snubbr.com

You talk about noise, but you dont understand it!
Most people talk about noise from sensors, but lack fundamental knowledge about noise ..

Therefor I suggest that you all read the following article:.

Http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/index.html.

A700 /img/avatar9.jpg..

Comments (22)

A really interesting article. It will require several readings to comprehend it all.Thank you for providing the link..

Art.

'There is no limit to what a man can do so long as he does not care who gets the credit.'-Philip Hyde (1922-2006)http://yorkiedad.smugmug.com..

Comment #1

He forgot to say how to combine all these noises to make it look like random film grain..

Http://jaylifotoweddings.blogspot.com..

Comment #2

I do NOT need to read yet another article on noise, THANK YOU VERY MUCH!.

Heres the formula I use, if I can SEE noise in an image, that = BAD!!.

If I do NOT see noise in an image that = GOOD.

Pretty simple HUH!.

Try it some time, I think you'll be surprised just how EFFECTIVE this little formula works..

On the other hand, if you'd prefer to run to your computer and GOOGLE noise and read 1000 different articles EVERY TIME you SEE noise in an image, JUST SO YOU CAN UNDERSTAND WHAT YOU ARE LOOKING AT, please feel to do so, the rest of use have better things to do with our time! .

Thanks for posting, I feel sooooo much better now.......

Comment #3

Nice, basic article on noise as it relates to digital / electronic sensors..

They did mention the issue of pattern noise, and related one example (from the Canon 20D). It would be nice to see an "amplified" (no pun intended) discussion of pattern noise and how it relates to specific sensor +amplifier technologies...

Comment #4

Snapshot32 wrote:.

Heres the formula I use, if I can SEE noise in an image, that = BAD!!.

If I do NOT see noise in an image that = GOOD.

This applies to you but some people actually prefer some noise in their pictures. Depends on the style you want but a noisefree picture is not necessary a good picture..

Http://ricoh-gr-diary.blogspot.com/http://www.flickr.com/photos/cristiansorega..

Comment #5

One weird thing I noticed it that he first says that the Panasonic FZ50 sensor have average light collection performance, then he says that because it has less stops to saturation from middle gray compared to DSLR's (due to the lower saturation levels of the smaller pixlels), it's efficiency figures shall be considered higher by a factor 1.4 to 2.0. This reasoning seems odd to me. Anyone care to explain?.

Regards /Mats..

Comment #6

Mats_b wrote:.

One weird thing I noticed it that he first says that the PanasonicFZ50 sensor have average light collection performance,.

...yields a gain g of roughly .29 photons/12-bit ADU at ISO 400....

...which yields .072 photons per ADU per square micron, comfortably in the middle of the efficiency table above....

Then he says that because it has less stops to saturation from middle graycompared to DSLR's (due to the lower saturation levels of the smallerpixlels),.

...digicams put middle gray about 0.5-1.0 stop closer to raw saturation, due to their lower dynamic range. This means that their ISO calibration is about 0.5-1.0 stops understated in relation to DSLR ISO calibration...

Its efficiency figures shall be considered higher by a factor 1.4 to 2.0.This reasoning seems odd to me. Anyone care to explain?.

If you read the first quote, it says "at ISO 400"..

The second quote says that "digicam ISO400 is not equal DSLR ISO 400, but more like DSLR ISO 560-800". If you put digicam and DSLR on equal ground ISO-calibration wise, you get that 1.4-2.0 factor. Not odd to me..

Hope this helps,Stefan.

Regards /Mats..

Comment #7

The odd part is that it is not about actual efficiency but about conversion of values..

With a DSLR we could simulate having a lower saturation level by adjusting exposure and clip values exceeding the saturation value of the FZ50 and then do a conversion similar to the FZ50 on the data. The DSLR would seem to have become more sensitive, but the actual efficiency would not have changed at all..

Regards /Mats..

Comment #8

So it doesn't matter to you if noise reduction reduces detail and even creates a "plastic" looking image as long as there's no visible noise? Simple rules are often not very useful past a certain simplistic generalization.BJ NichollsSLC, UT..

Comment #9

Moire wrote:.

Most people talk about noise from sensors, but lack fundamentalknowledge about noise ..

Therefor I suggest that you all read the following article:.

Http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/index.html.

Most don't care about it's cause or origin, just it's severity..

Image control:Zoom outZoom 100%Zoom inExpand AllOpen in new window.

'I cried because I had no E-3. Then I met a man with no E-510'.

Olympus E-420, E-330, Pentax K20D. 40 lenses of various types..

Comment #10

That's a great article. It's very well written for a semi-technical piece..

Thanks!..

Comment #11

Mats_b wrote:.

The odd part is that it is not about actual efficiency but aboutconversion of values..

No, it's about the actual efficiency (converting the incoming photons into voltage/digital signal). The numerical value of the efficiency (in table 1) depends on the ISO, as it is essentially the voltage after the gain (which is determined by ISO). And the FZ50's stated ISO and DSLR's ISO are not the same (for the reasons of lower dynamic range of FZ50). If we use the real common ground ISO, the FZ50 turns out (somewhat surprisingly, but not unbelievably) more efficient. Big deal, there are possibly tricks you can to with small sensors that has not yet been transferred to bigger ones (and there might be undesirable tradeoffs)..

With a DSLR we could simulate having a lower saturation level byadjusting exposure and clip values exceeding the saturation value ofthe FZ50 and then do a conversion similar to the FZ50 on the data..

Its not how it works. It is more like this: FZ50 knows it can't keep the whole dynamic range of a picture taken by a DSLR, so it chooses to underexpose (by 0.5-1 stop, by having stated ISO 1.4-2x smaller then the true ISO is uses) to save the highlights and sacrifice the shadows..

The DSLR would seem to have become more sensitive, but the actualefficiency would not have changed at all..

That's not what the article is saying..

You are still missing the fact that the sensitivity value is _FOR FIXED ISO_, once you change ISO, you need to properly normalize the sensitivity value. The 0.72 photons per ADU per micron^2 for FZ50 at FXZ50'd stated ISO400 has to be compared to the photons per ADU per micron^2 for a DSLR at ISO 560-800,. Table 1 shows the sensitivity of DSLRs at ISO 400, so you need to divide those numbers by 1.4-2 to be on equal ground comparing with FZ50's 0.72. If you take D3's 0.113, it is cca 1.57x more then FZ50's 0.72, i.e. well within 1.4-2 range. No magic needed for FZ50..

Stefan.

Regards /Mats..

Comment #12

Snapshot32 wrote:.

I do NOT need to read yet another article on noise, THANK YOU VERY MUCH!.

Heres the formula I use, if I can SEE noise in an image, that = BAD!!.

If I do NOT see noise in an image that = GOOD.

Pretty simple HUH!.

Here's the formula I use, if I can see someone SHOUTING in a post, that = bad!!.

If I do not see someone SHOUTING in a post that = good..

Pretty simple huh!..

Comment #13

VladimirV wrote:.

Snapshot32 wrote:.

Heres the formula I use, if I can SEE noise in an image, that = BAD!!.

If I do NOT see noise in an image that = GOOD.

This applies to you but some people actually prefer some noise intheir pictures. Depends on the style you want but a noisefree pictureis not necessary a good picture..

Does it make the image worse? It's easier to add grain without losing quality then it is to subtract noise. Even with Noise Ninja to help, this remains true..

Interesting article. And a thank you for the OP for posting it - But overall, I agree with Snapshot. His high tech formula works pretty well... .

Dave..

Comment #14

Jaylifoto wrote:.

He forgot to say how to combine all these noises to make it look likerandom film grain..

Maybe not, but that wasn't the focus of the article..

If you want to emulate film grain, or things like it, only your imagination is the limit. There are any number of ways to distort an image at the sub-pixel level to lose it's cartesian grid..

John.

Image control:Zoom outZoom 100%Zoom inExpand AllOpen in new window..

Comment #15

John W Peterson wrote:.

Nice, basic article on noise as it relates to digital / electronicsensors.They did mention the issue of pattern noise, and related one example(from the Canon 20D). It would be nice to see an "amplified" (no punintended) discussion of pattern noise and how it relates to specificsensor +amplifier technologies..

One of the most important things to realize about pattern noise is that it's presence is not necessarily obvious. When the Canon 20D first came out, for example, people that upgraded from the 10D started to see a lot of horizontal line noise at high ISOs, and thought that this was a new noise. However, the 10D had stronger line noises, but they were hidden by higher pixel noise. Same holds true for many Nikons of the same era; they had more line noise than the 20D, but had so much more random pixel noise that it was not noticeable. And that's the big problem with pattern noises; they are more visible, even when an image is printed very small, or when an image is downsampled to a small web-image; pattern noises reduce very little, compared to random pixel noise..

John.

Image control:Zoom outZoom 100%Zoom inExpand AllOpen in new window..

Comment #16

Mats_b wrote:.

One weird thing I noticed it that he first says that the PanasonicFZ50 sensor have average light collection performance, then he saysthat because it has less stops to saturation from middle graycompared to DSLR's (due to the lower saturation levels of the smallerpixlels), it's efficiency figures shall be considered higher by afactor 1.4 to 2.0. This reasoning seems odd to me. Anyone care toexplain?.

The point is that ISO settings on cameras, even if true to metering, can be totally arbitrary in terms of RAW sensitivity. The FZ50 has a base ISO of 100, with about 2.5 stops of headroom from middle gray up to RAW clipping. That same sensor capture and gain and digitization could be used for ISO 50 with 1.5 stops of headroom, ISO 200 with 2.5 stops of headroom, ISO 400 with 4.5 stops of headroom, etc. The way that most DSLRs would use it would be as ISO 200 with 3.5 stops of headroom, and in that sense, by DSLR standards, the base ISO 100 of the FZ50 would be ISO 200..

This does nothing to improve the DR of the camera at base ISO; the absolute sensitivity is irrelevant, but for absolute sensitivity itself, the noise figures as they would blindly be measured, assuming that noise measurements from DSLRs and small-sensor cameras like the FZ50 have equal meaning, are wrong. The shot noise is 1/2 stop lower for the FZ50, and the read noise a full stop lower than a figure like "full well saturation of x electrons" or "x ADU read noise" might suggest..

This is easily overlooked because the FZ50 sensor is so much smaller than a DSLR sensor that it has more shot noise, regardless of the higher efficiency..

In fact, most of the truth relating to the myths about pixel density are hidden in this same circumstance..

John.

Image control:Zoom outZoom 100%Zoom inExpand AllOpen in new window..

Comment #17

Mats_b wrote:.

The odd part is that it is not about actual efficiency but aboutconversion of values..

With a DSLR we could simulate having a lower saturation level byadjusting exposure and clip values exceeding the saturation value ofthe FZ50 and then do a conversion similar to the FZ50 on the data.The DSLR would seem to have become more sensitive, but the actualefficiency would not have changed at all..

The difference is that at base ISO 100 RAW saturation (~2.5 stops higher than metered ISO 100 middle gray in the green channel), the FZ50 is collecting as many photons per square mm on the sensor as the DSLR. This higher native photon sensitivity is not a cheap trick performed by running out of room to collect electrons, such as in having a very small fill factor (percentage of sensitive sensor area); it is achieved by actually capturing more of the photons that hit the surface of the sensor, most likely due to more effective use of microlenses..

John.

Image control:Zoom outZoom 100%Zoom inExpand AllOpen in new window..

Comment #18

That s a nice considered response to someone sharing some info.......More kit than I probably need.http://www.flickr.com/photos/rb56/..

Comment #19

DonEsteban wrote:.

Its not how it works. It is more like this: FZ50 knows it can't keepthe whole dynamic range of a picture taken by a DSLR, so it choosesto underexpose (by 0.5-1 stop, by having stated ISO 1.4-2x smallerthen the true ISO is uses) to save the highlights and sacrifice theshadows..

That is backwards..

The FZ50 *overexposes* by a stop, so what would be a base ISO of 200 on a DSLR is pulled to 100.

The DSLR would seem to have become more sensitive, but the actualefficiency would not have changed at all..

No; with the same absolute exposure (lighting, and Av/Tv), the FZ50 collects more photons per unit of sensor area than all DSLRs (not just this one camera, BTW; it just serves as an example)..

John..

Comment #20

Are you saying that with equal light and exposure time,the FZ50 is converting 1.4 to 2 times as many photonsto electrons for an equal size area compared to a DSLR?That would truly be higher efficiency and no "trick"..

If that is the case I have misunderstood the article andagree that the efficiency of the FZ50 is higher..

Regards /Mats..

Comment #21

Yet, it's taken for what it's worth. Thanks..

...Bob, NYC.

Galleries: http://www.pbase.com/btullis.

Image control:Zoom outZoom 100%Zoom inExpand AllOpen in new window..

Comment #22

Click Here to View All...

Sponsored Amazon Deals:

1. Get big savings on Amazon warehouse deals.
2. Save up to 70% on Amazon Products.


This question was taken from a support group/message board and re-posted here so others can learn from it.

 

Categories: Home | Diet & Weight Management | Vitamins & Supplements | Herbs & Cleansing |

Sexual Health | Medifast Support | Nutrisystem Support | Medifast Questions |

Web Hosting | Web Hosts | Website Hosting | Hosting |

Web Hosting | GoDaddy | Digital Cameras | Best WebHosts |

Web Hosting FAQ | Web Hosts FAQ | Hosting FAQ | Hosting Group |

Hosting Questions | Camera Tips | Best Cameras To Buy | Best Cameras This Year |

Camera Q-A | Digital Cameras Q-A | Camera Forum | Nov 2010 - Cameras |

Oct 2010 - Cameras | Oct 2010 - DSLRs | Oct 2010 - Camera Tips | Sep 2010 - Cameras |

Sep 2010 - DSLRS | Sep 2010 - Camera Tips | Aug 2010 - Cameras | Aug 2010 - DSLR Tips |

Aug 2010 - Camera Tips | July 2010 - Cameras | July 2010 - Nikon Cameras | July 2010 - Canon Cameras |

July 2010 - Pentax Cameras | Medifast Recipes | Medifast Recipes Tips | Medifast Recipes Strategies |

Medifast Recipes Experiences | Medifast Recipes Group | Medifast Recipes Forum | Medifast Support Strategies |

Medifast Support Experiences |

 

(C) Copyright 2010 All rights reserved.