While going home with the Schneider 40 mm - a night shot

nicolas claris · Feb 29, 2008

On the go, going back home yesterday, I made a stop where Asher and our beloved had a magnificent meal some centuries ago…

Hy6 +Schneider Super-Angulon PQ 3.5/40 mm - lying on the fence (no tripod)

400 ISO, shutter 2.8 s, ƒ8 - no added light (obvious)

Quickly processed in tif 8bit with CaptureShop (curve/denoise/sharpness, no PP in CS

[EDIT] For the very same reasons than for "the men at work"
I have reworked the derawtization… please see below
I have not deleted the above image for the good understanding of the discussion. [/EDIT]

Georg R. Baumann · Feb 29, 2008

Nicolas Claris said:
Full sRGB JPEG res

Beautiful, I like it a lot!

What filesize is the full sRGB in your link?

nicolas claris · Feb 29, 2008

Georg Baumann said:
Beautiful, I like it a lot!

What filesize is the full sRGB in your link?

'bout 11 Mb (saved at 10 on the CS scale 0 to 12)

Full means full size…

nicolas claris · Feb 29, 2008

Looking back to the other side, the opposite of the river. The houses on the dock are at a distance of 500 meters/1600 feet (I know the length of the bridge nearby):

Schneider 40 mm - 400 ISO - 6.4s

100% crop:

Georg R. Baumann · Feb 29, 2008

OMG, this is a 100% crop? You could print form that! Awww.... NICE!

Tell me, I do not understand one thing, why would you choose ISO 400 on a longexposure shot?

I would think, a longer exposure and ISO 100 would create a better file, then again, I might be completly mistaken. Also, does the Hy6 have something like "darkframesubtraction" to avoid hot pixels with long exposures and noise to be an issue?

Where do you leave the camera at night? LOL Hehehe! <Slipping into camouflage Ninja Gear, climbing up the wall to the back of Nicolas house, patiently sitting on roof and waiting until lights go out, preparing darts with a strong liquid deep sleep agent....deep breath.... thhhhhwup thhhhhwup.... miauuu....Feck, I think I hit the cat.... falling of the roof laughing. LOL>

nicolas claris · Feb 29, 2008

Georg Baumann said:
Tell me, I do not understand one thing, why would you choose ISO 400 on a longexposure shot?

Testing my friend! 400 ISO is noise free...

Where do you leave the camera at night? LOL Hehehe! <Slipping into camouflage Ninja Gear, climbing up the wall to the back of Nicolas house, patiently sitting on roof and waiting until lights go out, preparing darts with a strong liquid deep sleep agent....deep breath.... thhhhhwup thhhhhwup.... miauuu....Feck, I think I hit the cat.... falling of the roof laughing. LOL>

LoL! you're welcome for climbing the wall: 5th floor…

does the Hy6 have something like "darkframesubtraction" to avoid hot pixels with long exposures and noise to be an issue?

This one is for Thierry!

John_Nevill · Feb 29, 2008

Jawdropping!

Its no good, the house goes on the market tomorrow and i'm going to sleep rough with a new sinar!....but then I'll need a lens

.

Georg R. Baumann · Feb 29, 2008

Nicolas Claris said:
Testing my friend! 400 ISO is noise free...

Amazing!

Do you use an external lightmeter, or meter from the Hy6 direct?

As for the images, is it only me? (uncalibrated CRT screen)

On all three shots there is a clearly to be seen "lighter vertical band" to the left and right edge of the shot, also on the 100%crop.

Do you see that as well?

Georg R. Baumann · Feb 29, 2008

John_Nevill said:
Jawdropping!

Its no good, the house goes on the market tomorrow and i'm going to sleep rough with a new sinar!....but then I'll need a lens .

John, are you any good in climbing walls up to the 5th floor? We could work together. LOLOLOL I'll put on my cherokee Regalia and do a rain dance in the front, distracting them while you sneak in and snatch her, but watch the cat. LOLOLOL

Seriously though, Nicolas, another major question came to my mind, how did you manage insuarance for such gear? I don't think that goes with a normal house/content policy. I would imagine this needs a special policy for all cases and that is also probably costly, but a must have.

nicolas claris · Feb 29, 2008

Bad news for you my friend Georg!
You'll have to do same as John and put your cottage on the market ;-):

- new camera
- new web hoster
- new monitor!

Good night! sleep well! nice dreams! LoL!

John_Nevill · Feb 29, 2008

Georg, i'm already climbing them looking at these!

BTW, I studied Shotokan for 10 years, perhaps if Nicolas was to walk this path regularly with the Sinar, then we could hop on the ferry and bump into him - nudge, nudge wink wink!

Anyhow, I couldn't see any banding on my monitor, but perhaps a hot pixel bottom right ~400 pixels in and 800pixels up on the first shot. Although could be a highlight.

Georg R. Baumann · Feb 29, 2008

Hehehe.... as for those vertical bands, you don't have them?

nicolas claris · Feb 29, 2008

Georg Baumann said:
Do you use an external lightmeter, or meter from the Hy6 direct?

No, no external lightmeter, average metering, next time I'll try centered weighted metering. If I remember well, this one was shot 1 stop underexposed, all cells are always fooled at night with such DR… I should maybe have pushed to 1.5 stop…

Georg R. Baumann · Feb 29, 2008

John_Nevill said:
Georg, i'm already climbing them looking at these!

BTW, I studied Shotokan for 10 years, perhaps if Nicolas was to walk this path regularly with the Sinar, then we could hop on the ferry and bump into him - nudge, nudge wink wink!

LOL, nope, let me handle him, I did Aikido many years, Nicolas is a friend I don't want him to get hurt! I'll send him sleep and you handle that darn cat. Hehehe.

Anyhow, I couldn't see any banding on my monitor, but perhaps a hot pixel bottom right ~400 pixels in and 800pixels up on the first shot. Although could be a highlight.

Strange, it is very clear to be seen here.

That insurance thing is really something I am wondering about, just out of curiousity, I am going to call my broker and ask him what they can come up with for gear on that scale.

John_Nevill · Feb 29, 2008

Joking aside, thanks for sharing these Nicolas, no doubt you'll master this tool over the coming months and I'll look forward to seeing some of your new work.

Georg R. Baumann · Feb 29, 2008

Yeah, Nicolas, thank you for doing all that work here! It is really great to learn about your first hand experience with this amazing camera.

Thierry Hagenauer · Mar 1, 2008

Georg Baumann said:
Tell me, I do not understand one thing, why would you choose ISO 400 on a longexposure shot?
I would think, a longer exposure and ISO 100 would create a better file, then again, I might be completly mistaken

Yes, basically it is always better to use the lowest possible ISO and increase the exposure time, rather then increasing the ISO with a shorter time. A higher ISO will always give more noise, but then, as long as it is not hot with temperatures around 30°C, you might be getting good results with ISO 400.

Georg Baumann said:
Also, does the Hy6 have something like "darkframesubtraction" to avoid hot pixels with long exposures and noise to be an issue?

Sure it has: the eMotion back creates "Black References" (at the start of the shooting, then every 15 min OR manually if one wants, OR THEN when the exposure time is changing more than a certain %). These files are saved and later applied (subtracted) automatically to the right images, denoising them.

Eric Hiss · Mar 1, 2008

Nicolas,
Looks like you found your tripod.
Eric

nicolas claris · Mar 1, 2008

Yes Eric… a very stable subject and a quite strong unmoving stone wall (to put the camera on)!

And also some improvement in "derawtization" which I'll check on the "man at work" pic.

What amazed me in the above picture is the distance of the subject from the lens (manual focus BTW) more than 500 meters! (about 1600 feet)

nicolas claris · Mar 1, 2008

Hi all
Thanks to Eric comments, I went back to work on the workflow…

This image is tricky because there are so many different light sources…
I have experienced in the past that in such case, even the excellent grey card from Whitbal cannot be of real help… so it comes all during the derawtization process (better than in PP on tif).

So here's a new try! colors look more natural to me and images seems less blotchy.

I have deleted the hires jpeg from the original post, but, certainly more easy for everybody, have added below a 100% crop. Sarting from a 6600 pixel image to "only" 1100 posted on the web creates a lot of indesirable issues…

100% crop:

The reflection you can see on the window are the houses on the dock 500 meters below the back of the photographer (see the 2nd image posted in this thread)

The green and pink halos you may guess in the image reflected lamps are not CA, these lamps are colored, hence the reflection on the water (La Garonne river) that you can obviously see in the 2nd image posted in this thread

Any comments? is it improved for you?

Asher, do you remind the good wine we had there, sat on the terrace with Wendy and Marine ?;-)

Almost same angle of view!

Bart_van_der_Wolf · Mar 1, 2008

Thierry Hagenauer said:
Yes, basically it is always better to use the lowest possible ISO and increase the exposure time, rather then increasing the ISO with a shorter time. A higher ISO will always give more noise, but then, as long as it is not hot with temperatures around 30°C, you might be getting good results with ISO 400.

I assume the Sinar backs are no different than other CCD-backs, they'll start to noticably accumulate 'Dark current' noise at exposure times of 1 second or longer. It then becomes a trade-off between dark current noise and amplifier noise (due to higher ISO settings). It would require a more controlled test sequence to tell where the tipping point is.

Sure it has: the eMotion back creates "Black References" (at the start of the shooting, then every 15 min OR manually if one wants, OR THEN when the exposure time is changing more than a certain %). These files are saved and later applied (subtracted) automatically to the right images, denoising them.

Does this darkframe subtraction use a single (e.g. most recent versus the actual exposure) frame or does it average several of those reference frames?

Bart

Thierry Hagenauer · Mar 1, 2008

No, it does not average several BR: basically each image has its own BR (most recent versus the actual exposure, as you write it), or at least all images taken between a period of time of 15 min. will use the same BR, until the next BR is created automatically. Or then one can create manually a BR, for each image and before the actual shot, or whenever one wishes.

But BRs are never averaged, that would probably give a big "mess" in terms of noise and de-noise.

Thierry

Bart_van_der_Wolf said:
Does this darkframe subtraction use a single (e.g. most recent versus the actual exposure) frame or does it average several of those reference frames?
Bart

Bart_van_der_Wolf · Mar 1, 2008

Thierry Hagenauer said:
But BRs are never averaged, that would probably give a big "mess" in terms of noise and de-noise.

On the contrary. It's standard best practice in astronomical photography. There it is common practice to average some 10 or more darkframes to create a 'master' darkframe. It results in much lower random noise, both in the master darkframe as in the final result. It will still remove the systematic noise, even more accurately so.

When only 1 darkframe is subtracted, systematic (pattern) noise wil reduce but random noise will increase (!) by a factor of sqrt(2) or +41%.

Bart

Georg R. Baumann · Mar 1, 2008

Interesting!

So a single 15 minutes exposure has more noise than 15x 1 minutes exposures stacked. The reason is that the background light (city lights, skyglow) overwhelms/saturates the sensor. Hence, the longer the exposure the more of that background light interfers with the data that images the object of the picture.

Here is an excellent article that explains what Bart mentioned:

http://www.qsimaging.com/ccd_noise.html

Thierry Hagenauer · Mar 2, 2008

Dear Bart,

I am not convinced this would work the same way in "normal" photography, unless you have a temperature stable environment as for astronomical photography (which is I believe the case).

Best regards,
Thierry

Bart_van_der_Wolf said:
On the contrary. It's standard best practice in astronomical photography. There it is common practice to average some 10 or more darkframes to create a 'master' darkframe. It results in much lower random noise, both in the master darkframe as in the final result. It will still remove the systematic noise, even more accurately so.

When only 1 darkframe is subtracted, systematic (pattern) noise wil reduce but random noise will increase (!) by a factor of sqrt(2) or +41%.

Bart

Bart_van_der_Wolf · Mar 2, 2008

Thierry Hagenauer said:
Dear Bart,

I am not convinced this would work the same way in "normal" photography, unless you have a temperature stable environment as for astronomical photography (which is I believe the case).

Thierry, first of all, thanks for your replies on OPF. It certainly helps to get a better understanding of the progress made by traditional LF manufacturers in the digital arena (and I certainly appreciate you representing Sinar, even outside(?) your geographical area of responsibility).

May I suggest you take the RF/Darkframe issue up with your research people. There may be some improvement opportunities lurking in the software as it (apparently) is.
We can learn a lot from the astro photographers, who have been (or rather needed) pushing the physical capture limits for several decades already (I've sold specific glass plate coated emulsions to a number of them, in the Netherlands all the way to Indonesia since 1978).

Digital capture has created new opportunities that can be utilized by us 'terrestrial' photographers quite well. The only difference is that our common exposures are not as photon starved as theirs, but there is something to be learned when we get into longer exposure times than, say, 1/30th of a second (or when dealing with light fall-off across the sensor array).

Kind regards,
Bart

Georg R. Baumann · Mar 2, 2008

Bart_van_der_Wolf said:
When only 1 darkframe is subtracted, systematic (pattern) noise wil reduce but random noise will increase (!) by a factor of sqrt(2) or +41%.

Hi Bart,

I trust "EMMA" keeps you awake as well? <smile>

When you refer to random noise, do you mean this:

Other sources of noise, such as periodic noise introduced by the camera’s electronics are more difficult to deal with because these types of noise build up just like the signal from the stars you’re trying to image

Thierry Hagenauer · Mar 2, 2008

Sure, Bart, I shall forward this suggestion: this is always welcome.

Best regards,
Thierry

Bart_van_der_Wolf said:
May I suggest you take the RF/Darkframe issue up with your research people. There may be some improvement opportunities lurking in the software as it (apparently) is.
Kind regards,
Bart

Bart_van_der_Wolf · Mar 2, 2008

Georg Baumann said:
Hi Bart,

I trust "EMMA" keeps you awake as well? <smile>

When you refer to random noise, do you mean this:

No, EMMA stopped rearranging the outdoor furniture last Saturday morning, I slept well the preceding night. Did she hit your place on Friday?

Noise comes in many 'qualities'. The truely random kind can be reduced by averaging multiple samples, it increases the Signal/Noise ratio. When the random noise is eliminated, we're left with the non-random noise (systematic or pattern noise, hot/dead pixels, amplifier glow, etc.), and last but not least the actual signal we want.

CCDs, and in particular large ones, also have specific noise patterns due to their 'bucket-brigade' type of reading the data (the read-noise increases with each sensel move in the readout direction). Thanks to Thierry I was pleased to learn that Sinar take precautions and make an attempt to remove that systematic noise. Unfortunately one can also increase random noise in that process of removing systematic noise, hence my remarks on how to improve that process.

Bart

Georg R. Baumann · Mar 2, 2008

Bart_van_der_Wolf said:
No, EMMA stopped rearranging the outdoor furniture last Saturday morning, I slept well the preceding night. Did she hit your place on Friday?

It was less than predicted, but bad enough. Being so close to the conveyor we often get the brunt first, in this case we were battered from Thursday afternoon, top speeds on Friday, and into saturday night, I would guess still gusts up to 130km/h. The worst we had so far in the past 4 years we are here was 2 month or so ago, 174 km/h, really severe. I was on the beach that day out of curiousity, was hoping to get a special shot, but to be honest, it was impossible, the gusts were so strong they nearly knocked me of my feet twice, and that's not funny balancing close to the cliffs. LOL - Won't do that again!-

Talk about climate change, I am just back from the beach, and guess what I found. LOL A flippin coconut, on a beach in Donegal! <grins> As fresh as they come from the palmtree, still full of juice.

Noise comes in many 'qualities'. The truely random kind can be reduced by averaging multiple samples, it increases the Signal/Noise ratio. When the random noise is eliminated, we're left with the non-random noise (systematic or pattern noise, hot/dead pixels, amplifier glow, etc.), and last but not least the actual signal we want.

CCDs, and in particular large ones, also have specific noise patterns due to their 'bucket-brigade' type of reading the data (the read-noise increases with each sensel move in the readout direction). Thanks to Thierry I was pleased to learn that Sinar take precautions and make an attempt to remove that systematic noise. Unfortunately one can also increase random noise in that process of removing systematic noise, hence my remarks on how to improve that process.

Thanks again Bart, I am learning tons from you here!

While going home with the Schneider 40 mm - a night shot

OPF Co-founder/Administrator

Inactive

OPF Co-founder/Administrator

OPF Co-founder/Administrator

Inactive

OPF Co-founder/Administrator

New member

Inactive

Inactive

OPF Co-founder/Administrator

New member

Inactive

OPF Co-founder/Administrator

Inactive

New member

Inactive

pro member

Member

OPF Co-founder/Administrator

OPF Co-founder/Administrator

pro member

pro member

pro member

Inactive

pro member

pro member

Inactive

pro member

pro member

Inactive