# What is signal resolution?

I'm trying to understand signal resolution, so therefore I am looking for a more clearer explanation. Primarily signal resolution is linked to an Analog-Digital converter.

Quoting from new electronics, http://www.newelectronics.co.uk/electronics-technology/whats-the-difference-between-resolution-and-accuracy/73740/, states: “The resolution of an A/D converter is defined as the smallest change in the value of an input signal that changes the value of the digital output by one count.”

In addition, I have read many more unclear definitions. Specifically, what is signal resolution?

• Welcome to EE.SE. What is not clear? Granted that those statements need the detailed specs of a given ADC in order to make sense. Resolution of the ADC itself is the number of bits it will output per each conversion. But each bit has a voltage 'weight', with the LSB having the smallest.
– user105652
Oct 18, 2018 at 19:55
• The resolution is, essentially, the difference between the smallest input that produces an output of 00000001 (assuming an eight-bit ADC), and the smallest input that produces an output of 00000010. It's the size of one "step" of the output. Oct 18, 2018 at 19:55

"Signal resolution" refers to how detailed the waveform is allowed to be. Analog signals (red line) have an infinite amount of points that can be attained between a "high" and "low" - this is a reflection of the real world, since we can always find an infinite amount of points between two finite points.

The problem comes when you need to store this data digitally. You can't give a computer an infinite amount of points, since it would require an infinite amount of memory. Therefore we choose a certain "resolution", or amount of attainable points between "high" and "low" in order to describe the waveform.

In the sense of an A/D converter, we need to represent this "infinite wave" in memory. We can use an 4-bit number (for example) to represent the points between "high" and "low". This would result in 16 possible (vertical) points - see the figure below. If we chose a 10-bit number, it would result in 1024 different possible points, and thereby a greater signal resolution.

“The resolution of an A/D converter is defined as the smallest change in the value of an input signal that changes the value of the digital output by one count.” - They are simply saying here that, if the A/D converter has a low resolution, then the red line needs to change quite a bit before the black line will jump to the next point vertically (in memory).

Note: "high" and "low" simply refer to the max and min possible values of the signal.

To keep it simple: resolution is a number of steps an ADC can produce.

Example: 8-bit resolution means that the ADC can sense 256 levels (2 to the power of 8). If the reference voltage is 1V it means that the ADC can roughly sense a voltage difference of 1/256 V = 0.0039 V.

Another analogy: it is the number of shades of gray the ADC can see.

• Notice the error-areas, where the steps are not exactly the sinusoid. If the input is a ramp, you will observe a repetitive pattern for any given slewrate. This repetition will be a TONE in the error, and will appear in the DAC_constructed output. This is one problem with digitizing MUSIC ---- error tones appear. Oct 19, 2018 at 4:08
• @analogsystemsrf That is what dither is for (you always need it if your objective is to have a linear system), it converts a stepwise approximation to a linear system with noise. The other source of spurious tones is limit cycles in DSMs, but again this has well known fixes. Oct 19, 2018 at 7:07