Context is everything. There are multiple domains you're generalizing that contribute to the 'problem' of poor playback experience.
For Video, specifically MPEG variants there are a series of frames that comprise the video and look of the form:
K1,P1,P2,P3,K2,P1,P2,P3 - K are I frames, P are P frames
If you have K1,P1,,P3 then your picture will 'glitch' out on part of the frame missing as it can't reconstruct it from missing data. the glitch can be blank space ('blackout'), frozen space from the last known frame ('freeze') (P1 repeats in this case until the next I frame to catch up). if you have partial/malformed data it can also screw up the subsequent DCT decoding and be 'corrupt' imagery. sometimes fast moving data will appear like this as the nature of MPEG can't deal with compression of rapid transients well (water, fire).
If you have K1,P1,P2, it can look 'choppy' as it misses the last frame before a full screen refresh much like the choppy water effect.
If you only see K,P1,, habitually like it is rendering some frames then cuts out that usually means the source data isn't arriving (network latency/bandwidth delay product) or being processed too slow (cpu or gpu depending on if hardware offloading support is present). The former is 'buffering' and the latter is 'video lag'
If the rendering is cpu/gpu based depending on the framerate and framesize your computer may not be fast enough to accomplish the workload required in the time you ask of it. If you find your playback shows fine for a few seconds then goes herky jerky and stalls a few seconds that usually means you're underpowered in terms of memory or cpu power. Reducing the number of running apps to relieve resource contention typically recovers those. Also restarting a player with a memory leak that has been running too long (Hi Flash and Silverlight!) is common.
In the domain of sound the phenomena depends on what is your reference. Is the volume modulating up and down or just sound like they're getting nearer/farther? That can happen as the codec selected (video can do this for adaptive playback) gets shifted based on the predictive available network bandwidth. mobile/voip in particular have these issues. that's why people ask you to call them on a land/hardline to ensure the entire call is full 64k (56k with robbed bit for the pedants is possible too).
For phones specifically it is probably codec shifting or codec mismatch. part of the call is in 64K mode, and another leg is in 8K mode, followed by another 64K leg and the constant lossy compression being applied and removed ends up in a mutant signal. On older telephone networks it could be a tandem stacking issue that introduced small latency that would manifest in sort of negative feedback loops. both manifest as the 'tunnel' effect.
All of these cases are generally summarized as missing data, slow data, or corrupt data. All of which are common to any communication conversation. The domain it applies to will stipulate how it manifests invariably as some form of 'glitch' depending on where the issue is and how the vendor implements playback in the presence of these errors. Some players endure hardship better than others.