Ambient is definately playing a part here. Some peoples results at +3EV are like a christmas tree and others are not. My +3EV results with LENR off are not that much worse than LENR on so im fairly happy. I didnt compare pre and post update but from previous shots taken pre-update there seems to be less colour noise to my eyes
xpfloyd wrote:
Ambient is definately playing a part here. Some peoples results at +3EV are like a christmas tree and others are not. My +3EV results with LENR off are not that much worse than LENR on so im fairly happy. I didnt compare pre and post update but from previous shots taken pre-update there seems to be less colour noise to my eyes
If ambient is that important, would you able to do say multiple 30s shots for 5 minutes and one 5-minute shot? The first series to see how long it would take for the sensor to heat up to a steady state (similar hot pixel pattern/amount). The one 5-minutes shot is more of a control and I expect it to have much more hot pixel. This is quite a realistic test for time-lapse, star-trail or stack-track. For single shot at night, I imagine it would be good to frame and turn the camera off with the LCD away from the body to cool it down as much as possible (if the two tests above pan out as expected) before taking your shot.
I'm not so sure ambient temperature plays such a huge role. It would certainly be one factor of many I suppose. So when I am out shooting this winter in -35C, I should see absolutely no noise or hot pixels?
Gary Clennan wrote:
I'm not so sure ambient temperature plays such a huge role. It would certainly be one factor of many I suppose. So when I am out shooting this winter in -35C, I should see absolutely no noise or hot pixels?
The temperature dependency for recent Canon sensors with very low thermal noise has been measured by R.N. Clark to about 5 degrees C per doubling of exposure time. So if you can get your camera and battery work at that temperature, you will be able to get very low thermal noise. At long exposures though, the heat buildup in the sensor will soon erode much of the benefit.
I also think this is a valid test because it is controlled with less variables when comparing to a regular shooting. The best way to test it would be with the camera cap on instead of lens cap to remove the lens variable as well. However, I did that same test with the camera cap on and the results came out identical.
To me the only variables are ambient and camera temperatures.
My curiosity was to see the differences between the firmware versions and from the test, it's noticeable that Sony altered how the long exposure files are processed. I was also able to confirm that I should still be using my A7R for astro or nightscapes.
Gary Clennan wrote:
I'm not so sure ambient temperature plays such a huge role. It would certainly be one factor of many I suppose. So when I am out shooting this winter in -35C, I should see absolutely no noise or hot pixels?
if you turn it on and take the very first picture, it should be with minimal noise and hot pixels (dark current noise at its lowest); read and photon noise should still be there. However, this is a continuous LiveView camera which the signal is streaming from the sensor. It would heat the sensor up in no time. Of course, there is an equilibrium point with the heat exchanging and I think that's what most would be interested in knowing, especially for a long night session.
Fred Miranda wrote:
I also think this is a valid test because it is controlled with less variables when comparing to a regular shooting. The best way to test it would be with the camera cap on instead of lens cap to remove the lens variable as well. However, I did that same test with the camera cap on and the results came out identical.
To me the only variables are ambient and camera temperatures.
My curiosity was to see the differences between the firmware versions and from the test, it's noticeable that Sony altered how the long exposure files are processed. I was also able to confirm that I should still be using my A7R for astro or nightscapes....Show more →
I understand the premise of this for sure . But I did just take some images last night and one thing is certain from ISO 100 to 640 Im seeing very good results with no hot pixels on a 30 second exposure and its hot here , maybe 82 last night. When I did hit 800 it was pretty good with a few hot pixels but after that going up it was pretty heavy. Funny thing is I did try ISO 3200 both NR on and OFF and have to say NR on was not to bad it still retained details fairly well. Not great for astro work because of the length of the black frame that I understand but overall not bad . Especially in this heat and shooting quite a bit so heat buildup was there for certain.
The biggest variable for your test I agree is heat and camera heat buildup. Hard to control those variables.
Just from what Im seeing not sure I see a big difference in firmware update but Im not sure either and I would not bet on it.
I can do. It will have to be at the weekend though as I wont get a chance until then hiepphotog wrote:
If ambient is that important, would you able to do say multiple 30s shots for 5 minutes and one 5-minute shot? The first series to see how long it would take for the sensor to heat up to a steady state (similar hot pixel pattern/amount). The one 5-minutes shot is more of a control and I expect it to have much more hot pixel. This is quite a realistic test for time-lapse, star-trail or stack-track. For single shot at night, I imagine it would be good to frame and turn the camera off with the LCD away from the body to cool it down as much as possible (if the two tests above pan out as expected) before taking your shot....Show more →
Not what I said. But if you think ambient is not playing a part how do you explain the difference in fred's and my results?
Genuine question. I'm not sure myself Gary Clennan wrote:
I'm not so sure ambient temperature plays such a huge role. It would certainly be one factor of many I suppose. So when I am out shooting this winter in -35C, I should see absolutely no noise or hot pixels?
Silly thought do you think Sony put the Chroma noise update in the NR on and not the off position. Second silly thought did the put anything in the menu in corrections. Have not looked. Let's face it they do weird stuff. lol
I had a similar thought. I was wondering if they added something in that activated something in lightroom upon import (the way the automatic chromatic aberration profile is applied) I.e they have done nothing to the way the raw is recorded but changed the way lightroom (and possibly c1) reads it?
GMPhotography wrote:
Silly thought do you think Sony put the Chroma noise update in the NR on and not the off position. Second silly thought did the put anything in the menu in corrections. Have not looked. Let's face it they do weird stuff. lol
Stupid question from somebody who has no degree in engineering: Did they just not get it right this time, but could possibly improve things via another fw update in the future or is this a hardware issue that cannot be resolved via updating the firmware?
Btw, re astro photography, the a7r works fine for wide field astro, not sure whether I would use a camera without EFCS to take pictures of deep sky objects through a tele lens or telescope.
martintyler wrote:
>snip<
Secondly, ISO 100 vs 3200 tests. I think the ISO 3200 tests are clouding things here. As the ISO noise is overpowering the hot pixels, which is why most of my tests have been at ISO 100.
>snip<
Remember, ISO 3200, pushed +3ev in LR is basically ISO 256000, so you would expect a lot of noise!
>snip<
Hi martintyler,
Understanding the thermal characteristics of one's sensor a various ISO's and temperatures and exposure durations is vital knowledge to have both in one's shot and post-processing considerations. By way of example, starscapes taken on a hot summer night at ISO 3200 will be noisier than those same exposures taken on a cold winter night. (This phenomenon being not particular to Sony sensors, only.)
What one should be paying attention to here is how in-camera LENR performs for their particular shooting parameters (ISO, exposure length) and environmental factors (as affects sensor temperature noise) and, as the purpose of this thread, how the firmware 1.10 makes a difference to one's output with or without LENR. Tests in this thread demonstrate those effects.
Of course, what would matter more here (IMO) are tests pitting in-camera LENR against dark frame subtraction in post. This article discusses some of the considerations regarding those two choices (using a D800)...
Chris_88 wrote:
Btw, re astro photography, the a7r works fine for wide field astro, not sure whether I would use a camera without EFCS to take pictures of deep sky objects through a tele lens or telescope.
The A7R without EFCS is fine for astro as we shoot with longer exposures and the initial vibration is no longer an issue.
The main issue without EFCS is between 1/4 to 1/100s exposure.
GMPhotography wrote:
Silly thought do you think Sony put the Chroma noise update in the NR on and not the off position. Second silly thought did the put anything in the menu in corrections. Have not looked. Let's face it they do weird stuff. lol
There is a possibility that the chroma noise fix is only triggered in Bulb mode but Sony specifically stated that the fix happens when LENR is "OFF". This is what they wrote on the firmware update page:
"Decreased chroma noise when long exposure NR is set off"
Sorry if this is not 1.0 vs 1.1 firmware, but it seems more relevant here than a new thread...
I noticed a flaw in my previous test - I forgot that continuous shooting mode triggers the 12 bit mode, and the high number of hot pixels at low ISO shown in my previous tests posted were done in continuous mode. So to clarify some of this I did a few more shots. 12/14 bit was confirmed using exiftool.
ISO 100, no lens, body cap on. 30s in manual.
First is single shot (14bit) vs continuous shot (12bit)
As described before there are two things going on, hot pixels, and general noise. It seems 14 bit has more general noise, but few hot pixels. Continuous 12 bit has less general noise, but lots of hot pixels... and strangely bulb mode 12 bit has the lower general noise and few hot pixels.
So, in this particular case bulb mode wins, but what else might you be losing by doing it?
A general point on LENR, or black frames, or NR in post - you may be able to reduce/remove the noise and/or hot pixels - but you are losing information that should be in those pixels, surely all these methods are just work arounds and not solutions. Likewise with the difference between LR and Capture1, LR can't be making up these hot pixels, it would be an odd bug - so what is C1 doing?
I also did another test, taking 13 30s shots. starting from cold and one after the other. No difference between first and last shots. I got bored after 13, maybe more would have shown something?
for anything below 30s, bulb might see some improvement.