• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

44100hz 16bit vs 44100hz 24bit - 16bit causes White/Static noise..

daftcombo

Major Contributor
Forum Donor
Joined
Feb 5, 2019
Messages
3,477
Likes
3,497
DidBd you miss the part about pausing the music and cranking up the gain? Lack of dither wouldn't affect silence.

Perhaps I missed something said by the OP.
Anyway, adding dither with a triangular processing will add noise way below -96 dB so I don't understand how someone would notice it. On the other hand, converting from 24bit to 16bit with noise shaping can give artefacts in the audible band, though very high, at 16 kHz or so. It can be heard more easily since it can be at -80dB or so. But it doesn't sound like a pink noise, it sounds like high pitched noises when you hear them. With bad amp/speaker, those can create IMD down the chain and become very audible. For instance, I can hear such things with PC speakers, but not with Aria 906 powered by an Ncore amp.

Sticking to 24bit is the way to go anyway.
 

mansr

Major Contributor
Joined
Oct 5, 2018
Messages
3,727
Likes
8,410
Location
Hampshire
Anyway, adding dither with a triangular processing will add noise way below -96 dB so I don't understand how someone would notice it.
Try the file I posted. If you turn up the volume enough, you will probably hear a faint hiss. Just remember to turn it back down before playing any music, or you might blow up your speakers and/or ears.
 

Vincent Kars

Addicted to Fun and Learning
Technical Expert
Joined
Mar 1, 2016
Messages
530
Likes
932
As Toslink is SPDIF over optical and SPDIF sends a 32 bit data packet containing a 8 bit preamble and 24 bit audio, best bet is to set Win to 24 bit anyway.
 
OP
A

all-air

Member
Joined
Feb 23, 2021
Messages
18
Likes
1
Location
Sweden
why not just keep it at 24 bit then?

do you hear the same noise when you use a headphone jack from another device? ( or the motherboard even) this would quickly determine whether you're hearing the 16 bit noise floor or not.

Just tested, my Beyer T5p 32ohm, fairly sensitive, couldn't pick up any noise through the motherboard back or front (chassi) 16bit... Only with DAC connected and amplified signal with an amp (gain etc).
Guess the motherboard can't amplify the noise (dither?) like a dedicated amp

Here's a 24-bit FLAC with 16-bit TPDF dither noise. If you play it with settings at 24-bit, does it sound the same as the noise you're hearing with the 16-bit setting?

Yes, it sounds exactly like that! The very same noise that I get when I use16bit with an external amp & dac :)
 
OP
A

all-air

Member
Joined
Feb 23, 2021
Messages
18
Likes
1
Location
Sweden
So you have the level meter right in front of you and all you have to do is watch it. Read the (peak) values in the 'pause/stop' mode. If you can hear it then you can measure it. If you can measure it (the level meters showing signal) you know that the Asus' digital output spits out such sounds in 16 bit mode.

Not sure I understand what you mean (I'm a novice sorry), but it seems my PC does not registre the noise/static
 

scott wurcer

Major Contributor
Audio Luminary
Technical Expert
Joined
Apr 24, 2019
Messages
1,503
Likes
2,766
Anyway, adding dither with a triangular processing will add noise way below -96 dB so I don't understand how someone would notice it.

No actually ordinary TPDF dither increases the noise floor above -96dB, still IME inaudible with ordinary listening. Turning up the volume during quiet or silent passages to hear the noise floor has little practical use.
 

daftcombo

Major Contributor
Forum Donor
Joined
Feb 5, 2019
Messages
3,477
Likes
3,497
No actually ordinary TPDF dither increases the noise floor above -96dB, still IME inaudible with ordinary listening. Turning up the volume during quiet or silent passages to hear the noise floor has little practical use.
If you take a 24bit music file and dither it down to 16bit, there is no way noise can become higher than -96dB. That's the purpose of dithering.

Edit: I meant "a lot higher".
 
Last edited:

MrPeabody

Addicted to Fun and Learning
Joined
Dec 19, 2020
Messages
657
Likes
844
Location
USA
If you take a 24bit music file and dither it down to 16bit, there is no way noise can become higher than -96dB. That's the purpose of dithering.

Since quantization noise cannot be higher than -96 dB with 16-bit quantization, the only apparent reason to do what you are describing would be to change the spectral makeup from whatever it is with quantization noise to something else. And in order for any noise at -96 dB to be audible, you have to be listening at a volume setting where the loud peaks recorded at the highest possible encoding level are very loud, and then be listening to a very quiet passage with the volume still turned up that high, and even then, the noise would be at the threshold of audibility, which means that you'd need to be in a setting with no background noise. No fans or compressors, no trains or roaming automobile subwoofers off in the distance, and most importantly, no cicadas crawling out of their holes in the ground and then climbing up in the trees. And speaking of trees, no wind.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
1,724
Likes
1,593
Location
.de
If you can plainly hear dithering noise in 16 bit mode, over headphones no less, amplifier + transducer gain is very likely to be excessive. It would be interesting to know what the sound device and amplifier volume settings are at normal volume. I would advise ca. 70% for the former and about 10 o'clock for the latter; reduce amplifier gain setting as needed.

Normally you can get by fine with 16-bit output if there's an amplifier with a volume control of its own following; requirements tend to be far greater if you have to accommodate a large variety of headphones with fixed analog stage gain. The Asus Xonar ST, for example, became known for exhibiting increased output noise in high sensitivity headphones when switching hardware sample rate to 44.1 kHz (which reduces loopback dynamic range to a "mere" 108 dB(A), still almost 18 bits effective). If you have somewhere on the order of 6 Vrms worth of maximum output as that card does, it's a bit hard to get away with less than about 120 dB(A) of dynamic range.

There is relatively little reason not to use 24-bit output these days. If nothing else, you can provide extra headroom for EQ without losing any dynamic range on 16-bit material (or any material, for that matter - even the most dynamic acoustic recordings top out at about 90 dB).
 

scott wurcer

Major Contributor
Audio Luminary
Technical Expert
Joined
Apr 24, 2019
Messages
1,503
Likes
2,766
If you take a 24bit music file and dither it down to 16bit, there is no way noise can become higher than -96dB. That's the purpose of dithering.

No this is wrong, the dither needs to be applied to the 16 bit result 1 LSB of TPDF dither at 16 bits raises the noise floor above -96dB (slightly). Try it yourself, generate a 24 bit -80dB sine wave with proper 24 bit dither and down sample it to 16 bits with all dither disabled.
 

daftcombo

Major Contributor
Forum Donor
Joined
Feb 5, 2019
Messages
3,477
Likes
3,497
No this is wrong, the dither needs to be applied to the 16 bit result 1 LSB of TPDF dither at 16 bits raises the noise floor above -96dB (slightly). Try it yourself, generate a 24 bit -80dB sine wave with proper 24 bit dither and down sample it to 16 bits with all dither disabled.

It should be inaudible, which could be different without dither. A bit higher than -96dB (you are right) with dither, but near -80dB in hi freq without.
 
Last edited:
Top Bottom