Jump to content
Please note: You can easily log in to MPN using your Facebook account!

Software rig - USB midi delay? (not audio latency)


bloodyMary

Recommended Posts

Posted

I've tried a laptop rig this saturday, on a rehearsal, and wanted to share some thoughts, as well as ask those of you who use laptop rigs ,a question.

 

The rig:

 

IBM laptop

Windows XP, clean an fast install.

Brainspawn Forte VSTi host.

Lexicon alpha usb-audio interface.

ESI ROMI/O usb-MIDI interface

low-end Yamaha S03 for controller.

 

 

I felt a great deal of latency, despite having ASIO latency set for 2.67ms, which is almost as low as it gets.

After doing some research I realized that the latency was coming from MIDI! And the worst part, there was jitter (latency deviation) I could feel - I mean, it sounded like latency was jumping all the time. One bar it's very good, next bar I can't stay on time because of it!

 

 

I figured, it's either my usb-midi interface, or the cheap/slow Yamaha keyboard's midi output.

 

 

Now, the question is - do you experience things like this too, and will my situation improve if I use USB keyboard, like one of these E-MU or Edirol controllers? (pretty much set for an M-Audio Axiom or E-MU X-board, thanks to this forum).

Stage: MOX6, V-machine, and Roland AX7

Rolls PM351 for IEMs.

Home/recording: Roland FP4, a few guitars

 

  • Replies 62
  • Created
  • Last Reply
Posted

I have had a couple of laptop setups, both of which worked with no variance in MIDI and with consistent latency. The older setup was a Pentium III 1Ghz Compaq notebook in a motorized docking station that had three PCI slots. There was an eMu 1212M audio/midi pair of boards in the docking station. I could run a Kurzweil K2000VP into the MIDI in/out with no problems, and I did run an eMu XBoard 49 through USB, also with not problems. I generally used the XBoard to run Native Instruments B4 or PRO52 software.

 

The more recent notebook is a Lenovo (IBM design) ThinkPad R52, using the eMu 1616M card-bus interface. I have run the eMu Xboard into this with no problems, also my Kurzweil K2661.

 

In addition, both of the eMu interfaces provided excellent audio for both mic and line level inputs. Note that I have not tried the newer eMu 0404 UBS interface, although, based on the performance of their other gear (I also have an 1820M PCI interface on my main DAW), I believe it would work well.

 

I'm not sure if you are getting problems from the usb/midi interface or the Yamaha keyboard. Quickest way to find out would be to hook whatever pro level midi keyboard you or a nearby friend has to it and see.

 

Jim

 

Howard Grand|Hamm SK1-73|Kurz PC2|PC2X|PC3|PC3X|PC361; QSC K10's

HP DAW|Epi Les Paul & LP 5-str bass|iPad mini2

"Now faith is the substance of things hoped for, the evidence of things not seen."

Jim

Posted
Thanks, I'll try to swap the Yamaha for Korg Karma next week, and see what happens.

Stage: MOX6, V-machine, and Roland AX7

Rolls PM351 for IEMs.

Home/recording: Roland FP4, a few guitars

 

Posted

I'm using a dual P4 3GHz desktop computer with 1GB ram, with an Evolution USB controller, and the latency drives my nuts. If I use the sounds built into my sound card, there is virtually no latency. But if I use any of my VSTs, even at the fastest setting, the latency is completely unacceptable. Until this can get to 0 ms delay, I will never use a software based system on stage.

 

I can hear the latency between two hardware boards connected with a 5' MIDI cable (very slight, but noticable) so the latency I'm getting is aggrevating at the very least. I expect the sound to begin when my finger hits the bottom of the key travel, not a few milliseconds later.

 

"In the beginning, Adam had the blues, 'cause he was lonesome.

So God helped him and created woman.

 

Now everybody's got the blues."

 

Willie Dixon

 

 

 

 

 

Posted

I figured, it's either my usb-midi interface, or the cheap/slow Yamaha keyboard's midi output.

 

 

Now, the question is - do you experience things like this too, and will my situation improve if I use USB keyboard, like one of these E-MU or Edirol controllers? (pretty much set for an M-Audio Axiom or E-MU X-board, thanks to this forum).

 

I think the problem is using USB. The Yamaha should be plenty fast with MIDI to not be a problem. Try the Karma to be sure, but I'm not sure that will help.

 

Prior to getting my computer, I had an AMD K6-2 533MHz w/ 320MB ram. My sound card had a gameport that I connected a MIDI cable into. I connected to the keyboard using the MIDI, and there was virtually no latency. Since I got my current system which has no gameport, I've used a Tascam control board, as well as the Evolution USB controller, and got the same latency out of both, and this is with USB2 which is supposed to faster.

 

 

"In the beginning, Adam had the blues, 'cause he was lonesome.

So God helped him and created woman.

 

Now everybody's got the blues."

 

Willie Dixon

 

 

 

 

 

Posted

I think the problem is using USB. The Yamaha should be plenty fast with MIDI to not be a problem. Try the Karma to be sure, but I'm not sure that will help.

 

Prior to getting my computer, I had an AMD K6-2 533MHz w/ 320MB ram. My sound card had a gameport that I connected a MIDI cable into. I connected to the keyboard using the MIDI, and there was virtually no latency. Since I got my current system which has no gameport, I've used a Tascam control board, as well as the Evolution USB controller, and got the same latency out of both, and this is with USB2 which is supposed to faster.

 

 

 

If it turns out that using Korg Karma produces same results, then..not good.

 

It might be a slow/unstable usb-midi box, because somehow many people gig with laptop setups!

 

The dude from Marillion, a great prog band, gigs with a PC and many KC members do, I guess.

 

I use a laptop, so no PCI/gameport for me. Another option may be a fireWire interface. In theory it allows less jitter, so it might cure the issues..I hope.

Stage: MOX6, V-machine, and Roland AX7

Rolls PM351 for IEMs.

Home/recording: Roland FP4, a few guitars

 

Posted

Check out this site: http://www.thesycon.de/deu/latency_check.shtml

 

They have a latency checker for deferred procedure calls. This will detect a bad driver. May not be your problem but it is one possibility.

 

Make sure you MIDI device and your audio device are on different USB controllers. If you go into Control Panel/System/Hardware/Device Manager/USB Devices, right click on each USB controller, go to Properties then Advanced you can check on bandwidth reservations for each controller. Though what you do if there is a problem, other than reinstalling the driver, I don't know.

 

The other thing you can check is if there are any stray applications or processors running (like virus scanning for example), or any "fast loaded" applications.

Posted

If you're running a plain vanilla Windows XP install then you've got a TON of crap running that you don't need/want, all of which will steal resources from your MIDI (and audio) processing.

 

Google for Windows XP Audio Configuration and read up on how to remove/disable all non-necessary services and startup processes. Drop your graphics settings to the bare minimum required (you do NOT require 24-bit graphics, for example).

 

All of these things will have a profound impact on the perceived performance of your system.

 

 

Posted

Thanks for replies, guys. I'm checking the bandwith/controller llocation right now.

 

Of course, I don't have nything running but the audio apps. The system has all the unnecessry processes turned off, and extra devices (WiFi, bluetooth, LPT port, cardbus controller, etc) are disabled.

 

The only thing I left on is IBM's "Active Protection", which monitors hard drive and shuts it down if the laptop is falling or shaking. I felt it's safer to leave it on. Do you think it may eat too much resources?

 

 

BTW, the latecy checker tool says everything's fine.

 

 

 

I just don't get it. There are a lot of people using laptop rigs, right? They hve no udible ltency, I suppose. How do I get the same? My ASIO drivers are set for 2.6ms latency, the system is optimized, but still, if I layer the software synth with h/w synth, and trigger them from the same keybord, there's delay between two sounds, even not considering the jitter I've been talking about.

Stage: MOX6, V-machine, and Roland AX7

Rolls PM351 for IEMs.

Home/recording: Roland FP4, a few guitars

 

Posted
But if I use any of my VSTs, even at the fastest setting, the latency is completely unacceptable.

Do your sound card has native ASIO drivers?

 

I expect the sound to begin when my finger hits the bottom of the key travel, not a few milliseconds later

Well, you already get few ms of latency just because of distance that sound waves have to travel from speakers before reaching your ears. So you're trying to achieve something unachievable.

Posted
I expect the sound to begin when my finger hits the bottom of the key travel, not a few milliseconds later

Well, you already get few ms of latency just because of distance that sound waves have to travel from speakers before reaching your ears. So you're trying to achieve something unachievable.

 

Yeah, when you're playing in a stadium, this might be true. No offense to you, Mr. Nighttime, but I doubt you live in a stadium... I also had lotsa latency eventhough I had ASIO4ALL installed. I bought a Mac. Now I have no more hearable latency (smug grin)

It's not a clone, it's a Suzuki.
Posted

I've had past issues with MIDI interfaces. Was using a MOTU FastLane USB (rest of chain was Kurz K2600 and Mac OSX) and had all sorts of latency, false and spurious MIDI data being interpreted, just all sorts of unreliable crap.

 

Long story short, I quickly found the problem was the FastLane MIDI interface. When I purchased my Novation Remote 61SL and started using that as the MIDI interface (K2600-->Novation--Via USB-->MacBook Pro), all the problems went away.

 

See if you can borrow or otherwise get your hands on a higher level MIDI interface (a MOTU MIDI Timepiece? A Firewire audio interface w/ MIDI?) and see if that solves the problem.

 

So to answer some of the questions: gig w/ laptop, no latency, very stable, using USB. Don't know how much of this is because I went Mac vs. Windows.

 

 

..
Posted
I expect the sound to begin when my finger hits the bottom of the key travel, not a few milliseconds later

Well, you already get few ms of latency just because of distance that sound waves have to travel from speakers before reaching your ears. So you're trying to achieve something unachievable.

 

Yeah, when you're playing in a stadium, this might be true. No offense to you, Mr. Nighttime, but I doubt you live in a stadium... I also had lotsa latency eventhough I had ASIO4ALL installed. I bought a Mac. Now I have no more hearable latency (smug grin)

No offense to you, Mo, but you don't know what you're talking about. And too bad about having to get a Mac, since Windows works fine for low latency (smug grin).

 

Sound travels at roughly one millisecond per foot. Unless you're wearing headphones or have your head inches from your monitor, you'll get a few milliseconds of latency due to the speed of sound. If you're noticing latency, it's more than "a few milliseconds". Most likely, 20 ms or more. Try an experiment with a normal hardware rig and see if you can tell the difference if you move your monitor 10 feet farther away. I sure can't!

 

Make sure you're really using the ASIO interface that's configured for low latency.

 

Please explain how you isolated the problem to MIDI, and tell us what softsynth you're using.

 

With MIDI cables, you'll get a latency of about 1 ms per note. That is, if you play a 5 note chord, hitting all notes at exactly the same time (yeah right), the 1st note will be delayed by 1 ms, the second by 2, etc. This is the serial "shifting delay" caused by the data rate on a MIDI cable and the length of a MIDI note-on message (3 bytes).

 

When using a USB cable rather than a MIDI cable, the data rate is far, far higher, and you won't get this kind of latency. The only cause for MIDI latency in this case would be delays inside the keyboard itself or in computer software (drivers, softsynth host, or possibly the softsynth itself).

 

BTW, if you're using Windows and the built-in wavetable softsynth, that softsynth has *lots* of latency that there's no way to avoid. It wasn't meant to be played by MIDI keyboard.

 

 

 

Posted
I expect the sound to begin when my finger hits the bottom of the key travel, not a few milliseconds later

Well, you already get few ms of latency just because of distance that sound waves have to travel from speakers before reaching your ears. So you're trying to achieve something unachievable.

 

Yeah, when you're playing in a stadium, this might be true. No offense to you, Mr. Nighttime, but I doubt you live in a stadium... I also had lotsa latency eventhough I had ASIO4ALL installed. I bought a Mac. Now I have no more hearable latency (smug grin)

No offense to you, Mo, but you don't know what you're talking about. And too bad about having to get a Mac, since Windows works fine for low latency (smug grin).

 

Sound travels at roughly one millisecond per foot. Unless you're wearing headphones or have your head inches from your monitor, you'll get a few milliseconds of latency due to the speed of sound. If you're noticing latency, it's more than "a few milliseconds". Most likely, 20 ms or more. Try an experiment with a normal hardware rig and see if you can tell the difference if you move your monitor 10 feet farther away. I sure can't!

 

Make sure you're really using the ASIO interface that's configured for low latency.

 

Please explain how you isolated the problem to MIDI, and tell us what softsynth you're using.

 

With MIDI cables, you'll get a latency of about 1 ms per note. That is, if you play a 5 note chord, hitting all notes at exactly the same time (yeah right), the 1st note will be delayed by 1 ms, the second by 2, etc. This is the serial "shifting delay" caused by the data rate on a MIDI cable and the length of a MIDI note-on message (3 bytes).

 

When using a USB cable rather than a MIDI cable, the data rate is far, far higher, and you won't get this kind of latency. The only cause for MIDI latency in this case would be delays inside the keyboard itself or in computer software (drivers, softsynth host, or possibly the softsynth itself).

 

BTW, if you're using Windows and the built-in wavetable softsynth, that softsynth has *lots* of latency that there's no way to avoid. It wasn't meant to be played by MIDI keyboard.

 

 

The MIDI latency I was talking about is when I play a note on my XB-2, with the N364 as the sound source, I can hear a difference in response compared to hitting the note on the N364, with a 5' MIDI cable connecting the two. I can live with that. (Although I'm not sure why, but the tone of the N364 is a bit muffled controlling from the XB-2 compared to from its own keyboard).

 

I'll have to see just what is causing my problem on the PC. Could be the sound card, but I got the same amount of latency coming from my Tascam as well. It's weird, but if I use the native GM sounds from the card, there's no noticable latency at all.

 

 

"In the beginning, Adam had the blues, 'cause he was lonesome.

So God helped him and created woman.

 

Now everybody's got the blues."

 

Willie Dixon

 

 

 

 

 

Posted
It's weird, but if I use the native GM sounds from the card, there's no noticable latency at all.

 

This is proof that you're looking at audio latency, not MIDI latency.

 

Posted

Learjeff, are you sure that using usb controller allows more bandwidth than plain midi? Will using a usb keyboard give less latency than midi+usb adapter?

 

Stage: MOX6, V-machine, and Roland AX7

Rolls PM351 for IEMs.

Home/recording: Roland FP4, a few guitars

 

Posted

 

I felt a great deal of latency, despite having ASIO latency set for 2.67ms, which is almost as low as it gets.

After doing some research I realized that the latency was coming from MIDI! And the worst part, there was jitter (latency deviation) I could feel - I mean, it sounded like latency was jumping all the time. One bar it's very good, next bar I can't stay on time because of it!

 

 

I figured, it's either my usb-midi interface, or the cheap/slow Yamaha keyboard's midi output.

 

 

Now, the question is - do you experience things like this too, and will my situation improve if I use USB keyboard, like one of these E-MU or Edirol controllers? (pretty much set for an M-Audio Axiom or E-MU X-board, thanks to this forum).

 

Hi !

 

Maybe you should read this 1st and download the midi test programs, Midi Test and Midi Time.

 

http://www.earthvegaconnection.com/

 

http://www.jay.fm/blog/pc-midi-timing-and-nuendo.html

 

I doubt, your midi keyboard(s) or the serial transmission of midi data in general is the cause of your probs. I have several midi interfaces connected to different machines,- Midex-8, 8Port SE (PC) and Miditimepiece II (Mac) as also Sycologic midi routers and a Miditemp PMM88E to control midified hardware racks including long cable runs. All works fine, regardless of it´s age, USB or not USB.

 

My experience is, it´s a question of the 3rd party drivers for the midi interfaces and it´s somewhat hardware related,- clock stability of the Mobo in a computer. You can do a selection for MoBos in a desktop machine, but not for a notebook unfortunally.

 

It´s hard to find PC notebooks which work well in respect on audio and midi latency/jitter.

 

If you want to go the virtual instruments route for gigging w/ PC, try to use a rackmount PC which is optimized for audio and midi performance. That means: Try to avoid USB for audio streaming as also sample streaming. If you decide to use a firewire solution, check the RME devices which come w/ a midi interface included.

Eventually, a PCIe based cardbus interface could be a good solution w/ a carefully selected laptop.

 

If you´re not one of the guys who´s hobby is fiddling around w/ PC tweaks, a Macbook might be your best choice.

 

I myself don´t trust computers on stage unless you go with a quality line conditioner and power stabilizer.

And I haven´t seen any pro keyboardist on stage w/ only a computer/virtual instruments setup. They all use combinations of hardware and software instruments or hardware only.

 

A.C.

Posted
If you're running a plain vanilla Windows XP install then you've got a TON of crap running that you don't need/want, all of which will steal resources from your MIDI (and audio) processing.

 

Google for Windows XP Audio Configuration and read up on how to remove/disable all non-necessary services and startup processes. Drop your graphics settings to the bare minimum required (you do NOT require 24-bit graphics, for example).

 

All of these things will have a profound impact on the perceived performance of your system.

 

 

Adding "tuning" to the search seems to help a lot. I found this web page to be particularly helpful. In fact I just checked and discovered I still had indexing enabled on the hard disk I recently added. That is a major performance hog.

 

[edit:spelling]

Posted
]

 

Yeah, when you're playing in a stadium, this might be true. No offense to you, Mr. Nighttime, but I doubt you live in a stadium... I also had lotsa latency eventhough I had ASIO4ALL installed. I bought a Mac. Now I have no more hearable latency (smug grin)

 

I don't think the latencies achievable with ASIO4ALL are very good. That's my experience at least. I don't know if it is sitting on top of DirectX but if so that would explain it - DirectX specs are very loose and it seems to do poorly in practice as well I have never attempted to do formal measurements but when Windows media player can't even keep your sound and video in synch, which you really notice if you record yourself playing and match the key presses to the sound, and given that the replay engine is pretty much directX, it clearly is not very good.

Posted
ASIO4ALL translates WDM into ASIO. WDM can be capable of pretty high perfomance. My built-in audio card can go to 1.33ms with ASIO4ALL, while the USB interface, Lexicon Alpha, goes to 2.67ms with latest native ASIO drivers.

Stage: MOX6, V-machine, and Roland AX7

Rolls PM351 for IEMs.

Home/recording: Roland FP4, a few guitars

 

Posted

Hi !

 

Maybe you should read this 1st and download the midi test programs, Midi Test and Midi Time.

 

http://www.earthvegaconnection.com/

 

http://www.jay.fm/blog/pc-midi-timing-and-nuendo.html

 

I doubt, your midi keyboard(s) or the serial transmission of midi data in general is the cause of your probs. I have several midi interfaces connected to different machines,- Midex-8, 8Port SE (PC) and Miditimepiece II (Mac) as also Sycologic midi routers and a Miditemp PMM88E to control midified hardware racks including long cable runs. All works fine, regardless of it´s age, USB or not USB.

 

My experience is, it´s a question of the 3rd party drivers for the midi interfaces and it´s somewhat hardware related,- clock stability of the Mobo in a computer. You can do a selection for MoBos in a desktop machine, but not for a notebook unfortunally.

 

It´s hard to find PC notebooks which work well in respect on audio and midi latency/jitter.

 

If you want to go the virtual instruments route for gigging w/ PC, try to use a rackmount PC which is optimized for audio and midi performance. That means: Try to avoid USB for audio streaming as also sample streaming. If you decide to use a firewire solution, check the RME devices which come w/ a midi interface included.

Eventually, a PCIe based cardbus interface could be a good solution w/ a carefully selected laptop.

 

If you´re not one of the guys who´s hobby is fiddling around w/ PC tweaks, a Macbook might be your best choice.

 

I myself don´t trust computers on stage unless you go with a quality line conditioner and power stabilizer.

And I haven´t seen any pro keyboardist on stage w/ only a computer/virtual instruments setup. They all use combinations of hardware and software instruments or hardware only.

 

A.C.

 

 

 

Hello and welcome to the forum!

 

First of all, I'm not very comfortable with bringing rackmount PC, UPS, monitor, keyboard and mouse to gigs. For me laptop rig is about quality and portbility.

 

 

I'm not enjoying tweaking the PC, but I definately can do it, and have tweaked the hell out of mine (beside overclocking, which I don't want on a music machine).

 

 

My laptop is an IBM, I believe it's a quality brand, and it has quality (=stable and fast) components inside.

 

Getting a firewire or PCcard interface with both MIDI and audio I/O may be an option, I suppose.

 

 

As I wrote, Marillion's player uses VSTi's on stage. Go to brainspawn.com (not an ad, it's just a great product) to read an article about him and his rig.

 

 

For now I'll probably start incorporating a laptop into my rig by having it take over few sounds (leads, Hammond), and more later. I'll control it with my Korg Karma, and use it's internal sounds too. Should be a smooth transition for me.

Stage: MOX6, V-machine, and Roland AX7

Rolls PM351 for IEMs.

Home/recording: Roland FP4, a few guitars

 

Posted

Hi !

 

Maybe you should read this 1st and download the midi test programs, Midi Test and Midi Time.

 

 

Hello and welcome to the forum!

 

As I wrote, Marillion's player uses VSTi's on stage. Go to brainspawn.com (not an ad, it's just a great product) to read an article about him and his rig.

 

 

For now I'll probably start incorporating a laptop into my rig by having it take over few sounds (leads, Hammond), and more later. I'll control it with my Korg Karma, and use it's internal sounds too. Should be a smooth transition for me.

 

Hello ´n thx for your welcome!

 

I know the B.F. program as also the story about Marillions keyboardist.

AFAIK, he´s using a rackmount 4HU PC, not a notebook.

And he uses several keyboards too which aren´t midi-controllers only but also tone generators, so you don´t know what he´s really doing w/ the VSTis and what he´s doing w/ the hardware instruments. Also, I don´t believe too much of the storys of endorsers. B.F. is not a perfect program and I´m sure this guy has VIP support, maybe a personal hotline.

 

As I said, if your notebook is the right one,- in the case you decide for a firewire solution check if the notebook has TI chips inside for the FiWi controller,- and if you have a clean system, you might have luck w/ midi-audio latency and running a few VSTis, how many depends on your cpu.

 

But my experience is, combining hardware instruments w/ virtual instruments results in an audible latency ever if you layer these, this regardless how good your computer works. This is very audible if these sounds have a percussive attack.

 

You mentioned the organ and I myself like the NI B4 (I own everything from NI except Traktor and Kore b.t.w.),- but playing the virtual organ like a real one, w/ complex fast runs and precise phrasing is timing-critical ´caused by latency and depends on tempo of the song too.

 

Let´s say, you´ll get a ideal ASIO driver latency of roundabout 2ms (out only)w/ your machine, plus midi drivers latency (an addition of Windows USB drivers and 3rd party ones for DirectMusic p.x.), plus midi delay ´cause it´s serial and plus the converters latency.

I doubt you end up w/ a real world delay under 10ms, maybe even more.

With this, you´re able to play sidelines, themes, chords and pads, but w/ some lead/soloing stuff, you probably end up feeling like a dancer on a dancefloor w/ some loam under his shoes.

 

Next prob is, if you use a standard midi-keyboard to play the virtual organ, you´ll recognize you must have the key all the way down to hear a note. With a real organ or the best clones, you only touch the key and the tone is there.

 

There´s the theory all around, if you have a latency of 10ms, this compares to standing away 10feet from your amp. That´s right, but in the real world, the distance to your amp exists in addition and if this is 10 feet too, you hear 20ms delay instead of 10 unless you use in ear monitoring.

 

A.C.

 

 

 

 

Posted
ASIO4ALL translates WDM into ASIO. WDM can be capable of pretty high perfomance. My built-in audio card can go to 1.33ms with ASIO4ALL, while the USB interface, Lexicon Alpha, goes to 2.67ms with latest native ASIO drivers.

 

Did you mean to write that ASIO4ALL sits on top of WDM. That would explain my poor results - I was presumably going through the kMixer. That is said to add a lot of latency. Were you able to turn that off? If so, how did you do it?

What are you running?

Posted
Learjeff, are you sure that using usb controller allows more bandwidth than plain midi? Will using a usb keyboard give less latency than midi+usb adapter?
In theory, yes, but not enough to matter. Are you really going to notice 1 msec of extra latency? No, just like you won't notice the delay if the speaker is 1 foot farther away.

 

Audio latency is the one you need to worry about, and anything below 10 msec should be fine, unless your monitors are rather far away in which case you might need to get it lower. Under 5 msec isn't hard for most computers to manage these days. Anyone worried about getting below 5 is most likely tilting at windmills: working the spec for the spec's sake rather than learning to trust their ears.

Posted

BTW, here's how to measure total latency.

 

Choose a nice percussive patch. Stick a mic close to a key. Put the speaker close to the mic. Plug the mic into a recorder, with record monitoring turned off. Hit record.

 

WHACK the key, in a way that it'll make a sound the mike will pick up -- along with the played patch. Do it once with the sound turned off so you can see/hear how long it takes between when your finger hits the key and when the key hits the bottom of travel. (To amplify this, hold a butter knife barely above the key surface before you whack it. In the recording, the click of your finger pushing the knife into the key will be distinct from the key hitting the bottom of travel.)

 

Inspect the recording. In the days of reel-to-reel tape, we'd 'scrub' to see where on the tape we hear the key whack and where the played note starts and measure the distance between, and divide by the tape speed. E.g., one inch at 15 IPS is 1/15 sec, or about 66 msec.

 

But with lovely digital recordings, we can just load them into our DAW, zoom in on the waveform, see the obvious key whack and note waveforms, and look on the timeline to measure the interval.

 

Bullet-proof, incontrovertible way to measure total latency for playing keyboards. Note that your ROMPler will have latency too: it's not a miracle device. A true analog synth will have latency too, but it will be far too small to be measurable this way.

 

Then go measure your Rhodes or your grand piano, at different velocities.

 

Then tell me how much 2 or 3 msec of softsynth latency matters.

 

:D

 

It's an important subject, but one that folks can go way off the deep end trying to achieve an impossible and unnecessary goal. Under 5 ms does the trick, for all practical purposes, and under 10 is good enough for most folks in most situations. 20 is noticeable -- tolerable by me but not many others; my timing isn't good enough to worry about it.

Posted
Did you mean to write that ASIO4ALL sits on top of WDM. That would explain my poor results - I was presumably going through the kMixer. That is said to add a lot of latency. Were you able to turn that off? If so, how did you do it?

What are you running?

Yes, ASIO4ALL presents an ASIO interface to other programs, providing that service using WDM services.
Posted
But my experience is, combining hardware instruments w/ virtual instruments results in an audible latency ever if you layer these, this regardless how good your computer works. This is very audible if these sounds have a percussive attack.

Good point. I find this is true even when combining MIDI analog synths with electromechanical instruments like Yamaha CP70. Layering percussive sounds is very tricky and generally best avoided if you want a tight sound. Admittedly, softsynths add yet another significant variable. In this case, a few milliseconds DOES make a difference. Not so much as a timing problem, but a tone/imaging problem, where the resulting sound is loose rather than tight.

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...