| Author | Thread |
|
|
05/15/2010 10:59:30 PM · #1 |
Originally posted by glockguy:
It would then be ILLEGAL to stack 10 images of the exact same composition, ie a field of flowers with mountains in the background and a cloudless sky, where each exposure has a different field of focus to then output a final composite with greatly extended depth of field? |
Under the current wording, this (extending DOF with the same somposition otherwise) is specifically LEGAL, just as the "same object" (e.g. start trails) appearing in multiple locations is ILLEGAL.
Originally posted by Advanced Rules (as of 5/15/2010):
The intent of allowing multiple captures is to enable such techniques as high dynamic range (HDR), noise reduction, increased DOF, etc., but not to permit a subject from one scene to be inserted into a different scene, nor is it intended to allow a subject to appear in multiple places within a scene. |
|
|
|
|
05/15/2010 10:42:55 PM · #2 |
Yeah. This is basically where I'm coming from. Things like star trails, tail/head light trails, etc. are pretty easy to interpret as uncontrollable. Random people in the background most likely ok. Closer and more prominent as a subject, possibly not. But then, I suppose this is where the SC would debate.
Originally posted by kirbic:
It *is* possible, even easy, however, to determine if the photographer took a continuous series of exposures to duplicate the effect of a single longer exposure, which I'd argue should be legal under the existing rules. |
|
|
|
|
05/15/2010 07:28:42 PM · #3 |
Originally posted by glockguy:
Just throwing something out there, and this has no effect on long exposure but, on the stacking images idea....
It would then be ILLEGAL to stack 10 images of the exact same composition, ie a field of flowers with mountains in the background and a cloudless sky, where each exposure has a different field of focus to then output a final composite with greatly extended depth of field?
Or is this a moot point because there is no apparent movement in the composition, and I am within the 10 image limit?? |
However people answer in this thread to appear consistent, that technique has been used and validated so you are good to go.
 |
|
|
|
05/15/2010 07:18:53 PM · #4 |
Just throwing something out there, and this has no effect on long exposure but, on the stacking images idea....
It would then be ILLEGAL to stack 10 images of the exact same composition, ie a field of flowers with mountains in the background and a cloudless sky, where each exposure has a different field of focus to then output a final composite with greatly extended depth of field?
Or is this a moot point because there is no apparent movement in the composition, and I am within the 10 image limit?? |
|
|
|
05/15/2010 06:21:19 PM · #5 |
Originally posted by kirbic:
Originally posted by slickwilly:
hey kirbic. i've been looking into stacking photos like you did in that first shot with the streaking clouds. how did you stack them?
I know there is a script for it from dr. brown but i can't get that working yet. |
The technique is really quite simple... load the images as a stack of layers. The order they are loaded in doesn't matter, unless you are planning to do an animation. If you are working in Ps CS4 (I think this works in CS3 as well) you can load images to a stack by selecting File>Scripts>Load Files into Stack.
Once you have the stack of files loaded, set the opacities of the layers as follows:
- Leave bottom layer at 100%
- Second from bottom at 50%
- Third from bottom at 33%
and so on. The opacity is always 1 divided by the layer number, counting from the bottom.
The end result will be equivalent to an exposure of the total length of the sub-exposures, with random noise reduced by a factor of N^(0.5) (that's the square root of N). |
Even if this technique is deemed by SC to be illegal for challenges, this is great info for life outside of DPC. Thanks for the info... |
|
|
|
05/15/2010 03:38:47 PM · #6 |
Originally posted by kirbic:
It *is* possible, even easy, however, to determine if the photographer took a continuous series of exposures to duplicate the effect of a single longer exposure, which I'd argue should be legal under the existing rules. |
I still believe the "continuous series of exposures to duplicate the effect of a single longer exposure" clause has to be added to the current rule for clarification.
As an illustration, take the case of a single cloud starting at 'frame center' and then moving across a featureless sky and out of the frame. The final composite would be a photo of a blur across the sky. However, the first frame would be of a cloud in the sky, and the final frame would be of a blank sky - a major compositional change. Adding your 'long exposure' clause to the current rules would eliminate the ambiguity.
|
|
|
|
05/15/2010 02:26:09 PM · #7 |
Originally posted by PGerst:
In all cases, it refers to the photographer who may not. The implication, as I understood it from the last thread, is to anything the photographer has direct control over. Earth rotations, movement of vehicles, water, etc. are not under the control of the photographer.
If the photographer moves a person, moves a table, bus, or camera, then, it is against the rules. |
First let me say that I would *love* the creativity that this interpretation would allow. Now the hard part; it has a flaw. There's no way to tell whether the photographer directed movement of some subjects. Also, the photographer may select exposure timing to produce duplicates of an object or objects, or to eliminate objects from a scene. Interpretation of photographer intent and control using just the source and final images is not always possible.
It *is* possible, even easy, however, to determine if the photographer took a continuous series of exposures to duplicate the effect of a single longer exposure, which I'd argue should be legal under the existing rules. |
|
|
|
05/15/2010 02:19:29 PM · #8 |
Originally posted by slickwilly:
hey kirbic. i've been looking into stacking photos like you did in that first shot with the streaking clouds. how did you stack them?
I know there is a script for it from dr. brown but i can't get that working yet. |
The technique is really quite simple... load the images as a stack of layers. The order they are loaded in doesn't matter, unless you are planning to do an animation. If you are working in Ps CS4 (I think this works in CS3 as well) you can load images to a stack by selecting File>Scripts>Load Files into Stack.
Once you have the stack of files loaded, set the opacities of the layers as follows:
- Leave bottom layer at 100%
- Second from bottom at 50%
- Third from bottom at 33%
and so on. The opacity is always 1 divided by the layer number, counting from the bottom.
The end result will be equivalent to an exposure of the total length of the sub-exposures, with random noise reduced by a factor of N^(0.5) (that's the square root of N). |
|
|
|
05/15/2010 01:55:28 PM · #9 |
I've read, well, skimmed through the responses below, and I can't figure out what the commotion is all about. This was discussed long ago, for the purposes of allowing HDR effects.
The wording seems pretty direct:
"You must create your entry from 1-10 captures of a single scene (defined as a scene whose composition/framing does not change)".
"You may not combine captures of different scenes, move or change a feature between frames, or combine different captures to create a new scene".
Now, reword the 2nd quote to put "you may not" before each listed item:
You may not combine captures of different scenes
You may not move or change a feature between frames
You may not combine different captures to create a new scene
In all cases, it refers to the photographer who may not. The implication, as I understood it from the last thread, is to anything the photographer has direct control over. Earth rotations, movement of vehicles, water, etc. are not under the control of the photographer.
If the photographer moves a person, moves a table, bus, or camera, then, it is against the rules. |
|
|
|
05/15/2010 12:58:55 PM · #10 |
hey kirbic. i've been looking into stacking photos like you did in that first shot with the streaking clouds. how did you stack them?
I know there is a script for it from dr. brown but i can't get that working yet.
|
|
|
|
05/15/2010 12:13:34 PM · #11 |
OK, a little later than expected, but here's my demo. This was shot in response to  GeneralE's post implying that a stack of exposures of moving clouds would not result in apparent smooth motion as would a long exposure. The test was performed using a 10-stop ND filter to achieve 10s exposures in mid-day sun. Ten exposures were made at as short an interval as possible. The exposures were stacked in Ps CS4 using techniques common to what would be used for star trail stacking. Here's the stack, and an animation of the individual frames:
 
Notice that in the stack, the clouds are perfectly smeared along the direction of travel, just like they would be in a 100s exposure. In fact, there is *no* difference between the results except the unavoidable and unnoticeable inter-frame delay (less than 1/2 second between frames). Finally, here are the individual frames:
    
    
An interesting thing about Frame 8 is that a bird glided casually through the frame, relatively near the camera, yet the bird is not visible on either the single frame nor on the composite. Notice how the movement of the branches in the composite is also more pronounced than in any single frame. Again, the captured movement is the same as what would have been captured in a 100s exposure.
This same idea goes for shooting moving water, and is a "best practice" technique for increasing apparent exposure time beyond what you are capable of achieving with an ND filter alone. |
|
|
|
05/15/2010 04:36:14 AM · #12 |
So, here's a very recent challenge photo that used exactly this technique:
And I quote "combined from 10 four minute exposures for star trails and to avoid sensor noise. Light painted during exposures with variuos(sic) colored gels on LED flashlight"
So, it's looks like we have a candidate to prove the SC's position on this. If it isn't DQ'd then this is apparently legal, if it is DQ'd, this is clearly a bad idea.. :) |
|
|
|
05/14/2010 09:05:37 PM · #13 |
Originally posted by DrAchoo:
I wonder if multiple exposures will avoid purple artifact from sensor heat? Or will it heat up anyway since it's open the same time. |
Excellent question. I just did some digging and found this: //www.cloudynights.com/ubbthreads/showflat.php/Cat/0/Number/3718026/page/5/view/collapsed/sb/5/o/all/fpart/1 In addition to making me really want to get my 5D MKII modified with a thermistor, it provides some very useful information. :) Check out his second graph. Those little spikes are 5 minute exposures with 15 seconds of dead time in between. Any drop in sensor temperature between frames while shooting consecutive exposures would be negligible, so you'd still have amp glow to deal with.
However, you could then shoot a dark frame (lens cap on, whatever) as your last exposure (when the sensor is hottest), throw that onto its own layer in Photoshop, set the blending mode to Darken, and try and get rid of the amp glow that way. At the very least, the multiple exposure technique would cut down on the amount of conventional noise you'd get.
Oh, and just another little tidbit (albeit slightly tangential) while we're talking about temperature. Several newer cameras (including the Canon 5D MKII) embed what's essentially the outside temperature in the EXIF data. If you don't already have it, grab exiftool from //www.sno.phy.queensu.ca/~phil/exiftool/ and check out the amazing wealth of information in your EXIF. :D For the 5D MKII, the temp value is "Camera Temperature". Not nearly as cool as knowing how hot the sensor is, but might be useful nevertheless. Besides, then you can enter into the exciting world of ThermoTagging ( //www.picurl.org/blog/tag/exif/ ). :) |
|
|
|
05/14/2010 05:42:36 PM · #14 |
Originally posted by kirbic:
Originally posted by DrAchoo:
My experience is that even four exposures stacked greatly reduce noise and in a way that's better that NI. |
Correct. Random noise (bot not fixed pattern noise) is reduced by the square root of the number of exposures. So four exposures gives a 2x reduction, 16 exposures gives a 4x reduction. Doing dark frames manually can give added benefit.
FWIW, for star trails, noise is much less an issue since they can and should be shot at base ISO, and you still will have to stop down to ensure you record good color in your stars. Remember that the "exposure" of the stars is affected not only by your ISO and aperture settings, but also by their rate of movement across the sensor (and so are affected, oddly enough, by focal length used). The trail brightness is *not* affected by exposure time, just the trail length. |
I wonder if multiple exposures will avoid purple artifact from sensor heat? Or will it heat up anyway since it's open the same time.
|
|
|
|
05/14/2010 05:00:09 PM · #15 |
| I shot a little daytime long-exposure demo today (moving clounds), we'll see how it comes out. I hope to get it posted later this evening. |
|
|
|
05/14/2010 04:59:00 PM · #16 |
Originally posted by DrAchoo:
My experience is that even four exposures stacked greatly reduce noise and in a way that's better that NI. |
Correct. Random noise (bot not fixed pattern noise) is reduced by the square root of the number of exposures. So four exposures gives a 2x reduction, 16 exposures gives a 4x reduction. Doing dark frames manually can give added benefit.
FWIW, for star trails, noise is much less an issue since they can and should be shot at base ISO, and you still will have to stop down to ensure you record good color in your stars. Remember that the "exposure" of the stars is affected not only by your ISO and aperture settings, but also by their rate of movement across the sensor (and so are affected, oddly enough, by focal length used). The trail brightness is *not* affected by exposure time, just the trail length. |
|
|
|
05/14/2010 03:49:21 PM · #17 |
Originally posted by coryboehne:
Originally posted by cutlassdude70:
Originally posted by GeneralE:
Moving clouds don't link up like a series of individual rail cars to form a picture of a train. I'm pretty sure that when you blend a series of pictures of moving clouds you get a composite which "looks like clouds" but not exactly the same as any of the source images. They don't form a "new shape" in nearly the same way as the star trails.
I'm just saying how I think the rule is now. No one is stopping anyone from posting a suggested new wording, which might be more helpful than just complaining about the current, obviously clarity-deficient rule ... |
Here you go:
You must:
* create your entry from 1-10 captures of a single scene (defined as a scene whose composition/framing does not change). All captures used must be shot within the challenge submission dates.
You may:
* split a long exposure into shorter consecutive exposures if you could obtain the same result with a single long exposure of equal total length.
This can obviously be refined, but gives us a starting point. Thoughts everyone? |
Do the math.. 10 exposures..
So what, for a 30 minute exposure, that's going to be 3 minutes per image? With no LENR? Maybe on your 5D, but not on much else.
If you want to do this for the proposed purpose, you need to significantly increase the 10 exposure limit, perhaps 200 exposures? Still though, that's probably insufficient.. It would be better if you were able to use say, 1800 or so images for your star trails.. That would be enough, as a 1 second exposure should be pretty nice as far as noise is concerned..
Or maybe unlimited exposures.
Then again, on the other side of the fence, this is DPC, and there should be some limit to the rules, and thankfully we are unrestricted by these rules in our day-to-day life...
In any case, I don't actually see the stars moving in the frame as a change of composition. Whatever. I have the patience for one exposure trails using LENR, so I don't really need to be able to stack trails for DPC. Would raise the bar a bit though. :) |
My experience is that even four exposures stacked greatly reduce noise and in a way that's better that NI.
|
|
|
|
05/14/2010 03:11:47 PM · #18 |
Originally posted by coryboehne:
If you want to do this for the proposed purpose, you need to significantly increase the 10 exposure limit, perhaps 200 exposures? Still though, that's probably insufficient.. It would be better if you were able to use say, 1800 or so images for your star trails.. That would be enough, as a 1 second exposure should be pretty nice as far as noise is concerned..
Or maybe unlimited exposures. |
Have fun submitting 1800 RAW files for validation within the 48-hour time limit ... ;-) |
|
|
|
05/14/2010 02:57:25 PM · #19 |
Originally posted by cutlassdude70:
Originally posted by GeneralE:
Moving clouds don't link up like a series of individual rail cars to form a picture of a train. I'm pretty sure that when you blend a series of pictures of moving clouds you get a composite which "looks like clouds" but not exactly the same as any of the source images. They don't form a "new shape" in nearly the same way as the star trails.
I'm just saying how I think the rule is now. No one is stopping anyone from posting a suggested new wording, which might be more helpful than just complaining about the current, obviously clarity-deficient rule ... |
Here you go:
You must:
* create your entry from 1-10 captures of a single scene (defined as a scene whose composition/framing does not change). All captures used must be shot within the challenge submission dates.
You may:
* split a long exposure into shorter consecutive exposures if you could obtain the same result with a single long exposure of equal total length.
This can obviously be refined, but gives us a starting point. Thoughts everyone? |
Do the math.. 10 exposures..
So what, for a 30 minute exposure, that's going to be 3 minutes per image? With no LENR? Maybe on your 5D, but not on much else.
If you want to do this for the proposed purpose, you need to significantly increase the 10 exposure limit, perhaps 200 exposures? Still though, that's probably insufficient.. It would be better if you were able to use say, 1800 or so images for your star trails.. That would be enough, as a 1 second exposure should be pretty nice as far as noise is concerned..
Or maybe unlimited exposures.
Then again, on the other side of the fence, this is DPC, and there should be some limit to the rules, and thankfully we are unrestricted by these rules in our day-to-day life...
In any case, I don't actually see the stars moving in the frame as a change of composition. Whatever. I have the patience for one exposure trails using LENR, so I don't really need to be able to stack trails for DPC. Would raise the bar a bit though. :) |
|
|
|
05/14/2010 02:41:50 PM · #20 |
Originally posted by GeneralE:
Moving clouds don't link up like a series of individual rail cars to form a picture of a train. I'm pretty sure that when you blend a series of pictures of moving clouds you get a composite which "looks like clouds" but not exactly the same as any of the source images. They don't form a "new shape" in nearly the same way as the star trails.
I'm just saying how I think the rule is now. No one is stopping anyone from posting a suggested new wording, which might be more helpful than just complaining about the current, obviously clarity-deficient rule ... |
Here you go:
You must:
* create your entry from 1-10 captures of a single scene (defined as a scene whose composition/framing does not change). All captures used must be shot within the challenge submission dates.
You may:
* split a long exposure into shorter consecutive exposures if you could obtain the same result with a single long exposure of equal total length.
This can obviously be refined, but gives us a starting point. Thoughts everyone? |
|
|
|
05/13/2010 08:06:43 PM · #21 |
Originally posted by GeneralE:
Moving clouds don't link up like a series of individual rail cars to form a picture of a train. I'm pretty sure that when you blend a series of pictures of moving clouds you get a composite which "looks like clouds" but not exactly the same as any of the source images. They don't form a "new shape" in nearly the same way as the star trails. |
You obviously haven't tried this, have you?
Originally posted by GeneralE:
I'm just saying how I think the rule is now. No one is stopping anyone from posting a suggested new wording, which might be more helpful than just complaining about the current, obviously clarity-deficient rule ... |
Paul, if you read my earlier posts, you'll realize that I *have* proposed a more logical way to interpret the rule. |
|
|
|
05/13/2010 07:04:10 PM · #22 |
Originally posted by GeneralE:
Moving clouds don't link up like a series of individual rail cars to form a picture of a train. I'm pretty sure that when you blend a series of pictures of moving clouds you get a composite which "looks like clouds" but not exactly the same as any of the source images. They don't form a "new shape" in nearly the same way as the star trails.
I'm just saying how I think the rule is now. No one is stopping anyone from posting a suggested new wording, which might be more helpful than just complaining about the current, obviously clarity-deficient rule ... |
So just to be clear. Are you saying that if you stacked all these images on top of one another the clouds would not "form a new shape" or would do so in some fundamentally different way than star trails?
For the record I think that idea is silly.
Message edited by author 2010-05-13 19:05:19.
|
|
|
|
05/13/2010 06:52:48 PM · #23 |
Moving clouds don't link up like a series of individual rail cars to form a picture of a train. I'm pretty sure that when you blend a series of pictures of moving clouds you get a composite which "looks like clouds" but not exactly the same as any of the source images. They don't form a "new shape" in nearly the same way as the star trails.
I'm just saying how I think the rule is now. No one is stopping anyone from posting a suggested new wording, which might be more helpful than just complaining about the current, obviously clarity-deficient rule ... |
|
|
|
05/13/2010 06:41:35 PM · #24 |
Originally posted by LadyK:
and its not necessarily for DPC so the legality doesnt matter but thanks everyone for covering that area as well:) |
What?? There's life outside DPC?? Thanks for the reminder Katherine! |
|
|
|
05/13/2010 06:33:34 PM · #25 |
woh, this thread has grown exponentially since i checked it yesterday! thank you everyone for the responses, ill try a photo tonight. thanks for the reminder on Mirror lockup, whoever mentioned it. and i have a stupid remote shuter clicker(it only works from like 5 inches away) and may need new batteries, but ill try and use it too. ill read the rest of the thread as well lol
and its not necessarily for DPC so the legality doesnt matter but thanks everyone for covering that area as well:) |
|
