Fine Tuning I/O ....observations in DP 9.5

Discussion of Digital Performer use, optimization, tips and techniques on MacOS.

Moderator: James Steele

Forum rules
This forum is for most discussion related to the use and optimization of Digital Performer [MacOS] and plug-ins as well as tips and techniques. It is NOT for troubleshooting technical issues, complaints, feature requests, or "Comparative DAW 101."
Post Reply
User avatar
Maxxy
Posts: 156
Joined: Tue Jul 31, 2012 1:32 am
Primary DAW OS: MacOS
Location: Bali, Indonesia

Fine Tuning I/O ....observations in DP 9.5

Post by Maxxy »

Hi

I’ve just been testing Fine Tuning I/O ….yes that old chestnut…..with DP 9.5 and had some interesting results

In my case, using a MOTU 16A interface and an outboard analog synth Korg MS-20 Desktop, I may have found a fast and accurate way to quickly calculate the Recording Offset needed for any Project Sample Rate with any Buffer Size …. ( N.B. All testing was done with 0 Playback Offset)

Divide the sample rate by 100 and add that to the Buffer Size

So, for example, if you are recording at 48khz at 128 sample buffer you would go: 48,000 divided by 100 = 480. Now add 128 = 608 … your (my) Recording Offset is 608

Try a different scenario: 44.1khz sample rate at 256 buffer = 441 + 256 = 697 Rec Offset

Or 176.4khz at 1024 Buffer = 2,788 Rec Offset

192k at 128 sample buffer = 2,048 Rec Offset

96khz at 128 Buffer =1088 Rec Offset

I don’t know how or if this translates to anyone else’s system… but for me it is uncannily accurate…in fact on any combination of Sample Rate and Buffer settings I don’t think I could make it more accurate! ( :

I just thought it might be a useful observation that might help others when recording external audio back into DP (such as from a hardware synth) that is being triggered from MIDI in DP.

Setting the Fine Tuning I/O correctly and getting maximum accuracy is really worthwhile …For right now, at least, this quick calculation will save me having to pull out a document which tables each and every Sample Rate /Buffer Size setting combination… IF this theory holds up over time ( : … I hope so

Being able to accurately and simply set the Fine Tuning Offset should make it easier to use outboard MIDI hardware under conditions where the CPU is being taxed and low buffer settings become unwieldy…. Note that under higher buffer settings the audio sounding from the synth and being heard will get later and later against the sequenced audio (Playback Offset is the other chapter in the Fine Tuning I/O story)…. but, most importantly, the synth’s audio will get recorded into DP precisely

Also note this is nothing to do with ‘playing along with the music’ in time (see Sync Recorded MIDI to Patch Thru) ….it is, again, only applicable when sequenced MIDI in DP is triggering external hardware that needs to be recorded as audio back into DP...

I may try this tomorrow and find it was just a dream…or a fluke … but I hope the dream works for you too

Best bleeats to all
Mac mini 2023 M2 Pro, 32g RAM, LG 32UD99 4k monitor on HDMI, OSX 13.6.1, DP 11.3, MOTU TB AVB 16A w/AVB Monitor 8 on ethernet cable, UAD Apollo x8, Netstor TB3 PCIe card expansion w/ UAD-2... 1 x Octo and 2 x Quad cards, Softube Console 1 MkII & Fader Control
User avatar
Phil O
Posts: 7231
Joined: Thu Jul 28, 2005 10:01 pm
Primary DAW OS: MacOS
Location: Scituate, MA

Re: Fine Tuning I/O ....observations in DP 9.5

Post by Phil O »

Thanks Maxxy. I'll give this a try the next time I'm using an external keyboard for sounds (which I do from time to time). I usually just nudge the soundbite after recording it, but if this works it will be a better solution. I should have looked into I/O timing years ago, but that's a feature I just have never explored.

Phil
DP 11.23, 2020 M1 Mac Mini [9,1] (16 Gig RAM), Mac Pro 3GHz 8 core [6,1] (16 Gig RAM), OS 14.3.1/11.6.2, Lynx Aurora (n) 8tb, MOTU 8pre-es, MOTU M6, MOTU 828, Apogee Rosetta 800, UAD-2 Satellite, a truckload of outboard gear and plug-ins, and a partridge in a pear tree.
User avatar
Maxxy
Posts: 156
Joined: Tue Jul 31, 2012 1:32 am
Primary DAW OS: MacOS
Location: Bali, Indonesia

Re: Fine Tuning I/O ....observations in DP 9.5

Post by Maxxy »

Thanks Phil

Having done some more on this I want to offer some more findings…. again I hope it may be helpful… it has been for me

I realise the ‘formula’ I spoke of seems, on the face of it, to be possibly nonsensical… why should this give reliable offset figures? I really don’t know why….but it does seem to hold up for me …with my system and using this particular outboard synth. I haven’t tested any other synths but I think that any individual variations of synth response time, whether faster or slower responding, may just require a slight offset in order to get the correct figure. This may beg the question then of why it wouldn’t just be better to find the right MIDI offset for each device and apply a MIDI plug in Time Shift

Also one could simply nudge the recorded audio into place as Phil does….this is how I have normally done it myself

But it’d be great if one could set DP up so that both DP-triggered MIDI outboard and DP audio tracks processed outboard could be recorded back into DP with great precision. Surely this is why FIne Tuning I/O exists

I decided to try again to understand it better in the context of the new DP 9.5

I hadn’t re-read the FIne Tuning I/O section in the Getting Started manual before I started …which was fortunate as it states there that one should begin with setting Playback Offset. Magic Dave has updated this information in "external MIDI instrument timing" thread where he explained that one should start with the Recording Offset to set the MIDI timing….and after that set the Playback Offset.

The problem for me is that the Playback Offset just cannot compensate for the amount the audio Out > In needs to be moved later after having been shifted so much earlier by the correct (for MIDI) Recording Offset. This is in the case of routing the audio track to an output of the interface and then routed straight back in ie. eg. Audio Track out the MOTU 16A Output #11-12 and directly patched back into the 16A’s Audio Input #12. In DP 9.5 this cannot be made to work for me… the closest I can get the audio back to square is about 403 samples ahead of the beat

However if you ignore the Playback Offset section altogether, leaving it set at 0, and instead use a Hardware Insert on the audio track routed in the same way, you can then bus that audio via a send to the input of the track you are recording to ….It works ‘perfectly’. Hardware Insert is working fantastically well now and the audio will be recorded back into DP precisely! …..Great! It works equally well being bussed to an Aux eg.for parallel compression

The other problem with the methods suggested is that the Recording Offset needs to be changed for every combination of Sample Rate and Buffer Size setting! …. what a drag….no wonder it is normally easier to just shift the recorded audio into place after the fact

I can now get both the MDI and the audio back into DP precisely and simultaneously … and with minimum fuss. Both MIDI and audio gets recorded into position consistently within about 6 or 7 samples of accuracy…very often within 2 or 3 samples… at 6 samples that is 1/10th of a millisecond!

It may be necessary sometimes to re-OK the Fine Tune I/O after opening and setting the Hardware Insert

I could post a couple of screen snaps in a day or so

In the context of a well set up FTIO it might be possible to just use MIDI Time Shift to finely adjust individual outboard MIDI gear which may trigger more slowly or quickly than the synth used for the basic Fine Tune I/O setup…. again, I haven’t tried any other synths yet

I am using the 16A on a Mac Pro 5,1 with USB connection
Mac mini 2023 M2 Pro, 32g RAM, LG 32UD99 4k monitor on HDMI, OSX 13.6.1, DP 11.3, MOTU TB AVB 16A w/AVB Monitor 8 on ethernet cable, UAD Apollo x8, Netstor TB3 PCIe card expansion w/ UAD-2... 1 x Octo and 2 x Quad cards, Softube Console 1 MkII & Fader Control
User avatar
HCMarkus
Posts: 9712
Joined: Tue Jan 10, 2006 9:01 am
Primary DAW OS: MacOS
Location: Rancho Bohemia, California
Contact:

Re: Fine Tuning I/O ....observations in DP 9.5

Post by HCMarkus »

Maxxy wrote:I am using the 16A on a Mac Pro 5,1 with USB connection
Hey Maxxy... thanks for the info. I did some testing a while back in an earlier version of DP (I believe it was 8.07 running under Mountain Lion), as discussed here:

viewtopic.php?f=1&t=63021&hilit=+Fine+T ... 30#p544921

As you have noted, with external synths, DP buffer size impacted timing.

I'm curious how solid the 16A has proven while running under USB, and if you have had the opportunity to compare computer performance against a MOTU Firewire interface. My reading suggests interfacing via USB makes more demands on the CPU than interfacing via Firewire, but I have not tested this.
HC Markus
M1 Mac Studio Ultra • 64GB RAM • 828es • macOS 13.6.4 • DP 11.31
User avatar
Shooshie
Posts: 19820
Joined: Sat Oct 16, 2004 10:01 pm
Primary DAW OS: MacOS
Location: Dallas
Contact:

Re: Fine Tuning I/O ....observations in DP 9.5

Post by Shooshie »

Wasn't the drawback with USB always that it was not a multiplexing protocol? Signals were sent one direction at a time, and timing in USB was iffy. Apparently not enough to worry MOTU, as they made it their standard. But back in the day that we were having these conversations, discussing the pros and cons of future interface protocols when the future was still looming and decisions were still being made, that was the primary reason that most of us were saying no to USB. Meanwhile, manufacturers were asking one question: is it cheaper than Firewire?

And yes, it was. So that was that.
Now there's USB 3, and Firewire has politely gotten old and died. If I were looking for millisecond or nanosecond variations, USB is where I would start.

Shooshie


PS: what is the opposite of multiplex? I've forgotten the language of modems and interfaces.
|l| OS X 10.12.6 |l| DP 10.0 |l| 2.4 GHz 12-Core MacPro Mid-2012 |l| 40GB RAM |l| Mach5.3 |l| Waves 9.x |l| Altiverb |l| Ivory 2 New York Steinway |l| Wallander WIVI 2.30 Winds, Brass, Saxes |l| Garritan Aria |l| VSL 5.3.1 and VSL Pro 2.3.1 |l| Yamaha WX-5 MIDI Wind Controller |l| Roland FC-300 |l|
David Polich
Posts: 4827
Joined: Tue Dec 07, 2004 10:01 pm
Primary DAW OS: MacOS
Location: Los Angeles, CA
Contact:

Re: Fine Tuning I/O ....observations in DP 9.5

Post by David Polich »

Here's a simpler number to remember - if you are recording MIDI for an external MIDI instrument,
record it and then shift it earlier by 25 ticks. You can either select the whole MIDI track and shift it, or shift the audio track after it has been recorded.

That is a "general" number, your mileage may vary (might be 27 ticks, or 21 ticks, or something between 18 and 35 ticks). Just experiment until you arrive at
an average. In my case, 25 ticks works for every piece of external kit I have.

Adding something like 600 to a buffer of 256 - nope, that doesn't work for me. Too weird. I'll stick with 256 or 512 for recording MIDI and audio, 1024 for mixing.
2019 Mac Pro 8-core, 32GB RAM, Mac OS Ventura, MIDI Express 128, Apogee Duet 3, DP 11.2x, Waves, Slate , Izotope, UAD, Amplitube 5, Tonex, Spectrasonics, Native Instruments, Pianoteq, Soniccouture, Arturia, Amplesound, Acustica, Reason Objekt, Plasmonic, Vital, Cherry Audio, Toontrack, BFD, Yamaha Motif XF6, Yamaha Montage M6, Korg Kronos X61, Alesis Ion,Sequential Prophet 6, Sequential OB-6, Hammond XK5, Yamaha Disklavier MK 3 piano.
http://www.davepolich.com
User avatar
Maxxy
Posts: 156
Joined: Tue Jul 31, 2012 1:32 am
Primary DAW OS: MacOS
Location: Bali, Indonesia

Re: Fine Tuning I/O ....observations in DP 9.5

Post by Maxxy »

Well…I think i’ve worked through this about as far as I can and I believe I have an approach that will benefit those who are seeking maximum possible accuracy of their external MIDI devices whilst minimising time and effort to setup

David I appreciate your method and have now experimented using only MIDI Time Shift as you favour…you are right that it is simple to understand and implement… and it gets you very good accuracy …beyond that, if one is very fussy, you can simply nudge the recorded audio file into place. As you say “simply experiment until you arrive at an average”, which in your case is 25 ticks, and just use that….fine

However I’m now thinking of it like this:

Use the FTIO Record Offset as I proposed in my first post here: Divide the sample rate by 100 and add that to the Buffer Size

Then use the MIDI Time Shift to make fine adjustments to each outboard synth/sampler etc. In my case the Korg MS-20M needed no MIDI TS adjustment at all (amazingly). My Korg Wavestation SR needs to play 1051 samples later so input 1051 samples (not ticks) in MIDI TS, Roland JX needs a different adjustment, old Korg MonoPoly w/ MIDI retrofit needs a different adjustment again. The point is that each synth will have its own definite adjustment. Once that adjustment is known it is simply applied in MTS and you’re done!

Advantages of the system as I see it:

1) one simple calculation and input to FTIO per SampleRate/Buffer change

2) one fixed MTS adjustment figure per synth

3) extremely accurate

4) no experimentation necessary …ever….and therefore …

5) much faster

6) recording to audio can be done at any time…eg. after initially playing the MIDI in at a low buffer setting you can freely use a higher buffer to carry on working and record the audio whenever

7) impervious to tempo changes

8 ) eliminates the need to check or nudge audio track

9) is an elegant approach that minimises time and embarrassment in front of a client

One could make a simple document with each synth’s unique adjustment figure…which could be kept as a DP clipping….always easy to hand

I don’t think it’s weird… and you would never add 600 to your buffer figure….unless you were working at 60k sample rate ( :

Of course I’d be interested and definitely pleased if anyone cares to try it … I hope it translates and is beneficial

Best to all

Maxxy
Mac mini 2023 M2 Pro, 32g RAM, LG 32UD99 4k monitor on HDMI, OSX 13.6.1, DP 11.3, MOTU TB AVB 16A w/AVB Monitor 8 on ethernet cable, UAD Apollo x8, Netstor TB3 PCIe card expansion w/ UAD-2... 1 x Octo and 2 x Quad cards, Softube Console 1 MkII & Fader Control
User avatar
Maxxy
Posts: 156
Joined: Tue Jul 31, 2012 1:32 am
Primary DAW OS: MacOS
Location: Bali, Indonesia

Re: Fine Tuning I/O ....observations in DP 9.5

Post by Maxxy »

Hi there

Thanks for your patience as I have progressed with this idea … I need to make an important revision to the concept

The first step is to input the Formula number into FTIO Recording Offset

The second step is to put the Unique Synth Offset (USO) into MIDI Time Shift plug…(actually a unique ‘synth + MIDI interface’ offset).

But in progressing with the concept I decided to check the actual accuracy of this method from the time perspective directly rather than thinking about it in samples or ticks (which have varying relationships to time depending on tempo, tick resolution, sample rate, and buffer settings). I have used different synths with a MIDI interface and with USB direct, as well as one chart using a different audio interface using Firewire400 …the latter to address the point I think Shooshie and HCMarkus were suggesting

Pease see attached chart https://www.dropbox.com/s/i4o4jirialafk ... t.pdf?dl=0

The results show that the theory is relevant. The unique offset of each synth in combination with its MIDI and audio interfaces is shown. As most people would be working with the same MIDI and audio interface hardware in their own setup a USO should then be easily established

But here’s the rub… The MIDI Time Shift plug-in only allows Time offset to be input to 1/100th of a second accuracy ….10 milliseconds. You cannot input a MTS to millisecond or 1/10th millisecond accuracy! ) : This seems odd when it can resolve to 1 tick ….which at 125 BPM and 480 PPQ is precisely 1ms …10 times greater resolution! If one were working at 140 BPM and using 960PPQ, 1 tick would represent less than 0.45ms! Or, logically, if working at 3840 PPQ and 140 BPM one tick would be about 1/10th ms. You get the idea … and DP says that one can work at a PPQ resolution up to 10,000….wow! That is accurate!

If this is correct… it seems that MTS cannot resolve *via Real Time input* to nearly enough for the purpose of compensating *in time* which I now understand is how each unique outboard synth/sampler needs to resolve to the FTIO formula basic setup for simple and consistent accuracy and is the number required for step 2 ) :

I think that’s what I think about it…. I’ve made some progress but then hit a wall … maybe that’s what goats do ( :

Regards and thanks to all
Mac mini 2023 M2 Pro, 32g RAM, LG 32UD99 4k monitor on HDMI, OSX 13.6.1, DP 11.3, MOTU TB AVB 16A w/AVB Monitor 8 on ethernet cable, UAD Apollo x8, Netstor TB3 PCIe card expansion w/ UAD-2... 1 x Octo and 2 x Quad cards, Softube Console 1 MkII & Fader Control
User avatar
toodamnhip
Posts: 3839
Joined: Fri Jan 07, 2005 10:01 pm
Primary DAW OS: MacOS
Contact:

Re: Fine Tuning I/O ....observations in DP 9.5

Post by toodamnhip »

any recent updates to this thread?
Mac Pro (Late 2013
2.7 GHz 12-Core Intel Xeon E5
64 GB 1866 MHz DDR3
Mojave
DP 10.13
MOTU 8pre, MTP AV, 828 mkII
Tons of VIS and plug ins. SSD hard drives etc
Post Reply