with native clock vs. QY-70 clock. If your going to record natively, you can set up the QY to be your metronome, so your playing is locked to it.

I use my MOTU MIDI Express to get clock from the QY, and send it to Ableton and to SCOPE simultaneously.

[dawman]

My ancient QX-1 uses 384ppqn and I have tried many different hardware sequencers over the years like the QY300, MC500-MKII, and Roland MV30.
For some reason the hardware sequencers w/ 96 or 384 allowed me to capture and quantize with feel much better than other DAW's and hardware units.
I can't explain why, but that's why I find it hard to part with a QY300 or QX-1 for music.
It's also probably because I have used these units since 1984....

If someone could just take the information from the minaiture LCD like Roland hardware samplers using the RGB outs did, I would be happy.
I can't believe the LCD and everything still works. I did have to replace the powered mains cable 5 years ago, but this Dog still Hunts.
They sure don't make hardware like they use to.

EDIT: Just tried using a DAW again for realtime MIDI Control of Kontakt 4.2 and decided to go back to hardware.
The reasons are simple, latency from the host compared to Standalone.

My current experiment will be syncing VDAT to playback audio since I still hear it's slight advantage over Cubase and Reaper, and hardware for the MIDI.
I have tried using my controllers and a host simultaneously to control MIDI and I just don't like the feel and timing.
Having the QX-1 as the MIDI playback device, mergerd through my DMC-MX8 with the realtime hardware surfaces suits me better.

It seems these apps just keep getting more and more gunk added to them.
Even Bidule is acting different now that it's x64. It asks me to activate my Kontakt instruments now, so my decision is based on performance, not conveniences. These apps are so worried about " workflow " and convenience that the pipes are all getting clogged up.

[spacef]

It seems these apps just keep getting more and more gunk added to them.

This is true, you can't beleive how hard it is for some developpers to stay focussed on the core functions while hundreds of users want to add this and that, and may be also that and this too....
May be one of the reason is that those "core features" are not the same for each user, and so in a attempt to satisfy the most, a lot of functions are added in the "basic" versions of the same soft.... It is something I regret sometimes, because adding a few features is often at the expense of improving what is already there.... On some other forums and mailing lists, I am feeling really alone sometimes because I often add "-1" for feature requests that seem to me as gadgets, but other users with deeper use of the same soft think they are essential (they must be right, for them, apparently). The struggle continues
But sometimes, I also see something i fought against, become one of my favorite feature . So all in all.... it's often a progress... and a cause of upgrading the computer... (arrg).
It is also cool to satisfy users' requests sometimes and see them happy with a tool that is almost "customized" for them that's also the job of being a developper (on scope or elsewhere, i wasn't talking about scope in fact). Resistance is futile It's one of the beauty and danger of software compared to hardware... software are easily updatable and easily available to all users for a small fraction of what it would cost to make and buy a new hardware...

I am pretty sure your hardware sequencer sounds great because it is not as precise/cold as a DAW while being very precise anyways ?

[Bud Weiser]

For what I have reading 480 ppq is good enough, but I have a doubt. If sync Cubase to QY70 midi clock, do that mean the driver will take the hardware clock instead of the Mobo?

I´ve read jksuperstar´s post above too and I´m not sure if it works like this.
If you use hardware transmitting MIDI clock to another sequencer or DAW, the MIDI clock does nothing else than positioning the 2nd sequencer/DAW to beats or fractions of beats.
Outgoing MIDI clock from a piece of hardware is always 24ppq (standard) which is much less resolution than any DAW offers internally.
http://en.wikipedia.org/wiki/MIDI_beat_clock

If you look at Notator/Creator from the past, you could switch it from 192ppq to 384ppq resolution after the Yammi QX-1 came out and it´s just only interpolation and outgoing MIDI clock was still 24ppq.

MC isn´t MTC which would be a better option.
http://en.wikipedia.org/wiki/MIDI_timecode

I don´t believe you really get better MIDI timing by feeding a external MIDI clock to a DAW because there´s something in between,- 1st the 3rd party hardware MIDI interface connected via USB, data running thru 3rd party MIDI driver and p.ex. DirectMusic Windows MIDI drivers or the emulated ones which is more worst.

In a DAW, incoming MIDI clock is translated to the clock the software is slaved to, and that´s QPC or TGT.
Incoming Midi clock tells the DAW program "now run from the beginning and at tempo X until the external MIDI clock stops" and the GUI of the DAW software runs in sync (visible) and tries to fire audio and MIDI as accurate as possible.
All you set up in the sync preferences is related to one of these clocks delivered by the motherboard.
Is all relative and that rules also for the time stamping of incoming and outgoing MIDI events.

It´s all system and your computers hardware dependend and you can have luck or not,- and who guarantees the front end MIDI clock from hardware, whichever it is, doesn´t introduce jitter already ?

Bud

[spacef]

ok, cool to know... That's why I was disapointed when i recorded the atari midi into cubase, slaving cubase to atari, it justs dimished the precision.... I tried without sync , but it didn't really work because clocks from various soft or computer are slightly different, so requantizing was difficult (i ended up recording a few stuff loop by loop, without sync, and rebuilding the projects...)..

a hardware clock can improve the stability of an incoming clock into scope though (i tested once with a basic clock from a yamaha toy electro drums 99€ or less, but again, it can have a slight difference (like 120.01 bpm instead of 120.00 for example...so i don't know if that's actually better... but i tried with a toy so...).

[Bud Weiser]

ok, cool to know... That's why I was disapointed when i recorded the atari midi into cubase, slaving cubase to atari, it justs dimished the precision...
a hardware clock can improve the stability of an incoming clock into scope though .

You mean, if you record into Scope VDAT (because you cannot record MIDI into Scope) ?
Or do you mean triggering Scope´s instruments w/ external MIDI ?
Or recording MIDI thru Scope into a native DAW runningh simultaneously on the same machine ?

Scope is another story,- I assume Scope generates it´s clock on it´s card and that´s why Scope´s MIDI timing is dead on and the whole thing is extremely low latency (if there´s latency at all) in addition if I use Scope instruments for realtime playing being triggered by external keyboards.

But the problems I described above in former posts come in again if you record MIDI over the Scope cards MIDI In thru it´s MIDI I/O devices into a sequencer running natively on the DAW.

Have also in mind MIDI and audio is a different handling in regards of syncing a DAW externaly.
Let´s say you have a tape MTR running, streaming SMPTE to an external synchronizer,- you have to connect MIDI Out from synchronizer to MIDI In of your DAW´s MIDI interface and adjust sync parameters in the DAW software´s preferences for MIDI as also feed a wordclock signal from the synchronizer to the DAW´s soundcard.
That´s why the audio is "slaved" to the wordclock and the position of the sequencer is slaved to MIDI Clock or MIDI Time Code.

The synchronizer itself affects and defines the quality of sync signals by it´s design, built quality and precision.
In most cases a clock is generated by a reference quartz (crystal) and the related electronics circuitry and this is also the case w/ computer motherboards.
If a consumer electronics device like a computer MoBo is designed to deliver extremely stable and precise clocks is hard to find out, if not impossible, for a customer.

For the more or less rare cases I have to sync my DAW to re-record tracks from old tapes w/ my tape MTR, I´m using a Steinberg Timelock Pro (which is a Rosendahl) which works perfectly together w/ my Cubase SX3 and Midex 8.
Well, that´s only a 32Bit OS solution and it´s limited to 16 / 24Bit and 44.1 or 48KHz SR, but that´s by far good enough for such tasks. (It also supports Digidesign Superclock b.t.w.)

To re-record (not import) MIDI or to trigger VSTis in a DAW running natively, you´re addicted to the process using a MIDI cable between 2 devices and it is like it is,- depending on your gear being set up.

Bud

[t_tangent]

Thanks for the very interesting info chaps. While hopefully keeping on topic I thought the following might be of interest. Related to Cubendo and certain motherboards & their MIDI timing, there was an interesting site that a chap called Jay Levitt wrote about this. Unfortunately the site doesnt exist now but I managed to find the info copied on another forum here http://www.ultimatemetal.com/forum/equi ... ost6315034 and although some of the data and motherboards are old hat now, but it still might be useful info. There was also another site that originally linked to this and had some very useful and related info http://miditest.earthvegaconnection.com/

======================================

[lagoausente]

EDIT: Just tried using a DAW again for realtime MIDI Control of Kontakt 4.2 and decided to go back to hardware.
The reasons are simple, latency from the host compared to Standalone.

My current experiment will be syncing VDAT to playback audio since I still hear it's slight advantage over Cubase and Reaper, and hardware for the MIDI.
I have tried using my controllers and a host simultaneously to control MIDI and I just don't like the feel and timing.
Having the QX-1 as the MIDI playback device, mergerd through my DMC-MX8 with the realtime hardware surfaces suits me better.

It seems these apps just keep getting more and more gunk added to them.
Even Bidule is acting different now that it's x64. It asks me to activate my Kontakt instruments now, so my decision is based on performance, not conveniences. These apps are so worried about " workflow " and convenience that the pipes are all getting clogged up.

I really didn´t understand much what you mean.
I also think that samplers as standalone perform better than inside Cubase. Do you mean that? Do you mean that you prerfer use Kontakt in standalone and record directly to Vdat instead of using Cubase/reaper and record inside de sequencer?

[lagoausente]

Thanks for the very interesting info chaps. While hopefully keeping on topic I thought the following might be of interest. Related to Cubendo and certain motherboards & their MIDI timing, there was an interesting site that a chap called Jay Levitt wrote about this. Unfortunately the site doesnt exist now but I managed to find the info copied on another forum here http://www.ultimatemetal.com/forum/equi ... ost6315034 and although some of the data and motherboards are old hat now, but it still might be useful info. There was also another site that originally linked to this and had some very useful and related info http://miditest.earthvegaconnection.com/

Thanks for the links, I think that are the info I attached on another thread on this forum some time ago. That´s what I refer to.

I´ve read jksuperstar´s post above too and I´m not sure if it works like this.

What I don´t know if this F8 data is coming into the midi signal from the external hardware sequencer. I´m not sure how the midi driver works then... If the midi signal from the external hardware sequencer comes with that F8 data, latency on windows will not take effect, since position should be determined by F8 data. But still I´m not sure if the midi driver will take that data in account or not.
In the worst case, and in relation what XITE-1/4LIVE tells (seems I´m looking for the same thing), recording externally on the QY70 could be a good choice. If any drops occur on a good take, I´ll have the midi backup up on the PY70. So I suppose I could import the midi files (I hope, haven´t read the manual yet) so should be a good midi backup without troubles.
But for playing back, I should record the wavs through VDAT if use standalone, or inside the sequencer (have to check that yet). The bad thing on VDAT is that waste a lot of hdd space even if you have tracks empty.
http://www.harfesoft.de/aixphysik/sound ... chor477082

[spacef]

>>>>>>>>>
You mean, if you record into Scope VDAT (because you cannot record MIDI into Scope) ?
Or do you mean triggering Scope´s instruments w/ external MIDI ?
Or recording MIDI thru Scope into a native DAW runningh simultaneously on the same machine ?
>>>>>>

About the clock that becomes stable (in fact, only a bit more stable) with external hardware clock, I was referring to the option to sync the tempo of tempo-based scope-plugins through the plugin's midi inputs , typically, to sync LFOs, when you choose "Ext" as a clock source on synths or Modular for example (which is known to give bad results if the clock source is cubase and many other software, like if the tempo was modulated by an LFO).

About the recording of atari to cubase, I don't remember at all if I used direct midi connections or if I used a small sync box in between, nor if it was scope's midi i/os or another interface.....

[jksuperstar]

In the original post, there was some blur between recording, playback, jitter, and latencies, so sorry for any confusion I added to the mix. In my original response, I tried to be specific about when mating external gear with internal sequenced stuff, usign an external hardware based clock tends to give best results.

When recording notes played by hand natively, they are asynchronous events, and will always get a system timestamp. The notes you play by hand don't really have anything to do with the MIDI clock or tempo. The sequencer used will then translate that system timestamp on your midi note to something that relates to the song, which can then involve tempo, and moving things around to match some level of PPQN resolution (960, 480, or whatever). The hardware clock can effect this translation, as the host calculates/interpolates where the sub-ppqn beats are. Whether that turns out better than whats possible natively is up to your software, hardware, drivers, mobo, yada yada. I can't speak for what exactly Cubase does here. If your tracks still have jitter that destroys the feel you're after, and you don't want to use quantization to fix the problem after the track is laid down, than using hardware to record those midi events might be best. Then transfer the MIDI to your sequencer as you mentioned (I have done this, but tend to quantize due to the style of music I make). For the QY-70, you'll need a serial port to do the transfer, or a USB>RS-232 device for most recent motherboards don't include RS-232 UARTS anymore. [Hmmmmm, this sounds like a great utility for MAX for Live to do automagically if I ever need it, record, transfer, and drop into a slot without all the hassle.]

For playback and real-time effects, however, the time those notes and tempo and clocks are played is definitely driven by the clock. Using a hardware clock in my experience will definitely give a lower jitter playback. Like Spacef commented, this is easy to hear in LFOs and other beat locked/repetitious patterns where you send the midi to an external piece of gear.

[Fluxpod]

I would buy one of these.http://www.innerclocksystems.com/New%20 ... -Lock.html Instead of some 80 tees sequencer.

[lagoausente]

In the original post, there was some blur between recording, playback, jitter, and latencies, so sorry for any confusion I added to the mix. In my original response, I tried to be specific about when mating external gear with internal sequenced stuff, usign an external hardware based clock tends to give best results.

When recording notes played by hand natively, they are asynchronous events, and will always get a system timestamp. The notes you play by hand don't really have anything to do with the MIDI clock or tempo. The sequencer used will then translate that system timestamp on your midi note to something that relates to the song, which can then involve tempo, and moving things around to match some level of PPQN resolution (960, 480, or whatever). The hardware clock can effect this translation, as the host calculates/interpolates where the sub-ppqn beats are. Whether that turns out better than whats possible natively is up to your software, hardware, drivers, mobo, yada yada. I can't speak for what exactly Cubase does here. If your tracks still have jitter that destroys the feel you're after, and you don't want to use quantization to fix the problem after the track is laid down, than using hardware to record those midi events might be best. Then transfer the MIDI to your sequencer as you mentioned (I have done this, but tend to quantize due to the style of music I make). For the QY-70, you'll need a serial port to do the transfer, or a USB>RS-232 device for most recent motherboards don't include RS-232 UARTS anymore. [Hmmmmm, this sounds like a great utility for MAX for Live to do automagically if I ever need it, record, transfer, and drop into a slot without all the hassle.]

For playback and real-time effects, however, the time those notes and tempo and clocks are played is definitely driven by the clock. Using a hardware clock in my experience will definitely give a lower jitter playback. Like Spacef commented, this is easy to hear in LFOs and other beat locked/repetitious patterns where you send the midi to an external piece of gear.

Nice, I´ll have to check what´s Max for Live..
Your explanation here helped a lot. The reason for recording midi is for example using BFD drums, to mix can go with lot of individual channels (overheads, direct...) or even change the snare, kick.. etc.
So for everyday work it´s enough just to record Stand-alone to VDat in Sterero, or 1 channel for each piece (eight for example) or doing the same with reaper that seems to allow record the Vst channel output directly to disk. Even jitter exists is much lower then midi. I confess I haven´t check midi jitter with my new laptop but I just don´t want waste time really. The jitter of the VST audio output is some somples, less than 1ms so nothing to worry about.
Having a midi backup on Qy-70 should allow to can rescue a good midi data if the "take" or the song worth it.., most of times will be jamming, playing... trying. For that is good record audio VST output or Standalone output. For a mix (if finally I can make a decent song) can be useful that midi from the QY-70, so I think I made a good purchase.
I really don´t use synths, LFO.., yet, I don´t have the time yet to learn how to use tem in depth, so I really will not use the clock, maybe only just for sync the transport between Qy70 and the recorder.

[dawman]

Fluxpod My Brotha'............
That's awesome stuff.
Thanks...

I would love to replace the old hardware sequencers, sooner or later they are going to croak.
26 years though is a testament to what vintage hardware was worth.
Never had to upgrade once.....I like that.

Great Link.

[Fluxpod]

Fluxpod My Brotha'............
That's awesome stuff.
Thanks...

I would love to replace the old hardware sequencers, sooner or later they are going to croak.
26 years though is a testament to what vintage hardware was worth.
Never had to upgrade once.....I like that.

Great Link.

Hey Jimmy.26 years is very good for any piece of hardware! I would like it if more things would work that long to not pile up electronic waste.
About that Sync-lock,i have seen it in action and it is one of those things that really just frikkin work.Not cheap for a 1 function "simple" sync device.You get what you pay for.And i bet that will live for another 26 years at least.

[lagoausente]

Fluxpod My Brotha'............
That's awesome stuff.
Thanks...

I would love to replace the old hardware sequencers, sooner or later they are going to croak.
26 years though is a testament to what vintage hardware was worth.
Never had to upgrade once.....I like that.

Great Link.

Hey Jimmy.26 years is very good for any piece of hardware! I would like it if more things would work that long to not pile up electronic waste.
About that Sync-lock,i have seen it in action and it is one of those things that really just frikkin work.Not cheap for a 1 function "simple" sync device.You get what you pay for.And i bet that will live for another 26 years at least.

Yes, that device looks very interesting. I haven´t read yet how it works in depth. Even I will not buy that for now, for curious in relation with the issue of recording the midi tracks inside Native-sofware it´s supposed will not help, probably could help syncing a hardware sequencer with the native for audio tracks. I read there that uses audio for sync, how is that?

[Bud Weiser]

Thanks for the very interesting info chaps. While hopefully keeping on topic I thought the following might be of interest. Related to Cubendo and certain motherboards & their MIDI timing, there was an interesting site that a chap called Jay Levitt wrote about this.

Yep,- I know the site,- I was in contact w/ Mr. Levit years ago because of MIDI timing issues.

Unfortunately the site doesnt exist now but I managed to find the info copied on another forum here http://www.ultimatemetal.com/forum/equi ... ost6315034 and although some of the data and motherboards are old hat now, but it still might be useful info. There was also another site that originally linked to this and had some very useful and related info http://miditest.earthvegaconnection.com/

In act, that´s the man who coded the excellent DirectMusic MIDI driver for the Opcode 8portSE 8x8 MIDI interface.
AFAIK Steini was in contact w/ him too and after that, the latest driver for Midex-8 came out (v1.92 I think) which is the best ever made for Midex-8.

Bud

[xite knight]

in relation with the issue of recording the midi tracks inside Native-sofware it´s supposed will not help, probably could help syncing a hardware sequencer with the native

I think you're right