When dealing with issues involving clock, especially from a DAW out to external sequencers, these two terms often come up: Jitter and Latency
Both are responsible for problems in trying to achieve an accurate clock, but they are not interchangeable. So what do they actually mean?
Jitter
This is a timing issue where the clock does not flow at a steady pace. Jitter is always present in digital devices, it's just a matter of severity.
Think of it like physical gears. If the gears are not manufactured precisely and do not lock together perfectly in a system, the movement of the overall system will be sloppy. If it's a severe enough problem, the movement can actually jerk forward and back.
It's the same for clock. There's a virtual "gear" in the DAW that sits between the running of audio and the sending of MIDI Clock. Those 2 things aren't coupled as tightly together as it may seem. Depending on the setup, there may even added virtual "gears" between them, such as when a class-compliant USB connection is involved.
Reducing Jitter as much as possible is the first step toward achieving a reliable clock. One of the best methods for clocking that will reduce jitter to nearly non-existent levels is "Sample Accurate Clock", where an audio track is the source for clock generation. This will often require an external device that works as a bridge between the audio clock generation and the MIDI Clock or CV Sync that the external hardware needs.
Latency
This just means there is a delay of some kind in the system. The clock an be reliable and have minimal Jitter, but the time it takes to go from the generation of a clock signal to the point where you experience a result may have a lag. In music creation using a DAW and external sequencing gear, it's often said to "not be on the grid". Latency is when every beat is ALWAYS the same number of milliseconds off the grid.
Reducing latency depends on where the latency is actually developing. It can happen in many places. This isn't going to be a comprehensive look at everywhere this can occur, but one most common area is the Audio Input for recording. The more of a buffer that the audio interface is configured to use, the more latency will exist as the sound is recorded. Most DAWs have a way to automatically compensate for this by nudging a newly recorded track back by a defined number of milliseconds.
But latency can also exist in monitoring. Nudging a track back after the recording is not real-time. For that, real-time latency can be reduced significantly by bypassing the DAW during monitoring so that it doesn't echo the input. Instead the audio interface sends the input directly to the speakers themselves (often referred to as "Zero-Latency" monitoring). Another way to achieve this is by placing a device between the DAW and the external sequencer which manipulates the clock signal in real-time, causing the external sequencer to play "early".
Solution
First is knowing exactly what kind of trouble your clock is experiencing. It does no good to be experiencing Latency to then have others recommend that you get a device that only solves Jitter or vice-versa.
We do make a device that can solve both Jitter and Latency when using external sequencers with a DAW called CLOCKstep:MULTI.