• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

16-bit... It really is enough!

Paianis

Member
Joined
Nov 2, 2018
Messages
25
Likes
19
I agree on that. I personally wish they picked 48K for Redbook sampling rate. This makes the mixdown from 96K 24-bit really simple and super accurate.
I suspect 48kHz wouldn't have permitted Beethoven's 9th to fit on a 12 inch disc.

44.1kHz is fine. The best resamplers available today are practically transparent.
 

Chromatischism

Major Contributor
Forum Donor
Joined
Jun 5, 2020
Messages
4,765
Likes
3,703
I agree dithered 16 bit is fine for playback. 44.1kHz is cutting it a bit fine, though, and definitely (strictly speaking) won't reproduce the full audible range for some, mostly younger, listeners (if this is captured in the recording in the first place).*

Personally, I see 16/44.1 as perfectly adequate for my own ears, which don't hear much beyond 17kHz.

*EDIT: and requires very steep anti-aliasing and reconstruction filters, even for listeners with just average hearing.
Interesting. So we should be more picky with our DAC choices if we have a lot of 16/44 content?

I get this is speaking technically. I'm not sure we'd notice the difference though. Would we?
 

Frank Dernie

Master Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
6,445
Likes
15,780
Location
Oxfordshire
Not to kick a dead horse, but while I totally agree with 44.1 (48) kHz for sampling, I think 24 bit still makes practical sense. While, 16bit is perfectly adequate dynamic range for "from below-audible noise floor to the loudest", it does not leave much room for mastering error/sloppiness. While the 24bit dynamic range is much more forgiving... IM-audio-amateur-HO.
This made me smile.
When I went from analogue tape to digital (DAT 48/16) it was the ease of setting levels which was the big shock. With tape, getting a balance between tape overload and audible noise was a bit of a fiddle on wide dynamic range music, on 16-bit digital it was a doddle, with 24 bit my dog would be as good at setting levels as an expert!
It leaves more margin for dicking about mixing multi track though.
 
OP
M

MusicNBeer

Senior Member
Forum Donor
Joined
Jan 20, 2019
Messages
333
Likes
478
Not to kick a dead horse, but while I totally agree with 44.1 (48) kHz for sampling, I think 24 bit still makes practical sense. While, 16bit is perfectly adequate dynamic range for "from below-audible noise floor to the loudest", it does not leave much room for mastering error/sloppiness. While the 24bit dynamic range is much more forgiving... IM-audio-amateur-HO.
I completely agree for everything other than final playback. And really I don't even care for final playback. I just get really annoyed when everyone is saying their brickwalled hi-rez download sounds so much better than the supposedly equivalent CD version.
 

Atanasi

Addicted to Fun and Learning
Forum Donor
Joined
Jan 8, 2019
Messages
713
Likes
792
Again, how does that work?
Genuinely curious...
When signal is rounded to 16 bits, there is some rounding error. The nature of the rounding error depends on how the rounding is done. If it is ordinary, arithmetic rounding, the rounding error has frequency components depending on the original signal, that is, the rounding introduces distortion.
If on the other hand, the rounding method is random, the rounding error is noise-like, which is more audibly pleasing, although on the average, the error is a bit larger than with arithmetic rounding. Randomness may be evenly distributed, which produces white noise, or it may be shaped to chosen frequencies, which makes the noise even less audible.
 

Frank Dernie

Master Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
6,445
Likes
15,780
Location
Oxfordshire

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
Not to kick a dead horse, but while I totally agree with 44.1 (48) kHz for sampling, I think 24 bit still makes practical sense. While 16bit is perfectly adequate dynamic range for "from below-audible noise floor to the loudest", it does not leave much room for mastering error/sloppiness. While the 24bit dynamic range is much more forgiving... IM-audio-amateur-HO.

Absolutely. But a recording that is mixed/mastered at 24+ bits and then mixed-down and dithered to 16-bit will pose no such risks, i.e. 16-bit is perfectly adequate as a playback medium.

Interesting. So we should be more picky with our DAC choices if we have a lot of 16/44 content?

I get this is speaking technically. I'm not sure we'd notice the difference though. Would we?

I certainly wouldn't :) And I'm not too picky about DACs (SINAD of >90dB and a proper reconstruction filter does the job IMHO).
 

MakeMineVinyl

Major Contributor
Joined
Jun 5, 2020
Messages
3,558
Likes
5,871
Location
Santa Fe, NM
Well for me, my reasons are nostalgia, some 70s and early 80s records are mastered really good, and the CD loudness war.

Loudness war is a damn shame and people think it's CDs fault. :mad:
Nah, the loudness wars were raging far back in the 70s on FM radio. Such devices as the "Optimod" manipulated the signal in elaborate fashion to squeeze every last bit of loudness out of the music. Actually you can go back even further, to the 50s where 'audiophile demo' records of the time by Enoch Light were compressed to within an inch of their lives. This was his 'sound' - super compressed and in your face.

The L1 Sonic Maximizer by Waves was one of the first plug-ins for ProTools which performed the currently popular digital version of compression where the audio is squashed as close as possible to 0dB FS. In the early 90s when I was doing sound design for motion pictures, I used the L1 to heavily compress gun shots. The compression made the guns sound extremely aggressive and 'mean'.
 

waynel

Major Contributor
Forum Donor
Joined
Nov 14, 2019
Messages
1,035
Likes
1,288
I agree that 16 bit 44.1KHz is generally enough for music distribution but not for recording , editing, mastering or playback. For music production it's helpful to have extra headroom as the peak may be unknown as well as the additive noise with multitracking discussed above. For playback using digital volume control, higher bit depth helps keep the 16 bit dynamic range intact over a range of playback volumes. Sampling rates above 44.1KHz/48KHz are beyond useless for audio.
 

tomtoo

Major Contributor
Joined
Nov 20, 2019
Messages
3,607
Likes
4,514
Location
Germany
16/44 is enough. But....just to stop this discussions go 24/96. And if then there are still people that imagine to hear its digital. Then there is absolutly no reason not to call the people with that speacial white jackets.
 

MontanaAndy

New Member
Joined
Jan 15, 2021
Messages
1
Likes
0
16/44 is enough. But....just to stop this discussions go 24/96. And if then there are still people that imagine to hear its digital. Then there is absolutly no reason not to call the people with that speacial white jackets.
New guy in need of some understanding.... I like my old vinyl, and have a hobby of digitizing much of it. I record it at 24/48Khz. Here's why. If I put new strings on my guitar and strum a chord, the sound rings until the strings quit vibrating. As I understand, when we record a signal electronically, the sound level reaches a low point where it no longer "triggers" the recording capture, and the (what I think of as) headroom of a well mastered vinyl recording is just lost. I agree that 16/44.1 is fine for most presentation, but for critical listening my ears tell me there is more to hear than a CD quality gives me. That is a reason why studio recording is done more often at the higher bit/sampling rates... Are my old ears imagining this or is this line of thought accurate? Thanks for your input...
 

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
An example would be when you're reducing bit depth from 24 bit to 16 bit with no dithering.

Here's the process for each sample FYI:
1) Express 24-bit value in 16-bit integer range (two's complement), which is 1 sign bit, 15 integer bits, and 8 fractional bits.
2) Round the 8 fractional bits to the nearest integer. This is done by simply adding 128 (which is 0.5 with the sample expressed as indicated in step 1) to each 24 bit value.
3) Take the upper 16-bits of the 24-bit value to use as your 16-bit sample.

So for periodic signals with frequency components at divisors of the sampling frequency (Fs/4, Fs/5, ...) this same rounding will occur each period of the signal. This adds non-random errors which will be non-white noise and has the property of having most it's energy at signal harmonics and intermodulation frequencies.

This generally isn't much of an issue with music which is not simple periodic signals, but it theoretically does add distortion along with the noise.

Best of all, dithering eliminates this problem. You just add a small amount of random noise energy to each 24-bit sample before doing the rounding procedure above. An RMS of half a bit (so 128 on average) completely eliminates the issue. Even 0.25 (64) works really well.

and @Atanasi... thanks for the replies.
I should have worded my question more precisely - I do have some understanding of how dither works, but the rudiments may be helpful to others :)

I was more focused on @MusicNBeer's points about the quantization error spectra, and the relationship of input signal and sampling rate to that spectra. This is something I had not considered previously...

At any rate, are we not talking about effects that are 93dB down in a 16-bit mix?
Not saying dither is always unimportant, only that context is as well.

FWIW, an interesting paper on quantization error analyses may be found here (I'm still digesting it myself).
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,399
New guy in need of some understanding.... I like my old vinyl, and have a hobby of digitizing much of it. I record it at 24/48Khz. Here's why. If I put new strings on my guitar and strum a chord, the sound rings until the strings quit vibrating. As I understand, when we record a signal electronically, the sound level reaches a low point where it no longer "triggers" the recording capture, and the (what I think of as) headroom of a well mastered vinyl recording is just lost. I agree that 16/44.1 is fine for most presentation, but for critical listening my ears tell me there is more to hear than a CD quality gives me. That is a reason why studio recording is done more often at the higher bit/sampling rates... Are my old ears imagining this or is this line of thought accurate? Thanks for your input...

The dynamic range of the best vinyl replayed over the best equipment is just over half that of dithered 16/44.1, I'm afraid :p
 

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
As I understand, when we record a signal electronically, the sound level reaches a low point where it no longer "triggers" the recording capture, and the (what I think of as) headroom of a well mastered vinyl recording is just lost.

No. The sound energy can decay into the noise-floor of the recording apparatus. The digitization of a mic's signal is not triggered on the presence or absence of input signal, but periodically in time (i.e. sampling rate).

but for critical listening my ears tell me there is more to hear than a CD quality gives me. That is a reason why studio recording is done more often at the higher bit/sampling rates...

Sampling rates > 44.1kHz and bit-depth > 16 bit are utilized in audio production to provide to provide headroom for downstream production processes.

is this line of thought accurate?
It is not.
 

Thomas savage

Grand Contributor
The Watchman
Forum Donor
Joined
Feb 24, 2016
Messages
10,260
Likes
16,298
Location
uk, taunton
For listening, more than you need, for recording and the likes more is useful.

However all thats irrelevant as the limits are being lifted for streaming so its utterly academic and only will become more so as the days months and years tick on bye..

The very question then is one for history.
 

Pluto

Addicted to Fun and Learning
Forum Donor
Joined
Sep 2, 2018
Messages
990
Likes
1,631
Location
Harrow, UK
a 24-track blank loop 16-bit master, every track added is a +3dB in noise floor
No! The first two tracks add 3dB but to add another 3dB you need to add two more tracks and to add a further 3dB you need to add 4 more to the former 4. In other words, doubling the number of tracks adds 3dB of noise (assuming each track contributes equally to the overall noise which is typically not the case with real recordings).

In summary, the more tracks you add the less significant each individual one becomes.
 
OP
M

MusicNBeer

Senior Member
Forum Donor
Joined
Jan 20, 2019
Messages
333
Likes
478
No! The first two tracks add 3dB but to add another 3dB you need to add two more tracks and to add a further 3dB you need to add 4 more to the former 4. In other words, doubling the number of tracks adds 3dB of noise (assuming each track contributes equally to the overall noise which is typically not the case with real recordings).

In summary, the more tracks you add the less significant each individual one becomes.

Yep, the RMS noise increases linearly with each added track. When you convert that to logarithmic dB, this is how it adds, so the equation would be: noiseIncreaseDb = 10*log10(numberTracks)
 
Top Bottom