Want to highlight a helpful answer? Upvote!

Did someone help you, or did an answer or User Tip resolve your issue? Upvote by selecting the upvote arrow. Your feedback helps others! Learn more about when to upvote >

Newsroom Update

The all-new Logic Pro for iPad and Mac delivers breakthrough music-making experiences. Learn more >

Looks like no one’s replied in a while. To start the conversation again, simply ask a new question.

Hexfix93 Author

Level 1

0 points

The daw war is over. midi timing! woooooo!

Midi jitter on external midi, sequencing synths and drums in real time from the computer. This is a big deal to those of us who like to sequence live and mix analog for better sound quality and tight notation. Logic 8 is the winner. Logic 8 with amt8 on osx is 1ms of midi jitter! this is as tight as notator on the atari st, which to this date held the record for tight midi on a computer. now logic 8 ties this record! KUDOS TO YOU APPLE AND EMAGIC! YOU HAVE WON THE MIDI WAR. AMT8 interface was used in my test. I triggered a high hat on my dr 880 drum machine, recorded it in, wrote down all the peak transients and subtracted the differences. yup 1 ms of jitter! so drums and bass lines can finally sound tight in real time thanks to apple and logic 8. I am so happy. i bought this software. logic 8 express! it has answered all my midi prayers! finally a daw that does it all, sounds great, and is feature laden. beats out all daws now. easily the king! i bet removing the dongle had a big impact of the midi timing. i cannot thank you enough, this has been a huge hassle for me over the last 10 years. i have pulled out hair, gone insane, gotten banned off the steinberg forums, and turned many against me because of this. its over. when i tested logic 7 with the amt8, i got 3 ms of jitter on osx. so logic 8 is even better!

PC MIDI: cubase sx with many interfaces including the midex8 on 4 different pcs i tested all were from 13 to 20 ms of midi jitter. Saw studio was 7 ms of midi jitter. Sonar 6 was 14 ms of midi jitter. Logic 5.5.1 with amt on xp(which was never stable) was 3ms of jitter but completely un usable because midi would just stop working, then you would have to reload logic and it would work for 5 minutes the quit again. so this is irrelevant.

Let me break down the daws for you in other ways. Saw has the best sound and worst interface with crap midi, sonar 6 has crap midi, second best sound and crap interface, cubase has crap sound and crap midi with an awesome interface. that is the pc side of things. on the Mac logic 7 had good midi, and decent sound, with an ok to decent interface. Now, logic 8 has a great interface, the best midi ever on a modern computer, and the sound quality is up there with sonar and samplitude. sonar had awesome time streching, now so does logic 8, cubase has awesome audio editing, but so does logic 8 now, 7 was no where close to cubase in audio editing, but now, its as good. finally logic 8 delivers on all fronts where all daws have failed before. LOGIC 8 express is 199! BUY IT NOW! Logic 8 studio is 499. BUY ONE OF THEM! cheaper than the competition, and way better in all regards! + NO MORE DONGLE!!!!!!! AWESOME!

taken from my web site here:

http://www.velvetacidchrist.com/gear.htm

macbook pro, Mac OS X (10.4.10)

Posted on Sep 20, 2007 6:38 PM

50 replies

iSchwartz

Level 5

4,321 points

Sep 23, 2007 4:16 PM in response to Targon

MIDI timing and MIDI jitter are two different things.

If you want to find out what the MIDI timing is like for an external sampler...

1. First, and most importantly, make sure that Logic's recording delay is properly calibrated.
2. See step 1
3. See steps 1 and 2

4. Load up your external sampler with a sound that has a very sharp transient, like a click or a closed HH that you're totally sure is tightly truncated to the attack of that sample.

5. Assign the sampler to a track and record a series of 1/4 notes (120 BPM should work fine). No need to record more than a few measure's worth. Quantize it to 1/4's (you know the drill...)

Also, edit the length of each note so that it's no more than an 8th note long.

6. Set your pre-roll to 1 measure and record your ext. sampler's audio output into a track in Logic. After you've done that...

7. Open the sample editor on that recording, set the view to samples (or milliseconds, whichever you prefer). Zoom way in and measure the distance from the bar/beat markings to the beginning of the start of the your click/HH sound. (If it's easier for you to measure to the peak, that's OK.) Write down the number of samples (or milliseconds) as displayed in the sample editor.

8. Move to the next beat and make the same measurement to the beginning of the next start point (or peak) of your HH track.

Repeat Steps 7 & 8 until you've measured at least 4 - 8 distances from beat marker to sample start (or peak).

9. Add up all of those numbers and divide by the number of measurements you've made. You now have the average MIDI timing response of your external sampler.

Targon

Level 1

55 points

Sep 23, 2007 7:47 PM in response to iSchwartz

Thanks dude, I already know all the mechanics of Midi. I just wanted to know if there was a difference in Timing and Jitter in Logic when Audio is enabled or completely disabled. In CubaseSX the Midi engine is dependent on the Audio driver an its buffer settings. If a higher buffer for instance is set, Midi playback is significantly impaired (delayed and ropey [considerable variance]). Of course Audio cannot be disabled in CubaseSX thus Midi is at the mercy of an Audio driver even tho a user may not even want to use any audio in the app.

So, if you playback a 4 to the floor beat on a external sampler or synth with Audio Disabled in Logic can you detect any difference compared to having Audio Enabled in Logic?

Thanks

Targon

Level 1

55 points

Oct 7, 2007 4:15 AM in response to iSchwartz

iSchwartz.

I followed your procedure above which resulted in somewhat as surprisingly positive figures. I slightly modified your method however. In step 5 (since i have Audio Disabled in Logic) i decided to record the analog out of my sampler into another Mac running Pro Tools. I find audio manipulation in PT be be far more adaptable to my style. I just find any other app impossible to do audio editing for even the most simple of tasks especially for linning up a transient to a beat mark as in this case. Anyways i had a few mixed results as i found at different tempo's the audio was consistently early by at least 300 samples to the beat. After playing with the Inspector "delay" track parameter in Logic i found some more consistent results.

I used my Powerbook G4 1.67ghz 2 gig ram Logic8 Pro, Audio disabled & AMT8 using drivers from the Logic 8 Install disk. The sampler was an E-MU E4XT (Non Ultra) for the results below. I did test my other E-MU, an E4XT Ultra, I observed slightly more accurate but no results posted here for the Ultra.

Below i table a few variations of my evaluation.

Average samples @ 174bpm @44.1k
4,2,11,-3,6,0,0,6=32/8=4 (0.09ms)
15,0,-1,1,3,9,19,0=48/8=6 (0.14ms)
32+48=80/16=5 (0.11ms)

Average samples @ 120bpm @44.1k
0,12,-20,-18,14,-19,-16,15=114/8=14.25 (0.32ms)

Average samples @ 120bpm @ 44.1k
(+1 tick delay on track)
16,0,1,18,18,19,19,3=94/8=11.75 (0.27ms)

Average samples @ 145bpm @ 44.1k
10,18,27,16,5,11,33,6=126/8=15.75 (0.36ms)

Average samples @ 145bpm @ 44.1k
(-1 tick delay on track)
22,3,0,3,5,14,3,3=53/8=6.625 (0.15ms)

In contrast to an old recording made with my Kurzweil K2500 and Atari Cubase.

Cubase Atari & K2500
Average samples @ 174bpm @ 44.1k
39,36,15,20,68,73,36,84=371/8=46.375 (1.05ms)

You may very well be aware of Logic's Midi capability. But all the same im very surprised at these results, and for me, this is incredibly tight which totally satisfies me. Im totally sold now to acquire 3 more AMT8's for the rest of my gear.

Thanks for your input into this thread iSchwartz and dispelling alot of the BS which has been posted here.

iSchwartz

Level 5

4,321 points

Oct 7, 2007 11:03 AM in response to Targon

Hi Targon,

Thanks for posting this. Will reply after I've had a few cups of coffee.

Just one thing tho... curious to know how you came up with the consistent number of 300 samples... did you synchronize Logic and PT somehow?

Targon

Level 1

55 points

Oct 7, 2007 1:46 PM in response to iSchwartz

Hey iSchwartz,

I should clarify, that point. Not exactly consistently 300, there was a slight waver, 303, 310, but max of 325. What meant to say is that it was running close to 300 or so samples early lowest i saw was 297 to 325. Absolutely no sync to PT. Im not sure exactly what the issue was there, once i adjusted the Track Delay parameter after a few times it hooked up close to 0 with some minor jitter as seen in the results above again between -3 to + 18 or there abouts. Interestingly before i decided to mess with the track delay parameter i decided to perform record the identical preset from my Emu Ultra, that device played slightly late but not near 300 samples like the other EMU first time without any Track Delay setting in Logic. After a bit i managed to set an -1 tick for the ultra and +1 tick delay on the Non ultra an both machines fired close to band on zero for several beats.

I really should have also triggered more than one note on a single channel and routed each voice out a different output and analyzed.

Anywayz absolutely no sync to PT audio straight in as if going to DAT.

Stay cool

iSchwartz

Level 5

4,321 points

Oct 7, 2007 7:00 PM in response to Targon

Hi Targon,

Thanks for clarifying, but I'm still not sure I understanding something here...

...i found at different tempo's the audio was consistently early by at least 300 samples to the beat...

If you were recording "free" into PT, how could you judge that the audio was recorded early? "Compared to what?" In other words, the first beat you recorded would have appeared at a random point in PT's tempo grid.

I see that you've been really methodical with your test, so maybe I'm just missing something here...

Targon

Level 1

55 points

Oct 8, 2007 3:39 AM in response to iSchwartz

I lined up the transient of the first beat, every beat there after (exactly the same transient sample point) was 300+ samples before the beat grid in PT. I think i may have an explanation for this. This was my first run, In retrospect i now remember I carelessly did not set up a pre roll in Logic. I hit play from bar 1 beat 1. This may explain the reason why I observed this on this single occasion. Perhaps there was some factor causing the first beat to be offset at initial playback and consequent beats after the first were realigned in time to where they should have been (hence the consistent 300 samples observation), I can see how this may have skewed the timeline. You did after all specifically say to pre roll a few bars, this is likely the reason why.

Sorry to have misrepresented this single point I mentioned. However, be assured, the figures I posted earlier were indeed derived using a 4 bar pre roll so the accuracy here is legit for those statistics. I actually took me a few minutes to figure out how to actually pre roll in Logic. I couldn't find a preference anywhere for this, only for record. In the end i decided to drag that little block at the start of the sequence in the time line to the left an noticed the transport start location go into negative figures. Oddly the other block at to the right (sequence length position?) appeared not to move according to the snap setting, which was set to Bar and it also did not adhere to the Absolute Position setting found also in Snap in the arrange. Oh well.

Hope this clears things up.

iSchwartz

Level 5

4,321 points

Oct 8, 2007 10:49 AM in response to Targon

Hi Targon,

Logic has always been funky when it comes to playing the first measure of MIDI perfectly in time. I didn't explain that aspect very clearly in my "how to" post (my bad). But what I was implying was that you should "give Logic 1 measure to get its **** together". So in your case, where you're recording audio to another system, you should let Logic play for 1 measure before your MIDI region starts (start playback on bar 1, have your first region appear on bar 2).

Oct 30, 2007 9:49 PM in response to iSchwartz

Hi,
After reading this thread I went and hunted down an AMT8 on eBay but does anyone know where I can get the latest drivers?
I found a driver on version tracker that actually seemed to work ... the amt8 showed up in audio midi setup, However, when I launched Logic I got a midi conflict error. I then found the 1.71 driver on the Apple site but now I cannot see the driver in audio midi setup or Logic.
Anyone?
Thanks,
grimley

Oct 30, 2007 10:12 PM in response to Armachian

After some more digging I discovered that I need version 1.9 of the drivers but Apple only have version 1.7 on their site. Would anyone be so kind as to email it to me?
Thanks!

ifwuh

Level 2

445 points

Oct 30, 2007 11:55 PM in response to Armachian

Do you mean the driver called "EmagicUSBMIDIDriver.plugin" ?
My experience is that the Logic 7.2.3 (Intel UB) install disk can install this driver.
It is version 1.11 and works pretty good with an AMT8.

Oct 31, 2007 12:40 AM in response to ifwuh

That's probably it. I'm new to Logic though so I don't have the Logic 7 CD. I'd really appreciate if someone would email this to me.
-grimley

ifwuh

Level 2

445 points

Oct 31, 2007 12:52 AM in response to Armachian

ok, you need to post your email or include your email in your profile settings of this forum
that way someone here can email you, you know?

Oct 31, 2007 9:42 AM in response to ifwuh

Ooops! My email address was hidden in my profile. Fixed! Thanks.

Nov 14, 2007 3:48 PM in response to Armachian

hi -

i was really interested to hear that Logic 8 had better timing than 7 ... but it would be nice to see the results for a "midi loopback test". I tried this using Logic 7 and an Amt8 midi interface ... outputting two bars of sixteenth notes from one port, and re-recording through another ... the results being measured in ticks (which isn't as good as milliseconds, as it varies with tempo)...BUT...it is alot easier than mucking around with a sampler editor if you want a quick preliminary test ...

So... my initial results depended on the direction of the wind at the time, and possibly the day of the week? At bpms of around 160 the accuracy was + or - 6 ticks, at bpm 120 it was + or - 3 ticks. Sometimes it performed the whole task perfectly, sometimes every note was late by one tick, mostly things were late between 1 and 3 ticks...

Anyway, it would be nice to hear from anyone who has tried Logic 8 using a loopback test (you would probably have to do it a dozen times to see the extent of the variation)...

Personally, for me, at 160... 6 ticks late, followed by six ticks early ... is too much for the relative placing of two notes ... i'm sure many people would think this was fine? ... for those that don't ... i would love to hear about the current state of affairs on logic8? ...

The daw war is over. midi timing! woooooo!