[BUG] Recording Midi at ANY latency on ANY system (sync)

if you are at any buffer above 64 samples.

And why would anyone want to do that?
And why would they need to? Because their system is not configured properly in the first place for any modern DAW.
Don’t tell me it works on X or Y DAW because the physics will not add up and furthermore it wouldn’t be relevant to this thread.
Jitter is a hardware expression. Originally “the jitters” comes from the physical shakes of alcoholism or being frightened of demons. Hardware jitter is the same but caused by solid componentry vibrating.
Software can’t do that. Software has “glitches”.
If we mix in the jitter rubbish we’ll get even further from solving any problem here. Another matter entirely.

Sigh…this is the same issue man, including the jitter, and I don’t think you are helping the cause by trying to shut down your allies. All these MIDI problems have to do with how Cubase interprets the time difference between when an event is heard and when it is triggered. Same internal code.

Hi Robotpriest

What you describe is EXACTLY what happens when you play a VSTi in time with a Cubase arrangement/click track. You play ahead of the beat because the sound you hear from the VSTi is delayed whilst you are playing it. You automatically compensate for it, maybe without even realising.

When you play the recording back, there is no such delay and so your ahead of-of-the-beat playing sounds just as you would expect - ahead of the beat.

it’s not Cubase’s fault…

Heh funny post, but guys seriously read this. This Cubase problem is ancient and it is absolutely related to the “early notes” problem you are finding now.

Thanks… I don’t know what else to say, your signature shows you run Nuendo also. Perhaps you can test the same thing on that system…

Do you have a optional sound card to try easily?

Do you have optional driver to try? (Sorry I’m not familiar with MAC architecture , ASIO/Core??)

Ok…its not Cubase’s fault in the sense that Cubase is a non-conscious entity that does what it is programmed. As I said, Cubase records the MIDI with the timing it was played at. But that does not match what you hear whilst you are playing. It is very difficult to play one thing and hear another! Don’t you think so?? Sit at a piano and play for a hour. Now go to Cubase and record something to a metronome and then play it back. Not good.

Steinberg should match the recorded MIDI to what you heard while playing. This is what Logic does and it works better.

Well how come several million users don’t have the problem?
You may mean well but you are not an ally here. And how, exactly, is Cubase supposed to know when YOU hear a note played? It only “hears” the trigger.
It is in that box and you are outside pressing keys.
“Same internal code” What’s that supposed to mean exactly? Of course it is. We all know that. Couldn’t be anythng else really, could it? :mrgreen:

Or at least make it an option which behavior you want. But I really doubt once anyone had tried having recorded MIDI aligned to what they heard while making the recording that anyone would want to go back to the old way. It is just too bizarre feeling to try and play a rhythm that sounds wrong while you are doing the recording.

Robot, I simply trying to simplify this ‘test’ for the purposes of narrowing down this problem to see how many are affected or not.

I decided that for this test we can take VSTi out of the equation even though recording to a vsti track will most likely exhibit the exact same symptoms if Cubase is not recording Midi as expect. Therefore, I would like to ask that for the simplicity of testing we eliminate various VSTi or complex Midi routing to a single Midi input device and a single audio device.

Not trying to disrespect anyone or discredit them.

Honestly, I think most of the users don’t have the chops to notice these problems strange as it sounds.

Cubase knows when I heard the note played; its whenever Cubase played the audio. Let me give you an example. Lets say buffer is 1024. At time 0 you press C#. Cubase plays the audio from the instrument at time 1024. Now when Cubase sets the time of the recorded MIDI event, should it use time 0 or time 1024? I believe the answer is 1024.

It’s because the common practice is to switch the buffer setting to ‘low latency mode’ when recording and switching back to high latency mode for mixing and heavy cpu intensive tracks… When switch ‘modes’, audio is always recorded “as you hear it” (ie. in time with the playback position).

Audio recording doesn’t suffer in high latency mode, so why does Midi suffer (on affected system)?

no prob :slight_smile:

Not sure it does. AIUI, Logic uses a neat system for providing very low latency for VSTis whilst running a much bigger buffer for the rest of the system. So when you record MIDI you have less delay to compensate for, and your timing is more accurate.

The problem really is latency in the system - most notably when you play a VSTi. As someone once observed, latency is like the sand in your sandwich.

And what that tells me is that any given DAW (and they all seem to be mantioned here) corrects errata sent by the relevant interface driver.
The jitter is in the hardware (soundcard), the driver compensates and the DAW compensates even more. And at the bottom of the box is the sensible solution of lowering the buffer rate to about 6ms. Furthermore, what is not mentioned is that the DAW also has to compensate for the PLAYER as well.
Don’t forget that the midi always goes through the WHOLE system even when you have no VSTi assigned!
The midi notes also come in and are recorded whether there’s any VSTi or external synth there or not. They are sent to them later.
Thankfully for me if not for some I have always found, timingwise, that most recording mediums have always been more accurate than myself. :mrgreen:

Actually, thinking about why people have not been screaming about this issue more, another reason might be that alot of the new users have been exclusively using softsynths and don’t remember what it used to feel like to play on the old rompler synths where the latency was near zero. So the standard for timing is so low now, the expectation for Cubase is also very low timing-wise.

Audio certainly does suffer from latency delays if you monitor through Cubase. This forced most of us to change the way we work with DAWs compared with tape/mixer - I used to monitor record channels through the mix channel - that way it sounds just the same when recording or playing back.

If you try that with any native DAW, the latency immediately bites you in the leg. And of course the talent hears themself with an echo.

No matter, I’ve grown to love Totalmix :smiley:

I’ve been playing many years. Players may not have chops but their ears usually do, as does any one who listens to music. It’s why they invented overdubbing.
Cubase hears it at 1024. Then processes it to the timescale of the Project events. It may log it at 1024 then should correct it according to the user settings.
As this thread is not a global issue I don’t think we should try to make it one.

– By knowing how long it takes for the VSTi note to start after the MIDI note-on appears. That (surely?) must be something Cubase has to know in order to be able to do the delay compensation during any subsequent playback. So if someone’s pressing the keys (say) 25 milliseconds early in order to make the VSTi sound in time with what’s being played back, that 25 ms will be perhaps wholly or, at least, partly determined by the known (to Cubase) time between the MIDI note-on and the corresponding sound starting in that particular VSTi. If there are other components contributing to the 25 seconds, it seems to me unlikely that either the programmers can’t determine what they are or Cubase doesn’t know their lengths - because mustn’t they all be known to Cubase for purposes of delay compensation? (I’m assuming the time taken for the sound to go from the monitors to the performer can be disregarded.)

But this is OT. Beerbong wants this thread to focus on Cubase’s placement of MIDI notes in circumstances where the performer isn’t hearing a VSTi, and therefore can’t be compensating for delays.

Sigh. It knows when you pressed the key. It cannot know whan you HEARD it. Could be half an hour later for all we know. :mrgreen:
I don’t know or can assume what beerbong wants. Doesn’t matter who hears what though if Cubase places something in the wrong place it’s got nothing to do with what you hear it’s all to do with what’s going through the system and causing the displacement. If your buffers are set right then you will hear it pretty much where you want to hear it.

Larger buffers and compensating for hearing the note incorrectly placed is down to building the DAW system properly.

Why not? Please explain; what am I missing? Cubase knows how long it takes from the note press til the VSTi starts making the sound (otherwise it couldn’t do the delay compensation during playback). The monitors are (assumed to be) close to the player (so only a few milliseconds for the sound to travel). What is is that accounts for the substantial unknown (to Cubase) delay that you imply can exist somewhere between the VSTi and the amp? (therefore making it impossible for Cubase to allow for)


And neither do I. When I said what Beerbong wanted his thread to focus on, I was just repeating what he had said. However, I’m content to believe he wasn’t lying until something happens to change my mind.


Isn’t that just an unfounded assertion? You might be absolutely right. But I don’t see how you can prove that the hypothesised performers’ compensation behaviour could have no impact on whether the MIDI notes appeared early. If you can’t prove that, I don’t see that you have any basis for saying that it doesn’t matter who hears what.