The Cost of a High Sampling Rate

Although it is true that high sampling rates produce better sound quality … that comes at a price.

That price translates into:

Higher processing load.
Less number of tracks.
Heavier audio files.
So you always give something in return. Professional studies can support higher sampling rates because they use better equipment.

But for most home studios, people often find that the standard 48 kHz configuration is the best.

Following…

4. Bit Depth

In order to understand what bit depth is, we first have to know what bits are.

A bit (or binary digit) is a single unit of binary code, with a value of 1 or 0.

The more bits, the more possible combinations. For example…

As you can see in the diagram below, 4 bits allow a total of 16 combinations.

4 bits

When used to encode information, each of these numbers is assigned a specific value.

As the number of bits increases, the possible values ​​grow exponentially.

4 Bits = 16 possible values
8 Bits = 256 possible values
16 Bits = 16,536 possible values
24 Bits = 16,777,215 possible values
With the bit depth in the digital audio, each value is assigned a specific amplitude of the waveform.

The greater the bit depth, the greater the volume increase between high and low … and a greater dynamic range in the recording.

A good rule of thumb is: for every extra bit, the dynamic range increases by 6dB.

For example:

4 Bits = 24 dB
8 Bits = 48 dB
16 Bits = 96 dB
24 Bits = 144 dB
In the end, what this means is that… the greater the bit depth, the less noise.

Because by adding more processing margin (or headroom), the useful signal (at the high end of the spectrum) can be recorded higher above the background noise (at the low end of the spectrum).

small vs large bit depths

Following…

5. Quantization Noise

Impressive that a 24-bit recording can result in almost 17 million possible values, right?

However, that remains much less than the infinite number of possible values ​​that exist in an analog signal.

Therefore, in almost all samples, the actual value is somewhere between two possible values. The solution of the converter is simply to round it or “quantify” it to the nearest value.

The resulting distortion, known as quantization noise, takes place in 2 phases of the recording process:

at the beginning, during the A / D conversion, and
at the end, during mastering
With mastering, the sampling frequency / bit depth of the final track is usually reduced by converting to the final digital format (CD, mp3, etc.).

When that happens, some of the information is erased and “re-quantized”, generating more distortion in the sound.

The most frequent solution to deal with this problem is …

6. Dither

When reducing a 24-bit file to 16 bits, the screen is used to mask much of the resulting distortion …

Adding a low level of “random noise” to the audio signal.

As it can be difficult to visualize the concept in audio, to explain it, we usually turn to the popular analogy of the screen plot.

Is that how it works:

When a color photo is converted to black and white, a mathematical estimate is made to determine if each color pixel should be “quantized” in a black pixel, or a white one …

As is the case when digital audio samples are quantized.

As you can see in this picture, the “before” photo is pretty bad, right?

dither

But with the plot …

a small number of white pixels are randomly distributed in black parts, and …
a small number of black pixels are randomly distributed in white parts …
By adding that “random noise” to the image, the “after” photo looks much better. Well, the screen in the audio works very similarly.

Following…

7. Latency

The GREAT PROBLEM of current digital studies is the amount of latency that accumulates in the signal chain, especially with DAWs.

With all the calculations that are processed, the audio signal takes time to leave the system between a few milliseconds and a few DOCENAS of milliseconds.

Between 0-11 ms of latency – it is short enough, so a normal person does not notice it.
Between 11-22 ms – an annoying delay is heard which it is difficult to get used to.
More than 22 ms – there is so much delay that it is impossible to play or sing at tempo with the track.
In a normal digital signal chain there are usually 4 phases that contribute to the total latency:

A / D conversion
DAW Buffer
Delay of the Plugins
A / D conversion
The A / D and D / A conversion are the least harmful, contributing to total latency with less than 5 ms.

But nevertheless…

The DAW buffer and certain plugins (including compressors and virtual instruments) can add up to 20, 30 or 40 ms or