• 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!

Can you tell 8-bit file from 16-bit file in a DBT? Listening test, RLJ

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,728
Location
Prague
A challenge - can you tell 8-bit file from 16-bit file?

Based on a discussion in my thread God of Sinad
I have prepared a listening test with an original 44.1kHz/16bit file and the derivative from that file with 8bit resolution, with appropriate dither

The beloved Rickie Lee Jones, both files coded as 16-bit, but one of them is only in 8-bit resolution. Appropriate dither used. Here is the link, and I take only results from a DBT test, foobar ABX report that can be verified.

https://pmacura.cz/rlj8_16.zip

So, can you hear a difference? If you say yes, then please add your foobar ABX report.

FYI, this is what Deltawave calculates as a difference in time domain.

1660252586807.png
 
Last edited:

ThatM1key

Major Contributor
Joined
Mar 27, 2020
Messages
1,048
Likes
882
Location
USA
I can tell a very very small bit of difference. The guitars notes are slightly more mashed together and have less punch. Although, the voices sounded pretty equal to me.

To be honest, I didn't do the ABX. I flipped between the 2 files over and over, and the difference is there but its very minimal. You would have to be told there's a difference just to notice it.
 

AdamG

Proving your point makes it “Science”.
Moderator
Forum Donor
Joined
Jan 3, 2021
Messages
4,637
Likes
14,924
Location
Reality
Oh heck my wife could tell the difference from the kitchen and I had headphones on! Easy work of that!o_O
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,985
Location
Riverview FL
One bit is enough to at least provide a recognizable playback.

https://soundcloud.com/ray-880875693%2Fonebitpaul
Amazingly, I first posted a longer sample and it was rejected for copyright infringement.


1660265510264.png


Method:

Amplify the digital file until all that is left is zero crossings to max positve and negative.

Attenutate that result to a reasonable level.

The file then has three digital levels - positive, negative, and a rare zero, or, as I see it "one bit". Ok, so maybe it is two bits 01, 00, and 11 (the negative 1 requires the sign bit).
 
Last edited:

DonR

Major Contributor
Joined
Jan 25, 2022
Messages
2,971
Likes
5,615
Location
Vancouver(ish)
One bit is enough to at least provide a recognizable playback.

https://soundcloud.com/ray-880875693%2Fonebitpaul
Amazingly, I first posted a longer sample and it was rejected for copyright infringement.


View attachment 223696

Method:

Amplify the digital file until all that is left is zero crossings to max positve and negative.

Attenutate that result to a reasonable level.

The file then has three digital levels - positive, negative, and a rare zero, or, as I see it "one bit". Ok, so maybe it is two bits 01, 00, and 11 (tje negative 1 requres the sign bit).
Wild.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
One bit is enough to at least provide a recognizable playback.

https://soundcloud.com/ray-880875693%2Fonebitpaul
Amazingly, I first posted a longer sample and it was rejected for copyright infringement.


View attachment 223696

Method:

Amplify the digital file until all that is left is zero crossings to max positve and negative.

Attenutate that result to a reasonable level.

The file then has three digital levels - positive, negative, and a rare zero, or, as I see it "one bit". Ok, so maybe it is two bits 01, 00, and 11 (tje negative 1 requres the sign bit).
I once was traveling in my 1967 Chevelle with rear seat speaker option. Listening to a live broadcast of AC/DC on a Chicago AM station from the Chicago concert on a cold dark night. I had it cranked up and finally my poor old speaker began to sound all fuzzy near the Tennessee state line. After that, when the Beatles came on, this example here is what it sounded like. I had to replace that speaker, and I think I used a Jensen in its place. Just mono so only one Jensen. Little did I know that I had experienced an early analog simulation of one bit clapping.
 
Last edited:
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,728
Location
Prague
8-bit resolution version was prepared this way:

1) the original 44.1/16 file was upsampled using noise shaping dither, to 96kHz/8bit
2) the 96/8 file was then downsampled to 44.1/16 file

Upsampling with noise shaping is crucial here to keep acceptable low audibility of background noise.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,497
Location
Seattle Area
I have prepared a listening test with an original 44.1kHz/16bit file and the derivative from that file with 8bit resolution, with appropriate dither
What do you mean by "appropriate dither?' Noise shaped?

1660280326557.png


To be able to tell the difference then requires people with acuity above 14 kHz.
 
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,728
Location
Prague
What do you mean by "appropriate dither?' Noise shaped?
Yes. As explained in post #9. It is the only way to get low audible noise level of the 8-bit file. With the TPD frequency independent dither, you immediately hear high noise from the 8-bit file. As there is not much usable sound information below -45dBFS in the original file (see histograms), even the 8-bit resolution with proper dithering method used results in acceptable sound quality. And 16 bits is more than enough. And this is still good quality of recording, talking about popular music, with DR=13 of the sample used.

Please let me add some screenshots from Paul's (@pkane ) Deltawave software and also histograms of the 8-bit and 16-bit file versions.

RLJ_raw.png


RLJ_spectrumofdelta.png


RLJ_pkmetrics.png


RLJ_8+16bit_histogram.png


And finally a histogram of the Deltawave difference file

RLJ_delta_histogram.png


Differences are spread mostly around some -43dBFS, as one might expect from the difference in resolution. Please note that we talk about something near -40dB, not about -80dB, not about -110dB or -120dB. This is my point. We have limits when going behind them is a useless effort.
 
Last edited:

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,422
Likes
2,407
Location
Sweden
It is not overly disturbing but yes. I just went 55 but the HF hearing is still ok for my age.

Skärmavbild 2022-08-12 kl. 09.53.23.png
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,497
Location
Seattle Area
Yes. As explained in post #9. It is the only way to get low audible noise level of the 8-bit file.
A point that should have been made at the outset. When dealing with music, you have no idea what dither is used, or if any dither is used in conversion of 24 bit material to 16 bits (or whatever the target is). It is for this reason that I like to sidestep the whole thing and get 24 bit masters.
 
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,728
Location
Prague
The difference extremely noticeable for me and i could identify which is which within seconds.

I agree with you, the difference in noise is clearly audible when listening to first 5 seconds only with high volume setting, that would be too much for the rest of the samples. Average rms power there is only -30 dBFS, so the noise compromise is audible.

Code:
foo_abx 2.0.2 report
foobar2000 v1.4.8
2022-08-12 09:54:15

File A: Rickie1.wav
SHA1: 1af7dad9413342c4c0eddab6182474bded3b02f0
File B: Rickie2.wav
SHA1: f23aede1fe13e20674ffde0e7e2f4d31880679c0

Output:
ASIO : Topping USB Audio Device
Crossfading: NO

09:54:15 : Test started.
09:54:43 : 01/01
09:54:49 : 02/02
09:54:55 : 03/03
09:55:08 : 04/04
09:55:16 : 05/05
09:55:22 : 06/06
09:55:34 : 07/07
09:55:47 : 08/08
09:55:53 : 09/09
09:56:08 : 10/10
09:56:17 : 11/11
09:56:26 : 12/12
09:56:38 : 13/13
09:56:47 : 14/14
09:56:56 : 15/15
09:57:05 : 16/16
09:57:05 : Test finished.

 ----------
Total: 16/16
Probability that you were guessing: 0.0%

 -- signature --
8b9e414650fbd6f914af71bb8b75b9b7ce16775f

It would be much more difficult if you tried more loud parts behind the first 6 seconds.

However, it is amazing what we get from 8-bit noise shaped file.
 
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,728
Location
Prague
When dealing with music, you have no idea what dither is used, or if any dither is used in conversion of 24 bit material to 16 bits (or whatever the target is). It is for this reason that I like to sidestep the whole thing and get 24 bit masters.

That's why it is good when we know what we are doing. BTW the dither used you find from the file spectrum analysis.
 

Mnyb

Major Contributor
Forum Donor
Joined
Aug 14, 2019
Messages
2,645
Likes
3,624
Location
Sweden, Västerås
A little of topic. if I'm excused:

This demonstrates how well dither actually works , if an 8bit file can be this good with dither , how good are then 16 and 24 bit .

There is some absurd opinions that dither does not really work among some audiophiles. this is a good demo of it.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,497
Location
Seattle Area
That's why it is good when we know what we are doing.
The concept of dither is so obscure that many in music production don't understand it either.
 
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,728
Location
Prague
I was curious what you guys have been hearing in this test if we used the usual SINAD chart method. So I took a 1kHz 44.1kHz/16bit TPD dithered sine and made the exactly same procedure as in the test to get the 8 bit resolution, same upsampling/dither, same downsampling. Below is the result. Wow, the SINAD is some 40 dB!!

signal_44_16b_dithered.png


and the 8-bit resolution
signal_44_8b_dithered.png


So we have a real life example of the theoretical 93dB SINAD vs. 40dB SINAD :D.

The result is better from psychoacoustic point of view if we decrease amplitude of the original 44.1/16 sine, below is the sine at -2.7dBFS. The noise in most sensitive part is much lower. It is because the processing was clipping here above. But, the single number SINAD that is used here as a comparative measure is not improved. Again we can see how low is the correlation of SINAD and hearing.

signal_44_8b_dithered_-2dB.png
 
Last edited:

charleski

Major Contributor
Joined
Dec 15, 2019
Messages
1,098
Likes
2,239
Location
Manchester UK
I suspect that a lot of the difference is masked by the fact that the basic recording here is, frankly, a bit ******. You can hear the mic preamp briefly overload at 36s, for instance. It also looks like the vocals were rolled off at 5kHz.
Screenshot 2022-08-12 102551.jpg
 
Top Bottom