When representing signed binary values in octal or hexadecimal, we pay no attention to the binary format and just convert the bits as we would with an unsigned binary number. Hence, we never see a '-' sign with octal or hexadecimal. The '-' sign is only used to indicate negative decimal values.

Hence, we can think of octal and hexadecimal as compressed representations of binary and nothing more. If given an octal or hexadecimal number, we must also be told what the binary format is in order to determine its value.

Likewise, to convert a signed decimal value to octal or hexadecimal, we must be told the binary format and number of bits. -7₁₀ is 11111000_{2, one's complement} = F8_{16, one's complement, but} 11111001_{2, two's complement} = F9_{16, two's complement}.

The main caveat is that we must be aware of how many bits the value contains in case it differs from the number of digits shown. Whatever octal or hex digits are shown are to be converted to binary. Any digits not shown are assumed to be 0. The examples below should clarify why this is important.

177_8 in 7-bit two's complement     = 1 111 111     = -1_10
177_8 in 8-bit two's complement     = 01 111 111    = +127_10
377_8 in 8-bit two's complement     = 11 111 111    = -1_10
377_8 in 16-bit two's complement    = 0 000 000 011 111 111 = +255_10

9F_16 in 8-bit two's complement     = 1001 1111     = -(01100000 + 1) = -97_10
9F_16 in 8-bit bias-127             = 1001 1111     = +32_10
9F_16 in 8-bit one's complement     = 1001 1111     = -01100000 = -96_10
9F_16 in 16-bit two's complement    = 0000 0000 1001 1111   = +159_10
9F_16 in 16-bit one's complement    = 0000 0000 1001 1111   = +159_10

1001000_2 7-bit two's complement    = 48_16 = 110_8
1001000_2 7-bit unsigned binary     = 48_16 = 110_8
1001000_2 7-bit bias-63             = 48_16 = 110_8