Daran wrote: aCuria wrote:
The Nyquist theorem (or Nyquist-Shannon sampling theorem) states that to accurately record or digitize a continuous analog signal without losing information, you must sample it at a rate greater than twice the highest frequency component present in that signal.
This means we want the Airy disk to cover at least 2 pixels on the sensor. If you use 1.5 you are losing information.
Uhm, that way you do loose information. You generally loose information as your airy disk gets bigger and also if you sample with less pixels. So your requirement of using no more than two pixels per airy disk is the wrong way around if you do not want to loose information. Meanwhile it is a valid approach to reduce Moire.
If the lens is not perfect you can think of this as the airy disc being larger than the calculated value.
Approximately if the lens is soft. CA means that the airy discs for different colors do not coincide. Coma means that they are not round.
I think you are right, we want to use a fast enough aperture such that the airy disk covers 2 or less pixels.
Daran wrote: aCuria wrote:
The Nyquist theorem (or Nyquist-Shannon sampling theorem) states that to accurately record or digitize a continuous analog signal without losing information, you must sample it at a rate greater than twice the highest frequency component present in that signal.
This means we want the Airy disk to cover at least 2 pixels on the sensor. If you use 1.5 you are losing information.
Uhm, that way you do loose information. You generally loose information as your airy disk gets bigger and also if you sample with less pixels. So your requirement of using no more than two pixels per airy disk is the wrong way around if you do not want to loose information. Meanwhile it is a valid approach to reduce Moire.
If the lens is not perfect you can think of this as the airy disc being larger than the calculated value.
Approximately if the lens is soft. CA means that the airy discs for different colors do not coincide. Coma means that they are not round.
I think you are right, we want to use a fast enough aperture such that the airy disk covers 2 or less pixels.
To maximize DOF without losing information then have the airy disk cover exactly 2 pixels
Daran wrote: aCuria wrote:
The Nyquist theorem (or Nyquist-Shannon sampling theorem) states that to accurately record or digitize a continuous analog signal without losing information, you must sample it at a rate greater than twice the highest frequency component present in that signal.
This means we want the Airy disk to cover at least 2 pixels on the sensor. If you use 1.5 you are losing information.
Uhm, that way you do loose information. You generally loose information as your airy disk gets bigger and also if you sample with less pixels. So your requirement of using no more than two pixels per airy disk is the wrong way around if you do not want to loose information. Meanwhile it is a valid approach to reduce Moire.
If the lens is not perfect you can think of this as the airy disc being larger than the calculated value.
Approximately if the lens is soft. CA means that the airy discs for different colors do not coincide. Coma means that they are not round.
I think you are right, we want to use a fast enough aperture such that the airy disk covers 2 or less pixels.
To maximize DOF without losing information then have the airy disk cover 2 pixels
Aug 19, 2026 at 10:44 PM
Previous versions of aCuria's message #17091967 « A7R6 and diffraction »
