We often encounter these concepts in our work. Because these concepts are very similar in meaning, they cause us a lot of confusion, and sometimes we confuse them. Today we will talk about these concepts, and hope to share with everyone. Understanding them can help.
Before describing these concepts, let's take a look at how a signal is generally transmitted over a channel.
As mentioned above, although we only want to transmit bit0~bit7, due to the limitation of the communication mechanism, we must add the start bit before bit0, add the stop bit after bit7, and other redundant parts. For reference, let's see how the above concepts are defined.
- Baud or Baud rate, the unit is bps (bits per second) or baud itself is the unit, that is, 1baud = 1bps, the baud rate is t1 in the figure, indicating how many symbols (symbols) can be transmitted in one second at most ).
- Symbol rate ( symbol rate , or code transmission rate , or symbol transmission rate), unit sps (symbol per second), symbol rate is equal to the baud rate, that is, the symbol rate (transmission rate) also corresponds to t1 in the figure.
- Bit rate (bit rate), unit bps (bits per second), bit rate is very close to symbol rate (transmission rate), like the situation in the picture, you can think that bit rate = symbol rate, because a symbol in the picture represents a bits. But there are many systems, one symbol can represent multiple bits, such as 4PSK (QPSK), one symbol represents 2 bits, so bit rate=2*symbol rate. There are also many systems where multiple symbols represent one bit, such as BLE coded PHY, where 8 symbols represent one bit. At this time, bit rate = symbol rate/8. Sometimes, the bit rate can also be used to represent the effective bit rate. For example, in the above figure, only 8 bits of data are actually transmitted during the entire t2 time (the other 2 bits of data are invalid data), so you can think that bit rate = 8/10 symbol rate
- Data rate , the unit is bps (bits per second). The specific connotation of data rate depends on the context. Sometimes data rate = symbol rate, and sometimes data rate = bit rate.
- Throughput rate (data throughput), in bps or Bps (bits/bytes per second), the throughput rate measures a relatively long period of time, and is the maximum "speed" that the user actually feels, just like t3 in the above figure, the t3 time All valid bytes transmitted divided by t3 is the throughput rate. When calculating the throughput rate, not only the redundant bits like start bit/stop bit should be deducted, but also the delay time between bytes and bytes, and the delay time between blocks, etc., so Throughput is much "slower" than data rate.
- Bandwidth . Bandwidth has multiple connotations. In the field of computers and networks, bandwidth is basically equivalent to throughput rate, and the unit is bps or Bps (bits/bytes per second). Please refer to the description of "throughput rate" above. In signal processing/RF/microelectronics, bandwidth is the difference between two extreme frequencies, as shown in the figure below, where B is the bandwidth .
The difference between the limit frequencies is the original meaning of bandwidth, and throughput rate can be said to be an extension of bandwidth. Why can bandwidth be extended to mean "throughput rate"? This is derived from the Nyquist theorem. In 1924, Nyquist deduced the limit transmission rate of a noise-free channel with limited bandwidth, which is called the Nyquist theorem. If the channel bandwidth is W (unit Hz), the maximum symbol rate B (unit: sps or baud) of Nyquist theorem is:
B=2W
It can be seen from the above formula that the bandwidth directly determines the maximum code transmission rate of the channel, so it is reasonable to use the bandwidth to represent the throughput rate.
Note: The Nyquist theorem is actually another expression of the Shannon sampling theorem, and the two are equivalent.