Resolution (digital technology)
In digital technology , the resolution indicates how finely an originally analog variable can be digitally represented, which is digitized with an analog-digital converter before further processing .
Discretized sizes
The resolution here can relate to different dimensions:
- the signal level , i.e. the measurable intensity of the size (e.g. the color depth or volume ),
- the spatial distance (e.g. the image resolution ) or
- the time interval (the sampling rate ).
- The limited resolution of the signal level results in so-called quantization deviations .
- If the temporal or spatial resolution is too low, aliasing effects arise .
The digital signal processing uses the term resolution in connection with the quantization in the sense of pitch . This is in accordance with the basic standard for measurement technology, which defines the term resolution as a quantitative indication of how far a measuring device can clearly distinguish between measured values that are close together.
The relationship between the stepless input signal and the stepped output signal is described by the quantization curve. With a linear quantization curve, the step size or width of a quantization level is constant. It results from the input signal range and from the number of levels or from the number of digits of the output value.
- Examples
- A measuring range of 0… 200 mV is resolved in 2000 steps. Then the step size is 0.1 mV.
- With a linear 8-bit ADC the relative step size corresponds to as much as one step in relation to all steps , the signal can then be resolved in steps of around 0.4% of the quantization range.
For certain applications (e.g. voice or image transmission) it can be advantageous to use a non-linear characteristic. The step size depends on the input value and can be different for each interval.
Applications
Sound engineering
The common term used for sound cards and audio software is simply "resolution". This is specified for the volume by the number of binary digits and for the time range by the sampling rate. For example: "16 Bit / 48 kHz".
Until about 1995 most sound cards worked with a resolution of 8 bits per sample , which meant that a slight background noise was still perceptible. With audio CDs and more modern sound cards, 16 bits per channel are now common; for audio and video DVDs up to 24 bits. In ISDN telephony, the analog input signal is sampled with 8 bits per sample, with the peculiarities of human perception being taken into account during the quantization .
Many music production programs work with 32-bit samples, but these can only be used with equipment that allows appropriate dynamics (e.g. microphones, amplifiers, loudspeakers, rooms).
Computer graphics
Computer graphics are broken down into three dimensions:
- the image resolution, i.e. in length and width,
- the color depth or, in the case of gray value images, the brightness .
In the case of a gray value image, a resolution of 8 bits is sufficient to obtain a natural-looking image with the 256 shades that result. A finer resolution is only necessary if the contrast is to be changed significantly later in order not to falsify the image result.
With color images, 256 colors lead to inadequate, grainy or comic-like images, so that nowadays each of the three color channels (red, green, blue) is usually resolved with 8 bits.
When it comes to image resolution, in addition to signal resolution, the resolving power must also be taken into account, the ability to detect objects that are close together as independent.
Video technology
When recording digital videos, the individual images are resolved as described in the Computer Graphics section. In addition, as with audio technology, there is a temporal resolution when scanning the images, the image frequency .
Individual evidence
- ↑ John G. Proakis, Dimitris G. Manolakis: Digital Signal Processing . 3. Edition. Prentice Hall, 1996, ISBN 0-13-394289-9 , Chapter 9.2, pp. 750 ff .
- ↑ DIN 1319-1 Fundamentals of measurement technology - basic terms . 1995.
- ↑ Thomas Waldraff: Digital image resolution . Springer, 2004.