Bandwidth alone isn't enough to ensure that a DSO can accurately capture a high-frequency signal. The goal of a DSO design is to have a maximally flat envelope delay (MFED). A frequency response of this type exhibits excellent pulse fidelity with minimum overshoot and ringing. Because a DSO or DPO is comprised of real amplifiers, attenuators, analog-to-digital converters (ADCs), and interconnections, the MFED response is a goal that can only be approached. Pulse fidelity varies considerably based on the model and manufacturer.
The sampling rate denotes how many samples per second the ADCs—and, therefore, the DSO—can acquire. Maximum sampling rates are usually specified in megasamples per second (Msamples/s). The faster the DSO can sample, the more accurately it can represent fine details in a fast signal. If the user needs to view slowly changing signals over long periods of time, the minimum sampling rate may also be important.
The record length of a DSO denotes how many waveform points the instrument can acquire for one waveform record. Some DSOs let the user adjust the record length. The maximum record length depends on the amount of memory in the DSO and its ability to combine memory length from unused channels. Because the DSO can only store a finite number of waveform points, there's a tradeoff between record detail and record length. A user can acquire either a detailed picture of a signal for a short period of time (the DSO "fills up" on waveform points quickly), or a less detailed picture for a longer period of time. Some DSOs make it possible for the user to add more memory to increase the record length for special applications.