DSLR Bias Frames

[laser_jock99]

Are these ISO setting dependent? And if so why?

Thanks in adavce

[red dwalf]

no idea but as there are so easy to take i take them at the same iso rate

[andrewluck]

Yes, definitely ISO dependent. The readout noise from the sensor will be amplified depending upon the ISO setting.

To simplify things I always shoot at the same ISO setting when using a DSLR. Then I only need one set of calibration frames.

Andrew

[MND]

Do you only need one set of bias not new ones each session with dslr?

[red dwalf]

as long as they are at the same iso rating and your camera and scope setup hasn`t changed then i can`t see the point of doing them every session, they are not temperature dependant, if you move or change anything in the imaging chain then you need to take new ones

[Coco]

Ive always found bias make more vertical noise patterns so I never use them..

[red dwalf]

must admit i`ve processed images without darks and they have come out fine, to my untrained eye anyway, you add darks, flats, bias and the finished image doesn`t always look that better for it

[dph1nm]

Yes, definitely ISO dependent. The readout noise from the sensor will be amplified depending upon the ISO setting.

May also use different electronics. I have read that on the 1000D different circuitry is used for ISO1600 than for lower ISOs. Certianly the bias looks very different at ISO1600 on mine (much cleaner!) than at ISO200/400/800 (which have horrible vertical banding).

NigelM

[andrewluck]

No need to re-do bias frames if the imaging train is changed. They're taken with the lens cap on.

You need to stack  a lot of bias frames as the noise in them is imprinted on every light frame. 50+ is good.

Flat frames are essential. I was rather ambivalent towards darks with a DSLR for two reasons: 1) The temperature isn't regulated; 2) The camera electronics mess with the frames (even if they're raw). Dither between exposures, use hot pixel removal and stack a lot of subframes instead.

Just my opinion :-)

Andrew

[Alien 13]

I have allways been confused with camera noise what does it look like? i have tried bias darks etc in the past but cant see any difference same with ISO settings the lower ones seem a little more grainy too me.

Alan

[wxsatuser]

Are these ISO setting dependent? And if so why?

Thanks in adavce

Same ISO and temperature as the lights.

In theory the bias is in any darks so using darks alone should be ok.

[wxsatuser]

I have allways been confused with camera noise what does it look like? i have tried bias darks etc in the past but cant see any difference same with ISO settings the lower ones seem a little more grainy too me.

Alan

Here is a bias frame from my 60Da, stretched a fair bit in Imagesplus.

[wxsatuser]

Here are 3 dark frames from my 60Da.......180/300/600 seconds respectively, all at 23C ISO 800.

All roughly cropped out of the centre of each frame and stretched a reasonable amount in Imagesplus.

180secs

300sec

600sec

[Alien 13]

Thanks so much for that wxsatuser, it realy makes it obvious now.

Alan

[bruno]

I have read somewhere that there are no need for bais frames, as dark frames contain bais

[sharkmelley]

Yes, bias frames are ISO dependent but I don't shoot bias frames for my Canon.  Instead I use a synthetic bias frame with constant pixel values.  The constant value is set to be the average of pixels in a typical bias frame.  The reason for not using "real" bias frames is that my Canons do not have repeatable bias frames - every time I switch the camera on I get a bias frame with different "patterning" or banding.  Using mismatched bias frames during image calibration then increases the pattern noise instead of reducing it.

On acquisition I use dithering to "average" out any pattern noise and then in the final stacked image I use a "remove Canon Banding" macro to kill any residual banding (which might be either vertical or horizontal).

Mark

[sharkmelley]

I have read somewhere that there are no need for bais frames, as dark frames contain bais

Yes dark frames contain bias but when using a Canon DSLR you still need separate bias frames to correct your flats.  When I shoot a set of darks (which is rare) I simultaneously (i.e. without switching the camera off) shoot a set of bias frames so I can subtract them from the darks and add the synthetic flats back to the darks.  This gets round the problem of having a set of "library" darks with different pattern noise to my lights.

If banding is not a problem in your stacked images then ignore all the complications I've just described!

Mark

[Christopher Davenport]

I only use DARKS if I surpass my sensor (600D) noise thresholds for temperature and exposure time.

Analysis on Temp, ISO and Exp in seconds before dark noise is added to read noise.
(Excludes read noise which is always present.)

SN°C  I400,  I800,  I1600
0°       3494   1747   873.5
5°       1558   779     389.5
10°      714    357     178.5
15°      338    169     84.5  
20°      164     82      41     
 

See this great article http://www.christoph...de/Astro/DC.pdf

Edit - If I am within the above range I only remove the BIAS. (read noise)

On acquisition I use dithering to "average" out any pattern noise and then in the final stacked image I use a "remove Canon Banding" macro to kill any residual banding (which might be either vertical or horizontal).

Interesting indeed, any chance you can point me in the direction of how you do the "remove Cannon Banding" macro?

[sharkmelley]

Interesting indeed, any chance you can point me in the direction of how you do the "remove Cannon Banding" macro?

I use the PixInsight Canon Banding Reduction Tool.   I believe there is also a "Noel's Action" for Photoshop that removes banding.  There are probably others.

Hint:  A horizantal band removal tool will also remove vertical bands if you flip the image first.

Mark

[NorrinRadd]

Yes, bias frames are ISO dependent but I don't shoot bias frames for my Canon.  Instead I use a synthetic bias frame with constant pixel values.  The constant value is set to be the average of pixels in a typical bias frame.  The reason for not using "real" bias frames is that my Canons do not have repeatable bias frames - every time I switch the camera on I get a bias frame with different "patterning" or banding.  Using mismatched bias frames during image calibration then increases the pattern noise instead of reducing it.

On acquisition I use dithering to "average" out any pattern noise and then in the final stacked image I use a "remove Canon Banding" macro to kill any residual banding (which might be either vertical or horizontal).

Mark

Mark,

Can I ask what you use as a synthetic bias frame ie  did you make it and how etc.   

What you said makes a lot of sense to me and from what I've experienced in my limited efforts as I have been guilty of using too high ISO's so any noise offset is really interesting.

Simon

[sharkmelley]

Can I ask what you use as a synthetic bias frame ie  did you make it and how etc.   

The average bias level for a Canon will be a power of 2 e.g. 512, 1024, 2048 depending on which model.  The way to generate a synthetic bias depends on what processing package you use.  The easiest way is probably to shoot a bias frame then apply a strong blur function to give all pixels more or less the same value.  Or if you can do pixel maths then multiply all pixels by 0 then add the bias relevant to your camera.

Mark

[NorrinRadd]

Thanks Mark much appreciated .

Alas I'm a naive with the fine workings if offsetting noise ie how to improve it and it's workings which is why I found your information realt interesting.

I use a Cannon 650d fully modded , and stack in DSS and clean up in photoshop. I do have Maxim DL but gave not realy got into this yet.

So to create a synthetic bias would you recommend just taking a single frame and then blurring it lots in Photoshop or taking a few blurring them all and then create a master blurred bias frame.

Any further pointers on your great idea much appreciated :-)