16 bit AD/C

Pentaxor · 03-12-2010, 12:24 PM

I'm just wondering why some people seem to be too much into this.

what are they suppose to gain in a 16-bit over a 14-bit ADC?

the mathematical equation shows or explains that such bit-rate would produce or enter more data and somehow affects or increase the dynamic range. however, the question is, that such mathematics really show a significant and noticeable difference in the images resolution and detail? it does however post another question that such extra bit-rate would only produce unwanted noise rather than data? if an AA filter is absent, would this benefit the images? or did they purposely allowed such inorder to display a quite noisy presentation of the image?

Itai · 03-12-2010, 12:58 PM

In theory it produces finer data samples, simply said 16 bit is more precise than 14 bits by a factor of 4.

In practice nearly every display on the planet uses only 8 bits per channel (there are two 10-bit ones by they can't be driven yet by anything other than prototype hardware) so even if you get a 16 bit or 14 bit or 32 bit or however many bits per channel, the most you can see on a display is 8 bits.

More bits in the ADC converter means the camera has more precision to choose the 8 bits it outputs but that is only theoretical. No improvement has been noted even with 22 bits per channel.

- Itai
Digital Camera Buying Guide, Photography Articles and Reviews | NeoCamera.com

Pentaxor · 03-12-2010, 02:23 PM

Originally posted by Itai

In theory it produces finer data samples, simply said 16 bit is more precise than 14 bits by a factor of 4.

In practice nearly every display on the planet uses only 8 bits per channel (there are two 10-bit ones by they can't be driven yet by anything other than prototype hardware) so even if you get a 16 bit or 14 bit or 32 bit or however many bits per channel, the most you can see on a display is 8 bits.

More bits in the ADC converter means the camera has more precision to choose the 8 bits it outputs but that is only theoretical. No improvement has been noted even with 22 bits per channel.

- Itai
Digital Camera Buying Guide, Photography Articles and Reviews | NeoCamera.com

so it's more like paying extra for the price of a theory or promise of a better result?

anyway, I'm curious to see if there is really a noticeable significant difference between a 14bit and 16bit on any images presented (side by side).

kenyee · 03-12-2010, 03:21 PM

Originally posted by Pentaxor

anyway, I'm curious to see if there is really a noticeable significant difference between a 14bit and 16bit on any images presented (side by side).

You'd never get an apples to apples comparison (same camera with both ADCs).
This is highly dependent on the analog source and whether there actually is 16-bits of info in it (*after* accounting for noise in the signal paths), so once you change camera bodies, you're comparing apples to oranges. If there are 12-bits of info, it won't make a difference. If there are 14-bits of info, the 16-bit ADC might be more accurate (the 14-bit ADC will be spec'd w/ an accuracy rating over the last few bits of data)...

RolloR · 03-12-2010, 08:20 PM

our k-x and k-7 are 12-bit right?

sterretje · 03-12-2010, 10:29 PM

Originally posted by Pentaxor

it does however post another question that such extra bit-rate would only produce unwanted noise rather than data?

Just a correction there. A 16bit ADC does not (necessarily) produce more noise. It makes available noise (from the sensor) more visible.

A noise level that will result in zero with a 14bit (it did not 'trigger' the first bit), will result in a value between 0 and 3 (the values that can be stored in 2 bits) with a 16bit ADC.

jan rinze · 03-13-2010, 05:40 AM

the linear data needs to be gamma corrected. the bottom bits do have a significant role in the dark areas.
Since one stop equals 2x the amount of light the theoretical dynamic range of a linear sampled image equals the amount of bits.
16 bit therefore has a theoretical DR of 16 stops.
of course raising the number of bits of the ADC does not improve the signal output by the CMOS chip and therefore it needs to 'match' the DR of the CMOS chip.

deejjjaaaa · 03-14-2010, 06:33 PM

Originally posted by Pentaxor

I

what are they suppose to gain in a 16-bit over a 14-bit ADC?

there is a very clear explanation by Emil Martinec, who in addition teaches, so he can explain things clearly - read the whole thread, paying attention to what "ejmartin" writes:

http://luminous-landscape.com/forum/index.php?showtopic=42158

Pentaxor · 03-14-2010, 08:19 PM

Originally posted by deejjjaaaa

there is a very clear explanation by Emil Martinec, who in addition teaches, so he can explain things clearly - read the whole thread, paying attention to what "ejmartin" writes:

Dynamic range and DxO - Luminous Landscape Forum

thanks.

03-12-2010, 12:24 PM	#1
Pentaxor Veteran Member Join Date: May 2009 Location: Vancouver, B.C. Photos: Gallery Posts: 6,513	16 bit AD/C I'm just wondering why some people seem to be too much into this. what are they suppose to gain in a 16-bit over a 14-bit ADC? the mathematical equation shows or explains that such bit-rate would produce or enter more data and somehow affects or increase the dynamic range. however, the question is, that such mathematics really show a significant and noticeable difference in the images resolution and detail? it does however post another question that such extra bit-rate would only produce unwanted noise rather than data? if an AA filter is absent, would this benefit the images? or did they purposely allowed such inorder to display a quite noisy presentation of the image?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Windows 7 64 bit & CS4 64 bit ?	holdgaj	Digital Processing, Software, and Printing	2	10-20-2009 06:12 AM
People a bit playfull, just a bit	devisor	Post Your Photos!	4	10-15-2009 12:25 AM
24 bit or 48 bit scanning	photolady95	Digital Processing, Software, and Printing	3	05-29-2009 08:14 AM
12-bit, 14-bit and 48-bit capture	Clarkey	Troubleshooting and Beginner Help	1	02-19-2009 09:16 PM
what is the difference bwt exporting to 8-bit or 16-bit tiffs?	rdrum76	Photographic Technique	3	01-22-2009 01:51 PM

03-12-2010, 12:58 PM	#2
Itai Senior Member Join Date: Sep 2009 Location: Montreal Photos: Gallery Posts: 209	In theory it produces finer data samples, simply said 16 bit is more precise than 14 bits by a factor of 4. In practice nearly every display on the planet uses only 8 bits per channel (there are two 10-bit ones by they can't be driven yet by anything other than prototype hardware) so even if you get a 16 bit or 14 bit or 32 bit or however many bits per channel, the most you can see on a display is 8 bits. More bits in the ADC converter means the camera has more precision to choose the 8 bits it outputs but that is only theoretical. No improvement has been noted even with 22 bits per channel. - Itai Digital Camera Buying Guide, Photography Articles and Reviews \| NeoCamera.com

03-12-2010, 02:23 PM	#3
Pentaxor Veteran Member Join Date: May 2009 Location: Vancouver, B.C. Photos: Gallery Posts: 6,513 Original Poster	Originally posted by Itai In theory it produces finer data samples, simply said 16 bit is more precise than 14 bits by a factor of 4. In practice nearly every display on the planet uses only 8 bits per channel (there are two 10-bit ones by they can't be driven yet by anything other than prototype hardware) so even if you get a 16 bit or 14 bit or 32 bit or however many bits per channel, the most you can see on a display is 8 bits. More bits in the ADC converter means the camera has more precision to choose the 8 bits it outputs but that is only theoretical. No improvement has been noted even with 22 bits per channel. - Itai Digital Camera Buying Guide, Photography Articles and Reviews \| NeoCamera.com so it's more like paying extra for the price of a theory or promise of a better result? anyway, I'm curious to see if there is really a noticeable significant difference between a 14bit and 16bit on any images presented (side by side).

03-12-2010, 03:21 PM	#4
kenyee Veteran Member Join Date: Mar 2007 Location: Boston, PRofMA Photos: Albums Posts: 3,026	Originally posted by Pentaxor anyway, I'm curious to see if there is really a noticeable significant difference between a 14bit and 16bit on any images presented (side by side). You'd never get an apples to apples comparison (same camera with both ADCs). This is highly dependent on the analog source and whether there actually is 16-bits of info in it (after accounting for noise in the signal paths), so once you change camera bodies, you're comparing apples to oranges. If there are 12-bits of info, it won't make a difference. If there are 14-bits of info, the 16-bit ADC might be more accurate (the 14-bit ADC will be spec'd w/ an accuracy rating over the last few bits of data)...

03-12-2010, 08:20 PM	#5
RolloR Senior Member Join Date: Feb 2010 Photos: Albums Posts: 264	our k-x and k-7 are 12-bit right?

03-12-2010, 10:29 PM	#6
sterretje Veteran Member Join Date: Dec 2008 Location: Roodepoort, South Africa Photos: Gallery \| Albums Posts: 3,561	Originally posted by Pentaxor it does however post another question that such extra bit-rate would only produce unwanted noise rather than data? Just a correction there. A 16bit ADC does not (necessarily) produce more noise. It makes available noise (from the sensor) more visible. A noise level that will result in zero with a 14bit (it did not 'trigger' the first bit), will result in a value between 0 and 3 (the values that can be stored in 2 bits) with a 16bit ADC.

03-13-2010, 05:40 AM	#7
jan rinze Forum Member Join Date: Jul 2008 Location: Netherlands Photos: Gallery Posts: 61	the linear data needs to be gamma corrected. the bottom bits do have a significant role in the dark areas. Since one stop equals 2x the amount of light the theoretical dynamic range of a linear sampled image equals the amount of bits. 16 bit therefore has a theoretical DR of 16 stops. of course raising the number of bits of the ADC does not improve the signal output by the CMOS chip and therefore it needs to 'match' the DR of the CMOS chip.

03-14-2010, 06:33 PM	#8
deejjjaaaa Veteran Member Join Date: May 2007 Location: steel city / rust belt Posts: 2,043	Originally posted by Pentaxor I what are they suppose to gain in a 16-bit over a 14-bit ADC? there is a very clear explanation by Emil Martinec, who in addition teaches, so he can explain things clearly - read the whole thread, paying attention to what "ejmartin" writes: http://luminous-landscape.com/forum/index.php?showtopic=42158

03-14-2010, 08:19 PM	#9
Pentaxor Veteran Member Join Date: May 2009 Location: Vancouver, B.C. Photos: Gallery Posts: 6,513 Original Poster	Originally posted by deejjjaaaa there is a very clear explanation by Emil Martinec, who in addition teaches, so he can explain things clearly - read the whole thread, paying attention to what "ejmartin" writes: Dynamic range and DxO - Luminous Landscape Forum thanks.