Have you ever considered what it takes to digitize text and sound?
In terms of text, it is quite simple and fortunately there are not many tradeoffs. Since written language is already symbolic it can be accurately digitized.
For example, if you are typing in text, each letter can be mapped to a numerical code known as ASCII or Unicode. Unicode being the universal code that is able to represent characters from all languages and symbols. ASCII is a subset and only represents English characters, amounting to 128 characters that can be used.
These numbers are then able to be stored quite simply by the computer in binary; as the digital world is relies on 0s and 1s.
The challenging part, and where tradeoffs occur, is when text is handwritten or printed. Scanners with character recognition can identify shapes into characters, however, due to the analog nature of human writing errors are more likely to occur.
There is a greater chance of characters being misread and misidentified if someone has hard to read handwriting or if an odd font was used for print.
When it comes to storing text, plain text can be stored in tiny files with the loss of their format. Whereas file types like DOCs or PDFs can preserve the layout of text with an increased file size.
Considering all these issues, digitizing text is almost lossless but relying on character recognition when it comes to print and handwritten text always introduces the potential for errors.
Digitizing sound is much more complex. Sound waves are traveling compression waves, they are the pressure through air; with compression the air is bunched together and refractions are where the air is missing. Additionally, sound is impacted by more variables and circumstances than text, singing has a more continuous sound pattern than talking.
There is also the analog aspect of continuous variations in voices. When applied to the concept of telegraphs or telephones, there must be an interval decided for when to measure the sound wave. In this interval air is being compressed and expanded, while this process occurs, noise electrons are also being created; these electrons give off heat.
When these sound waves are run through an analog amplifier in the hopes of amplifying the voice and making it louder, the noise is being amplified as well.
The longer the interval, the more detail you can capture but, the more noise will be picked up.
In digitizing sound you must sample and precisely measure the sample. The great advantage digital has is the ability to separate signal from the noise. A digital telephone is still picking up the noise an analog telephone/amplifier would, yet it has the ability to get rid of the noise and keep the signal.
Since binary is still being used to represent the waves in the digital world, the more numbers found in a sample leads to the waves being better represented. This means sampling should be done more often, especially because it is always possible to increase the quality of a digital signal!
Circling back to the tradeoffs, for digital sound you need a certain amount of numbers to play music. As an example, if you need 1000 numbers, the music will stop and will not continue until it gets the required amount of numbers; so it needs to be transmitted quickly.
It can take a long time to get those numbers and there is a very delicate balance in measuring the tradeoff between how many numbers are needed to represent something and the quality of what is being represented.
In order to increase the quality, the amount of numbers needed must be increased.
One last note to recap, it truly is impossible to achieve perfect digitization for analog sound because it is infinite in detail. Variables like the hearing limits of humans, bandwidth, and storage impact the balance sound waves are being digitized in. Even an object’s intended use affects the digitization process, on a cellphone the priority is to ensure voices are clearly heard which is why hold music is almost never at a good quality.