Experimental Photo Production SX110/120/130, A470

  • 18 Replies
  • 2090 Views
Experimental Photo Production SX110/120/130, A470
« on: 27 / October / 2016, 13:17:05 »
Advertisements
Hi,

I've been posting some of my CHDK inspired images to the Facebook forum and the suggestion was made to post them here as well. Constructive criticism always welcome.

Today's photo is the sunset at Mangawhai, New Zealand, April 4, 2015. 600 shots on an SX120 with an ordinary but accurate intervalometer, interval: 5 s, remetering in between. These shots were grouped by exposure settings and combined using "star trailing" or maximum pixel method described at https://iso.500px.com/time-stack-photo-tutorial/.

Post production involved intensifying contrast and saturation (http://wp.me/p5wbO2-nF) and Advanced Tone Mapping so that the colours are off the chart. The interval should be reduced to make the cloud trails smoother, possibly using continuous shooting with rawopint in CHDK 1.4, which was not yet available when I took this series.



Larger image here: http://www.skeptic.de/CHDK/Expt/Sunset Boulevard.png

Enjoy, Lee

Re: Experimental Photo Production SX110/120/130, A470
« Reply #1 on: 28 / October / 2016, 10:01:16 »
Thanks Lee.  Most of the conversations on this forum are either developers wrestling with the latest camera issues or newbies trying to figure out how to use CHDK for the first time.  Nice to see some actual examples of things done with CHDK.
Ported :   A1200    SD940   G10    Powershot N    G16

Re: Experimental Photo Production SX110/120/130, A470
« Reply #2 on: 13 / November / 2016, 10:20:18 »
I was taking some sunset series with the maximum pixel method and noticed when I was shooting the rising moon, that because the exposure time increased as it became darker, the moon gradually became overexposed, somewhat spoiling the effect:


Then I changed the intervalometer to allow fixed exposure series and took a series of 980 shots, with the moon well exposed in each. The first and last shots:



Now I took a pixel wide strip of each of these shots containing the maximum width of the moon and mounted them from right to left.


Finally  I added back the right hand side from the first picture and the left hand side of the last.



These day2night pictures are somewhat in fashion and you can probably get more out of city photos, but there you go.
« Last Edit: 13 / November / 2016, 10:24:38 by SkepticaLee »

Re: Experimental Photo Production SX110/120/130, A470
« Reply #3 on: 26 / January / 2017, 13:46:52 »
Another project I was glad to finally get done was star-trailing real stars. Most long exposures will be ruined because the camera is warm and this is a source of infrared which is exactly what the ccd picks up. Canon cameras will usually try to compensate this with longer exposures by creating a dark frame of exactly the same length and subtracting it from the image. This means that continuous shooting results in  a gap between shots of the length of taking the dark frame and of processing it.

CHDK allows this mechanism to be switched off. However the shots show clearly the effect of the camera warmth on the sensor:



This glow can be removed by taking the median pixel of a couple of images before the image being shot, the image itself, and a couple of images after the image being shot. This cleans things up considerably:



Some images had to be tidied up because of clouds appearing in the (almost) full moon night. The resulting images can then be stacked using the maximum pixel.



There is a bit of false colour in the regions where the glow was removed. The median pixel method is not so good for photographing the northern sky, as it will remove Polaris, but it works well enough with the southern sky, because there is no south pole star.

Technical details: Canon SX120, CHDK 1.3, ISO 400, 15 seconds, f2.8, intervalometer set to 0 seconds and let to run until the batteries ran out, 405 shots in total.


Re: Experimental Photo Production SX110/120/130, A470
« Reply #4 on: 26 / January / 2017, 17:37:20 »
However the shots show clearly the effect of the camera warmth on the sensor:
The purple glow in the upper left hand corner has been discussed many times on this forum and had always been attributed to an effect called "amp glow" and not camera temperature.   

http://darkerview.com/CCDProblems/ampglow.php

Regardless, this is some nice work on post processing to eliminate it.
Ported :   A1200    SD940   G10    Powershot N    G16

Re: Experimental Photo Production SX110/120/130, A470
« Reply #5 on: 11 / August / 2017, 12:42:25 »
Here are some results for a Powershot A470 mounted on a cell phone gimbal, shooting continuously using rawopint to dynamically adjust exposure. Additional deshaking was applied digitally.

/>

*

Offline reyalp

  • ******
  • 11202
Re: Experimental Photo Production SX110/120/130, A470
« Reply #6 on: 13 / August / 2017, 16:16:25 »
Here are some results for a Powershot A470 mounted on a cell phone gimbal, shooting continuously using rawopint to dynamically adjust exposure. Additional deshaking was applied digitally.
Nice, that looks like it worked quite well. Would benefit from higher resolution, but I guess card space might have been an issue?

Edit:
If you happen to still have settings you used, that might be useful for other users of the script. They are recorded in the first line of the log file.
« Last Edit: 13 / August / 2017, 16:21:45 by reyalp »
Don't forget what the H stands for.

Re: Experimental Photo Production SX110/120/130, A470
« Reply #7 on: 16 / August / 2017, 05:22:17 »
The settings were:
07/08/2017,11:39:57,474520,2,,,,,,,,,,2651,41,41,503272,173,31289408,74,370,0.002843,812,8.264,585,1,1051,,,,,,,,,,init:ev_change_max=96,ev_shift=0,ev_use_initial=true,bv_ev_shift_pct=0,bv_ev_shift_base_bv=false,tv96_long_limit=542,tv96_short_limit=1276,tv96_sv_thresh=542,tv96_nd_thresh=957,sv96_max=699,sv96_target=380,meter_width_pct=75,meter_height_pct=75,meter_step=29,meter_high_thresh=96,meter_high_limit=168,meter_high_limit_weight=200,meter_low_thresh=-168,meter_low_limit=-264,meter_low_limit_weight=200,over_margin_ev=24,over_thresh_frac=30000,under_margin_ev=384,under_thresh_frac=100000,over_weight_max=200,under_weight_max=200,over_prio=0,under_prio=0,histo_step=15,do_draw=false,draw_meter=none,draw_gauge_y_pct=0,smooth=true,meter_top=300,meter_left=400,meter_width=2320,meter_height=1740,meter_x_count=80,meter_y_count=60,histo_samples=31724,/,rawopint,v:0.21,/,platform:a470-102c-1.4.1-4772,Mar,19,2017,08:08:34,/,interval:0,/,cont_mode,/,tv,over,nd:133,/,sd:65535,af_ok:false,fl:6300,efl:38017,zoom_pos:0

As for the resolution, I really wanted to shoot at the highest resolution, but... So much to do, will try the next time!


Re: Experimental Photo Production SX110/120/130, A470
« Reply #8 on: 17 / August / 2017, 01:46:06 »

interval:0,/,cont_mode

With interval = 0 you get a varying interval. Continuous movements will then jerky. The interval depends on the exposure time and the write speed of the SD card. I would take with interval always the fastest possible plus 0.1s - 0.2s.

How great it is in your case and whether it would have led to an improvement of the video is difficult to say. If you are interested, please post the complete log file (rawopint.csv). Then I would make you a plot of the actual interval time.


Re: Experimental Photo Production SX110/120/130, A470
« Reply #9 on: 17 / August / 2017, 16:15:57 »

interval:0,/,cont_mode

With interval = 0 you get a varying interval. Continuous movements will then jerky. The interval depends on the exposure time and the write speed of the SD card. I would take with interval always the fastest possible plus 0.1s - 0.2s.

How great it is in your case and whether it would have led to an improvement of the video is difficult to say. If you are interested, please post the complete log file (rawopint.csv). Then I would make you a plot of the actual interval time.



I don't quite get it. Adding 0.1-0.2 s would only smooth out the relative time interval, say with a shooting interval of 0.6 s:

0.6 + exposure time + 0.2 vs 0.6 + exposure time
exposure time = 0.001 s gives
0.8001 vs 0.6001 (33.33%)
exposure time = 0.02 s (my maximum exposure time) gives
0.82 vs 0.62 (32.26%)

I would have thought that that was negligible.

The only way out that I can think of is setting an absolute interval, but do not know if that is possible with rawopint.

Anyway, the 7zipped file is attached (original file was too large). Will be interested in what you come up with.  Cheers

 

Related Topics