从“m”个字节中获取“n”个位数

Question

I have a problem with extracting specific amount of bits from specific bytes. Lets say we have case like:

(I assume that I start counting bytes/bits from 0) I receive 3 bytes from stream and want to read 11 bits starting from 1st byte's 7th bit: finishing at 2nd byte's 5nd bit, and I want to store it somewhere

Some visualisation:

     0th byte          1st byte         2nd byte
|7|6|5|4|3|2|1|0|#|7|6|5|4|3|2|1|0|#|7|6|5|4|3|2|1|0|
                   ^ ^ ^ ^ ^ ^ ^ ^   ^ ^ ^

I know that bit shifting is a direction to go, but I lack experience in that topic, and Im really curious on the opinion from experienced developers

EDIT: Sorry for not specifying it before - the language is C++, but I cannot use libraries like "bitset".

The "n" and "m" can be any number, eg.: n = 17 m = 5. Another assumption is that data will need to be continous - it means that in the end I need to have eg 11 bits of continous data.

What I was trying to do was something like this:

buf[0] << ((byteCount - 1) - index) * stepWidth;

Which gave me full first/second byte. In the end readed bits should be "merged" as one variable, somehing like:

int buf[1] = variable with 8 bits;
int buf[2] = variable with 3 bits;
mergedBuf  = buf[1] + buf[1] // "+" means merging, not adding values

Answer 1

C++ has nice class in standart library:

https://www.cplusplus.com/reference/bitset/bitset/

If C is the case, you can use bit mask and & (and) operation. like:

unsigned char buf[3]; // let's say this is your buffer;
int _2nd_byte_value = buf[1] & 0xFF; // since you want to get all bits of 2nd byte;
int _3rd_byte_value = (buf[2] & 0xE0) >> 5; // first mask then shift; 

Then you can manipulate these values as you wish.