What is level of service in a traffic capacity study?

Level of service, or LOS, is the letter grade your capacity analysis spits out at the end. A through F, same scale a lot of us learned in school, except here F doesn't mean "failing" exactly. It means demand is bumping up against capacity and delay is climbing past what the Highway Capacity Manual calls acceptable for that facility type.
The grade comes out of HCS, Synchro, or whatever software your shop runs, but the number underneath it is the volume-to-capacity ratio, v/c for short. Feed in your traffic counts, your geometry, your signal timing if there is any, and the model works out how much of the link's or intersection's theoretical capacity is actually being used during the analysis hour. Low v/c, free flow, LOS A or B. v/c creeping past 0.9, you're looking at D or E. Over capacity, F, and now you're the one explaining to the MPO why the signal warrant study is suddenly on the table.
How the grade gets set
For signalized intersections, HCM ties LOS to average control delay per vehicle, measured in seconds. Roughly: LOS A is under 10 seconds of delay, LOS F is over 80. Unsignalized intersections and two-lane highways use different delay thresholds because the failure mode is different; a stop-controlled approach backing up behaves nothing like a signal cycling through phases. Freeway segments and basic links run off density, passenger cars per mile per lane, because on a freeway nobody's waiting at a red light, they're just packing in tighter until flow breaks down.
None of that changes the fact that every one of these analyses starts in the same place: a turning movement count or a directional link count for the peak hour you're studying. The LOS letter is only as good as the volume you handed the model. Run a capacity analysis off a count that's two years stale, or off a seasonal factor you eyeballed, and the grade on page 14 of your report is a grade for a road that doesn't exist anymore.
Why the count matters more than the software
A lot of capacity studies live or die on a handful of links where nobody has current volumes. Maybe it's a collector that never got a permanent count station, or a corridor segment between two signals where the DOT only counts at the intersections themselves. Whatever the reason, you end up either sending a crew to lay tubes for a week, which costs money and tells you almost nothing about peak-hour directional splits by lane, or you run with an old count and footnote the assumption.
Vehicle Counting exists for that gap, the link nobody instrumented and nobody wants to send a crew to. One scheduled pass over the corridor, one image, and you get counts by link and direction back as a CSV you can drop straight into your capacity worksheet, plus the marked-up frame the counts came from so you can sanity-check the detection yourself before it goes in the file.
LOS as a grading scale isn't complicated once you've run a few. The part that eats a project's budget is getting a defensible count for the link that's been sitting uninstrumented for a decade. If that's the piece you're stuck on for an upcoming signal timing or capacity study, it's worth a look.