Which sound recording system do you think has the best sound quality?

[Mgeterno11]

As the title says, let's discuss, which sounds recording system has the best quality?

 

[Zwiebelchen]

This isn't something to discuss, just a matter of facts. From the perspective of audio fidelity, the CD takes the clear victory.

Tape and Vinyl have their places for purists who like the sound coloration that the carrier and (much more importantly) the playback devices produce, but that's it. Applying only physics here, there is absolutely no comparison.

 

[Dr Super Good]

From the perspective of audio fidelity, the CD takes the clear victory.

Not necessarily since it suffers from discretization. LP is purely analogue and Cassette can also be analogue. The discretization limits the accuracy in wave shape reproduction and in theory introduces high frequency artefacts unless filtered before playback. Additionally CDs are very limited as to the format and quality of audio they store where as the others are potentially less limited (although in reality they are likely worse).

In theory purely digital is the best recording system. Some professional sound capture devices can sample up to 200,000Hz, giving a huge range of non-audible frequencies for attenuation of the anti-aliasing filter. Sound samples can be 24 or even 32 bit for even higher precision. Sub 20 Hz and post 20,000 Hz frequencies can be perfectly filtered using DFT filters. The results can then be accurately resampled into more appropriate playback forms. The resulting sound file can then be lossy compressed for better distribution.

 

[Zwiebelchen]

Not necessarily since it suffers from discretization. LP is purely analogue and Cassette can also be analogue. The discretization limits the accuracy in wave shape reproduction and in theory introduces high frequency artefacts unless filtered before playback. Additionally CDs are very limited as to the format and quality of audio they store where as the others are potentially less limited (although in reality they are likely worse).

Both 16 bit in dynamic range and 44.1 Khz (which fits the Nyquist theorem) is above the threshold of what humans can destinguish by ear.
So, no, nobody is able to hear a difference except for the caviar-smokers. Discretization isn't a problem with 16 bit either, as 16 bit of dynamic range can still go beyond the noise floor of most conventional PA equipment.
24 bit will only be used during the production process to give extra headroom for mixing and applying effects. Once the mix is done, you will convert back to 16 bit, since the extra 8 bits are useless (if you are a human, that is).
The reason why we see recording companies distribute more and more 24 and even 32 bit recordings is because gullible people pay more for that.

Analogue devices suffer from physical limitations that CDs don't have. For example, Vinyl can not accurately reproduce low frequencies beyond 120Hz. Also, the Needle will be affected by natural oscillation and inertia, which also limits the high frequencies to somewhere around 14-16khz.

The reason why conventional CDs sounds worse than for example a 24 bit lossless digital audio file is not because that is what CDs allow, but because what we pressed on the CD. The production standards of the last 30 years summed all audio signal lower than 200 Hertz to mono, because most people will listen to CDs at home and conventional listeners don't have a PA system (nor a sound environment) that can actually reproduce stereo bass.
Same goes for the EQing and general mixing. Mixes got intentionally compressed way more than what was required by the audio medium simply because the 08/15 consumer did not have the gear nor personal interest to listen to music at a high dynamic range.

And you would think that this trend gets better with modern technology. Nope. People get dumber every year. Most consumers nowadays are fine listening to music on crappy laptop or phone speakers. Most consumers will "turn up the bass" way beyond what their gear allows or what sounds good just to be the cool guy who turns up the bass.

Whenever the discussion goes about audio fidelity and bit depth, it's good to apply a reality check every now and then and see for whom most music gets made for.


Music is dead.

 

[Dr Super Good]

Both 16 bit in dynamic range and 44.1 Khz (which fits the Nyquist theorem) is above the threshold of what humans can destinguish by ear.

Except you need an anti-aliasing filter, an analogue filter which is far from perfect. This is why 44.1 kHz was traditionally used since it gives 2.05 KHz for anti-aliasing filter attenuation. Additionally you can sample at 192 kHz and use a much weaker (cheaper) anti-aliasing filter (56 kHz spare) followed by digital filtering (which can have near perfect response). This probably means that the 192 kHz sampling approach is cheaper and more accurate than the 44.1 kHz approach when it comes to sampling audible sound.

24 bit will only be used during the production process to give extra headroom for mixing and applying effects. Once the mix is done, you will convert back to 16 bit, since the extra 8 bits are useless (if you are a human, that is).
The reason why we see recording companies distribute more and more 24 and even 32 bit recordings is because gullible people pay more for that.

Mostly they would be used for audio processing. If one is manipulating or processing the wave forms one wants them as high precision as possible so that the discretization error does not get compounded through your processing. This is why 24bit is used for that "extra headroom".

The reason 32bit is used is most likely purely for convenience from a programming perspective because you have no 24bit primitive type in most languages because most processors do not have instructions to process 24bit types. As such a 24bit audio sample will be processed most efficiently if stored in 32bits and while using 32bits one might as well take advantage of the free 8 bits of precision. For example to process a 24bit sample some instruction sets might have to process 3 separate byte memory read instructions and then merge the result into a 24 bit sample (stored in 32bits) where as they could execute a single 32bit memory read for a 32bit sample. Digital audio playback systems should not have this problem as they would be designed to accept 24bit sampled data (likely determined by a configuration register). I know this is kind of off-topic lol.

Analogue devices suffer from physical limitations that CDs don't have. For example, Vinyl can not accurately reproduce low frequencies beyond 120Hz. Also, the Needle will be affected by natural oscillation and inertia, which also limits the high frequencies to somewhere around 14-16khz.

However you then go on to say that audio equipment cannot output the full range of sound...

The reason why conventional CDs sounds worse than for example a 24 bit lossless digital audio file is not because that is what CDs allow, but because what we pressed on the CD. The production standards of the last 30 years summed all audio signal lower than 200 Hertz to mono, because most people will listen to CDs at home and conventional listeners don't have a PA system (nor a sound environment) that can actually reproduce stereo bass.
Same goes for the EQing and general mixing. Mixes got intentionally compressed way more than what was required by the audio medium simply because the 08/15 consumer did not have the gear nor personal interest to listen to music at a high dynamic range.

So with the obvious exception of mono vs stereo, both could potentially sound the same to a user because any imperfections are lost due to the quality of his loudspeakers.

And you would think that this trend gets better with modern technology. Nope. People get dumber every year. Most consumers nowadays are fine listening to music on crappy laptop or phone speakers. Most consumers will "turn up the bass" way beyond what their gear allows or what sounds good just to be the cool guy who turns up the bass.

The main problem is that a lot of consumer level audio devices are not properly engineered. Sure some engineers design them, but they clearly are not good engineers. Practically every loudspeaker, headphone and microphone lacks a proper technical specification document which documents the frequency response of the input and output (because analogue equipment is not perfect and so has natural filter qualities). In theory one would use such a document to setup mixers such that all frequencies are output at the intended levels. However trying to find what range of frequencies your speakers even support can be nightmare, let alone the frequency response graph.

This is in stark contrast to electronic engineering. Even something as simple as a £0.10 transistor has a specification document showing the frequency responses, temperature responses, voltage responses etc. How can a £50-£200 sound system not?

Whenever the discussion goes about audio fidelity and bit depth, it's good to apply a reality check every now and then and see for whom most music gets made for.

A lot of the audio processing technology used by the music industry might not be purely intended for human use. There are situations where people want to capture infrasound/ultrasound or very high dynamic range sound, especially when it comes to animal research or audio analysis. The music industry might just be using a small subset of the capabilities of such equipment and hence why the fidelity might seem pointless.

 

[hypsandar]

I'd recommend AudaCity
You can record and edit the sound in depth in it.
Removing all noise reductions for example.
I find that very useful when making some video types.

Unless you meant both screen and sound recording, than i'd recommend using OBS

(Both programs are free to use)