Bit Depth and Sample Rates

[Jason Willey]

On my computer it goes from 16-bit 44100 Hz to 24-bit 192000 Hz

 

I am only connecting via the headphone out jack to RCA into my speakers

 

Not doing any studio work but do play games, films and internet streaming services as well as MP3s

 

Wish Linus would do a video about this and what it all means, which setting should you have it on and would it actually benefit sound quality?

[mariushm]

Go with : 24 bit  48000 Hz

 

Music from audio CDs is 16bit 44100 Hz

Music from DVDs, Blurays etc is 16 bit or 24 bit and 48000 Hz

 

Sound can be converted with practically no loss from 16 bit to 24 bit. Resampling from 44100 to 48000 is fairly fast and the loss of quality is practically not noticeable.

In general, 48000 Hz is going to be used more than 44100 so it's better to set the default to this - windows will convert anything else to this format.

Higher than 24bit (ex 32bit float) would make sense only if you'd create music or edit movies with lots of audio tracks and overlapped sound effects .. for any other reason there's no benefit.

More than 48000 Hz also doesn't make sense - 48000 Hz means sound card and speakers receive a waveform that can reproduce sounds up to 24000 Hz when in real world human ears can perceive sounds only up to 18000 Hz, maybe up to 20000 Hz when you're very young or if you're a dog.  So more than 48000 Hz won't improve sound quality.

[SSL]

4 hours ago, mariushm said:

So more than 48000 Hz won't improve sound quality.

 

It'll actually degrade it. The Windows Direct Sound upsampling filter is very low quality and causes audible roll-off when upsampling to higher rates.

[Jason Willey]

my onboard sound card defaults to 24 bit 48000 Hz

 

what do games and Netflix and Amazon Prime output to?

[rgemmell82]

Unless you are specifically listening to 24 bit 48/96 kHz (48,000 and 96,0000 Hz) files, then 16 Bit 44.1kHz (44,100Hz) will suffice. Almost all content for streaming is compressed and increasing the playback quality won't give you any more fidelity than what the audio source encoded bit/sample rate. 

The quick and dirty on Bit depth, bit rate, and sample rate:
Bit Depth is more in terms of recording for signal to noise, dynamic range, along with other elements. 

Sample Rate is the frequency rate of the recording. Human hearing covers 20 Hz to 20,000 Hz (20kHz). 44.1 kHz refers to the 44,100 samples captured per second. This is considered as the basic coverage for the hearing spectrum. Doubling the amount of samples was required to achieve usable fidelity.
Bit Rate is the data rate that we are familiar with. 128/192/360+ kbits/s are the common ones that are used. Different encoders will yield different results with the same bit rate based on their algorithms. For perspective: CDs (16 Bit) bit rate is 1,411 kbits/s at 44.1 kHz sample rate in Stereo. 

 

I'm glossing over a lot of information, but I didn't want to overload with information.

[Pomfinator]

For Windows DirectSound, I recommend the following.

 

• Set your bit-depth to 24-bit or higher, this means that if you use computer audio to control volume, the bits that are shaved off to lower the volume does not effect any regular 16-bit music.
• Set your sampling rate to 44.1khz, while 48khz is more used in video and gaming, resampling does not effect the experience nearly as much as listening to music.  This means that any 44.1khz file (most music) will simply be sent to the DAC without any resampling

[anothertom]

32 minutes ago, Pomfinator said:

• Set your bit-depth to 24-bit or higher, this means that if you use computer audio to control volume, the bits that are shaved off to lower the volume does not effect any regular 16-bit music.

That is so wrong it's funny.

 

All you're changing is the resolution with which windows can output the voltage range. You don't get an extra 8bits of volume to chop off before you start affecting the signal because that would mean the volume would reach it's maximum level when it's set to 42.5% (if we take a single bit to determine positive or negative you get a maximum of 15^2 / 23^2 =0.42533...).

 

All of the steps present with 16 bit resolution are present in 24bit resolution, you may get slightly better accuracy at small volume percentages but not that you would be able to notice or measure unless you'be got some very accurate and precise equipment. You are simply wasting the energy it takes to convert it, albeit a very small amount.

[seduce_me]

Use ASIO or WASAPI if your DAC supports it, else set it to the bit depth and rate of your music. Basically, windows resampler sucks so it's best to not let it do anything.

 

Sound quality wise, it doesn't matter much because your speakers will affect it way more.