Ad
Ad
Ad
Pages: « 1 2 [3] 4 5 »   Bottom of Page
Print
Author Topic: Today's DSLR should have another exposure mode  (Read 27660 times)
Ray
Sr. Member
****
Offline Offline

Posts: 8943


« Reply #40 on: May 26, 2006, 10:30:02 AM »
ReplyReply

Quote
I find it difficult to believe that values of 255 in ACR with -4EC do not represent blown channels.


Bill,
I'm not qualified to comment, except to say one might argue equally that, 'I find it difficult to believe that BB's default linear conversion is not accurate'.

I simply don't know. I'm mainly interested in practical results and this sunrise seems to be closely enough exposed to the right for me. Fortunately (or unfortunately depending on perspective), the foreground in the shadows is uninteresting, drought scorched grass. There's no need to bring out the shadows in this shot.
Logged
jani
Sr. Member
****
Offline Offline

Posts: 1604



WWW
« Reply #41 on: May 26, 2006, 11:23:57 AM »
ReplyReply

Quote
Well - reread the article. Pay extra attention to the rationale for the technique. It's about bits dedicated to the image data.
And what has "bits dedicated to the image data" got do do with detail?

It only affects detail to the extent that detail is affected by exposure.

I think we have a terminology problem here, and that we're not communicating quite well, but I'm obviously not in possession of the necessary vocabulary to improve on the situation.
Logged

Jan
Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #42 on: May 26, 2006, 11:47:11 AM »
ReplyReply

Quote
The latest version of CDRaw has a -D switch, which causes totally raw output with no scaling. There is no demosaicing and the output is gray scale 0..4095. If you want RGB output, the above switches with -o 0 should do the trick.
Bill,

thanks for your explanations. Meanwhile, I did some testing an reading of your postings in the Adobe Forum. You say, you use the following options:

-m -n

So, in one case, we agree, since -m is equal to -o 0. See here, and testing prooves this true.

But in the other case, I think your conversion is faulty. The one reason is obvious:

-n is equal to -H 1, which means, that clipped channels are filled with shades of pink, which is not usefull when evaluating highlights. As long as you have no clipped channels in you selection - no problemo. But I don't like that.

For the other reason, I have to go further into detail. It's all about the multipliers. If you set the option -v, you'll see a line like this during conversion:
Code:
Scaling with black=0, pre_mul[] = 2.270731 1.000000 1.273082 1.000000
In lack of a documentation from Dave Coffin, I interpret those values as the multipliers of the direct values to the sensor data, seen as a square crop containing four pixels:

R G B G

My emphasis is on direct, so if you are out for the unbiased sensor data, you'll have to use the option set

-r 1 1 1 1

If you use the -i option, some data of the image file is listed. At the bottom, there are two interesting lines:
Code:
Daylight multipliers: 2.094750 0.922500 1.174418
Camera multipliers: 271.000000 256.000000 751.000000 256.000000
These are the values from a Raw file of a Konica Minolta A2.

The first line defines the multipliers for the R G B values, which are supposed to balance the colors with daylight. dcraw sets the smallest value to 1 and adjusts the other colors. This multiplier set is used, if you convert without any (relevant) options set. The computed multipliers displayed are the above mentioned:

0.922500 scaled to 1.000000 (G)
2.094750 : 0.922500 = 2.270731 for R and
1.174418 : 0.922500 = 1.273082 for B

In the second line, you find the values for R G B G according to the set WB. It's the same scheme as above:

256.000000 scaled to 1.000000
271 : 256 = 1.058594
751 : 256 = 2.933594 (pretty extreme value, AWB at tungsten)

Exactly those values, you'll see if using the -w option (Use the color balance specified by the camera).

No, you are using the -n option, which gives the following multipliers:
Code:
1.000000 0.440387 0.560648 0.440387
This is a remarkable notation, since usually, the smallest value is set to 1.000000. So even if you type an option like this:

-r 0.5 1 1 1

it's converted to this form:
Code:
1.000000 2.000000 2.000000 2.000000
If you set the smallest multiplier (0.440387) to 1, you'll see that the ratio is the same as with the daylight multipliers (edited). But now he sets the channel with the highest multiplier to one.

Anyway, you see, that using the -n option leads to false R G B values in the image, in the meaning of not as they were recorded. If you want all channels exactly as they were recorded by the sensor, you'll have to use the

-r 1 1 1 1

option, which BTW overrides the -n option.

Quote
The output is scaled by a factor of 16 to convert from 0..4095 (12 bit) to 0..65535 for display at 16 bits. The 12 bit output is very dark in Photoshop 15+1 display (0..32768).
Oh, yes, thanks for this hint. I was confused, since the values shifted considerably due to the sRGB conversion. The -o 0 or -m otpion solved the problem.

To sum it up, I recommend the following set for linear unbiased conversion:

dcraw.exe -3 -m -r 1 1 1 1 {file name}

Any comments?
« Last Edit: May 26, 2006, 05:31:30 PM by Dennis » Logged

Best Regards

Dennis.
John Sheehy
Sr. Member
****
Offline Offline

Posts: 838


« Reply #43 on: May 26, 2006, 01:40:24 PM »
ReplyReply

Quote
Well - reread the article. Pay extra attention to the rationale for the technique. It's about bits dedicated to the image data.

That may be what the article emphasizes, but there is more to it than that.  The signal-to-noise ratio is higher in the upper RAW ranges of a given ISO, including the distracting banding noise that some cameras exhibit in under-exposures.

Quote
It's an overexposure technique geared towards heavy post-processing that benefits from extra bits dedicated to image information, not a guide to correct exposure.

Actually, there really shouldn't be that much post-prcoessing to handle a high RAW exposure.  At least in theory, the exposure slider in a converter merely needs to be moved to the left (not all of them work like this, though).  What requires lots of post-processing sometimes is a "normal" exposure (swhat the camera's metering thinks is normal), and trying topreserve the RAW highlights without darkening the image.
Logged
John Sheehy
Sr. Member
****
Offline Offline

Posts: 838


« Reply #44 on: May 26, 2006, 01:46:50 PM »
ReplyReply

Quote
Oh. Then it's not blown  Sorry for my remark.

I obviously assumed that "white" meant all 255s.
[a href=\"index.php?act=findpost&pid=66594\"][{POST_SNAPBACK}][/a]

All 214/214/214 over an area does indicate clipping; something was clipped somewhere in the process, and then darkened in the final output.
Logged
John Sheehy
Sr. Member
****
Offline Offline

Posts: 838


« Reply #45 on: May 26, 2006, 05:06:02 PM »
ReplyReply

Quote
At base ISO most digital cameras set the gain so that sensor saturation results in a data number near full scale in the analog to digital converter. For example, with my Nikon D200, sensor saturation results in RGB values of 254, 249 and 253 expressed as 8 bits and 4064, 3984, and 4048 in 12 bits as determined by conversion with DCRaw with demosaicing and conversion to an RGB image. In this case, clipping occurs in the sensor. Full scale of 4095 is not reached at base ISO, but AD overflow with a maximum data numberof 4095 would occur at higher ISOs and the clipping would be in the AD converter at 4095.
[a href=\"index.php?act=findpost&pid=66634\"][{POST_SNAPBACK}][/a]

It is still possible that there is another reason for that.  Many DLSRs clip at less than 4095 for various reasons.  Canon 1DmkII cameras slip in the low 3700s, depending on the ISO; 5D cameras clip at 3962 at every ISO, etc.  I've looked at samples of clipped D200 NEFs that had 4095 at ISO 100 in some verticle lines, but a max of 4024 or thereabouts in alternating lines.  The Canon 10D clips lower at ISO 100 and increases slightly up to 800, etc.
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2880



« Reply #46 on: May 26, 2006, 05:20:57 PM »
ReplyReply

see following post
« Last Edit: May 26, 2006, 05:40:55 PM by bjanes » Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2880



« Reply #47 on: May 26, 2006, 05:26:18 PM »
ReplyReply

Quote
Bill,


To sum it up, I recommend the following set for linear unbiased conversion:

dcraw.exe -3 -m -r 1 1 1 1 {file name}

Any comments?
[a href=\"index.php?act=findpost&pid=66644\"][{POST_SNAPBACK}][/a]

Dennis,

Yes, I think those switches are fine for our purpose, but personally I use the -H 1 option (same as -n),  since I do not want any clipping. It seems to me, if you want to evaluate the status of the raw file, you do not want any clippping. The default is -H 0.

In testing with a heavily overexposed Macbeth CC, I saw little, if any, difference between -H 1 and -H 0 switches when using multipliers of 1 and raw output (-o 0). The clipped high lights did not fill with pink with the -H 1 option. I superimposed the two conversions in Photoshop and used the difference blending option, and the screen was totally black--no visible difference. Here is the -H 1 conversion converted to 8 bits and downsized. As you can see, the left two neutral patches of the color checker are blown to white.

[attachment=619:attachment]

I then converted the same files but to sRGB (-o 1) and "as shot" white balance (-w) with and without the highlight clipping (-H 0 and -H 1). Here there is a big difference. Also, the multipliers are different. I think the attachments are self explanatory.

-H 0 -w -o 1

[attachment=621:attachment]

-H 1 -w -o 1

[attachment=620:attachment]

[attachment=625:attachment]
« Last Edit: May 26, 2006, 08:31:59 PM by bjanes » Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2880



« Reply #48 on: May 26, 2006, 05:31:34 PM »
ReplyReply

Quote
All 214/214/214 over an area does indicate clipping; something was clipped somewhere in the process, and then darkened in the final output.
[a href=\"index.php?act=findpost&pid=66651\"][{POST_SNAPBACK}][/a]

Exactly what I said in an earlier post. The sun's disc would most likely have been red in the actual scene and it has been clipped to white. I'm glad someone agrees with me.  
Logged
Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #49 on: May 26, 2006, 08:07:36 PM »
ReplyReply

Quote
Yes, I think those switches are fine for our purpose, but personally I use the -H 1 option (same as -n),  since I do not want any clipping.
But if you have some clipping, your color values in the histogram are corrupt due to pink tinting introduced with the -n or -H option activated.

Quote
In testing with a heavily overexposed Macbeth CC, I saw little, if any, difference between -H 1 and -H 0 switches when using multipliers of 1 and raw output (-o 0)
There is no difference, since any -r setting overwrites any -a, -w or -H (and thus -n) settings. That's what I tried to explain: With -a, -w and -H dcraw set the multipliers to a certain value. But if you set the -r option, those switches are useless, since the -r option now sets the multipliers. So any set like

-n -r 1 1 1 1

or

-H 0 -r 1 1 1 1

or

-w -r 1 1 1 1

are redundand, since -r is stronger than the others.

Quote
I then converted the same files but to sRGB (-o 1) and "as shot" white balance (-w) with and without the highlight clipping (-H 0 and -H 1). Here there is a big difference. Also, the multipliers are different. I think the attachments are self explanatory.
Here, something is obviously wrong. Your cmd screenshot shows, that you processed two different images, not the same with different settings. Again please check the issue with the -r switch in combination with -w or -H.
« Last Edit: May 26, 2006, 08:09:18 PM by Dennis » Logged

Best Regards

Dennis.
bjanes
Sr. Member
****
Offline Offline

Posts: 2880



« Reply #50 on: May 26, 2006, 09:32:38 PM »
ReplyReply

Quote
But if you have some clipping, your color values in the histogram are corrupt due to pink tinting introduced with the -n or -H option activated.

There is no difference, since any -r setting overwrites any -a, -w or -H (and thus -n) settings. That's what I tried to explain: With -a, -w and -H dcraw set the multipliers to a certain value. But if you set the -r option, those switches are useless, since the -r option now sets the multipliers. So any set like

-n -r 1 1 1 1

or

-H 0 -r 1 1 1 1

or

-w -r 1 1 1 1

are redundand, since -r is stronger than the others.

[a href=\"index.php?act=findpost&pid=66677\"][{POST_SNAPBACK}][/a]

I think you are correct here. However, some of the command line switches have changed recently between the version I was previously using (v 8.05) and the most recent one (v 8.18, May 18, 2006). This led to some confusion on my part.  

In the previous version, -m -n -3 caused the file to be converted as raw without white balance and it was not necessary to use the -r 1 1 1 1. In fact, the -r switch previously set the red multiplier only.

In the current version, it is necessary to use the -r 1 1 1 1 switch for raw output without white balance, and -m is no longer sufficient by itself. As you have pointed out, the -n switch is overridden, and I don't see why you previously disagreed with its use in this situation; it is simply ignored. I have compiled the previous version as DCRaw.exe and the current as DCRawb.exe. Here are the verbose reports with the same switches and the same file being processed. As you can see, the results are quite different. At least, the old version showed the multipliers as 16 (for 16 bit output) and not 1 (for 12 bit). A bit confusing. I have not yet figured out a way to get 12 bit output with the -3 switch. Setting the -b to 0.0625 does not work, and one must divide the 15+1 Photoshop values by 8 to obtain the raw data number.

[attachment=626:attachment]

If you look at the valid switches as shown by typing DCRaw without a filename, the -m and -n switches are no longer shown as valid, but they are still processed. At least their meaning has not changed, in contrast to the -r switch.

[attachment=627:attachment]

Quote
Here, something is obviously wrong. Your cmd screenshot shows, that you processed two different images, not the same with different settings. Again please check the issue with the -r switch in combination with -w or -H.
[a href=\"index.php?act=findpost&pid=66677\"][{POST_SNAPBACK}][/a]

I inadvertently uploaded the wrong screen capture, but the same image was indeed processed with the shown results. My post has been edited to show the corrected screen capture. I agree the -w and -H are meaningless with the -r switch.
« Last Edit: May 26, 2006, 09:48:45 PM by bjanes » Logged
Serge Cashman
Full Member
***
Offline Offline

Posts: 200


« Reply #51 on: May 27, 2006, 12:54:51 AM »
ReplyReply

<edit> Post deleted by author.
« Last Edit: May 27, 2006, 01:00:31 AM by Serge Cashman » Logged
gochugogi
Jr. Member
**
Offline Offline

Posts: 80



WWW
« Reply #52 on: May 27, 2006, 10:10:53 PM »
ReplyReply

Cool, I envision a new pic mode "ER" placed on the Command Dial between Sports and Landscape modes of the EOS 30D MK II. It will be buried in the menus for Rebel and 1-series bodies.
Logged

[span style='font-family:Impact']I'm tryin' to think but nuttin' happens  -The Three Stooges
Ma Blessed Digs 'o Net[/span]
Ray
Sr. Member
****
Offline Offline

Posts: 8943


« Reply #53 on: May 31, 2006, 04:13:19 AM »
ReplyReply

Quote
All 214/214/214 over an area does indicate clipping; something was clipped somewhere in the process, and then darkened in the final output.
[a href=\"index.php?act=findpost&pid=66651\"][{POST_SNAPBACK}][/a]

Agreed, unless one happens to have an extremely neutral patch of gray in the image. It's unlikely the centre of the sun is going to be anything less than white, yellow or red, as in the following image of a setting sun which doesn't appear to be clipped.

[attachment=641:attachment]

Having experimented with a few conversions using BB's linear mode, I can see no advantage whatsoever in linear conversion. ACR and RSP seem to do at least an equal job of retaining highlight detail, if not slightly better, and without the hassle of devising and applying an appropriate tone curve.

Consider the following shot of a crocodile's response to a tasty pork chop. Detail at the back of the crocodile's mouth is moderately blown. With a straight conversion, the effect is as though the croc has swallowed a dozen light bulbs that are still working.

Using ACR and dual conversion (one at -4 EC) I think I've managed to tame the glow.

[attachment=642:attachment]

However, I was curious as to how ACR, RSP and BB's linear mode would handle these partially blown highlights. RSP actually seems to produce the most pleasing result. ACR seems to produce patches of almost neutral grey, perhaps due to greater contrast. BB's linear conversion seems to recover less highlight detail in general. (A word of warning: those of a nervous disposition should not click on 'enlarge').

[attachment=643:attachment]

Dave Coffin, the author of DCraw, seems to think that ACR attempts to reconstruct lost highlight detail. That's the impression I also get if I assume that BB's linear conversion mode is providing the true picture.

I'd be interested to learn from those who are using DCraw if they think it can provide greater recovery of highlight detail than commonly used converters like ACR or RSP.

A footnote: I neglected to turn off RSP's 'detail extractor', which is part of the reason the RSP conversion looks better.

All D60 images, BTW.
« Last Edit: May 31, 2006, 04:36:59 AM by Ray » Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2880



« Reply #54 on: May 31, 2006, 08:36:20 AM »
ReplyReply

Quote
Having experimented with a few conversions using BB's linear mode, I can see no advantage whatsoever in linear conversion. ACR and RSP seem to do at least an equal job of retaining highlight detail, if not slightly better, and without the hassle of devising and applying an appropriate tone curve.

However, I was curious as to how ACR, RSP and BB's linear mode would handle these partially blown highlights. RSP actually seems to produce the most pleasing result. ACR seems to produce patches of almost neutral grey, perhaps due to greater contrast. BB's linear conversion seems to recover less highlight detail in general. (A word of warning: those of a nervous disposition should not click on 'enlarge').

Dave Coffin, the author of DCraw, seems to think that ACR attempts to reconstruct lost highlight detail. That's the impression I also get if I assume that BB's linear conversion mode is providing the true picture.

All D60 images, BTW.
[a href=\"index.php?act=findpost&pid=66985\"][{POST_SNAPBACK}][/a]

Linear mapping of the tone values will not improve highlight detail, but will merely remap the lower tone values. Here are some Imitest produced characteristic curves for the Nikon D200 in linear space converted with DCRaw and in gamma 2.2 with Nikon Capture and Adobe Camera Raw (both with default normal tone curves). Also shown is literal encoding for gamma 2.2 and 1.8.

The highlight values are similar, but the three quarter tones in the gamma 2.2 spaces are considerably brighter, as expected. In the quarter tones both ACR and NC roll off the shadow response.

[attachment=644:attachment]

With the D200 and Adobe Camera Raw, when the highlights are set just below 255 in ACR, the raw data number is about 3550 in my tests, not 4095 as one might expect. This is consistent with what John Sheehy has also posted. Since the highlights are placed at 3550 in the raw file, there is some headroom.

As for highlight recovery, I am posting some observations for comment. Here are Stouffer stepwedge exposures in 0.33 EV increments. The linear raw conversion without white  balance (DCRaw) is on the left and the ACR conversion with default settings is on the right. Stof 03 Step 1 gives an ACR reading of 251 and Stof 05 and Stof 06 (with +EV 0.66 and +EV 1 exposure) are blown in ACR with a normal exposure setting. In the DCRaw conversion, Stof 06 is beginning to blow in the green channel, as indicated by the color shift toward white.

[attachment=645:attachment]

With Stof 05, highlight recovery in ACR is possible with a -0.66 EV correction as shown in this screen capture (posted as a GIF to show good text; please ignore the posterization):

[attachment=646:attachment]

With Sfof 06, it is not possible to set the highlights to less than 255 even with -4 EV correction and shadow tones are lost:

[attachment=647:attachment]

Because of the headroom in the raw file, highlight recovery is straight forward with slight "overexposure", since the raw channels are not blown. However, when the green channel starts to blow, highlight recovery is less successful. If one had used a magenta filter to balance the channels, I am not sure that highlight recovery would work very well.
« Last Edit: May 31, 2006, 08:42:33 AM by bjanes » Logged
Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #55 on: May 31, 2006, 12:03:49 PM »
ReplyReply

Quote
However, I was curious as to how ACR, RSP and BB's linear mode would handle these partially blown highlights. RSP actually seems to produce the most pleasing result. ACR seems to produce patches of almost neutral grey, perhaps due to greater contrast.
I see something different. RSP is rendering the clipped colors to grey. ACR is trying to keep a color and tints some areas in a fleshy pink tone, which are neutral in the RSP picture. Then there's a sudden jump to neutral grey patches in some small spots, where apparently the clipping reached a level, which didn't allow ACR to reconstruct any color. This is a known behavior of ACR. But AFAIK it does a better job in reconstructing highlights than any other raw converter.

Quote
Dave Coffin, the author of DCraw, seems to think that ACR attempts to reconstruct lost highlight detail.
And this is obviously the case. I just tested out the limits of my Minolta A2 und compared ACR and dcraw's highlight recovery abilities extensivly. My target was the white wall of a house, lit by cloudy sky. I bracketed in 1/3 f stops from +1.7 EV up to +4.0 EV. First, I zoomed to the wall to get  an white only frame for using dcraw's capability of analysing a picture and setting the right multipliers (option -a). Using these multipliers, I converted the bracketed shots. The last one with good highlight detail was the +2.7 EV exposure. The next exposure (+3.0 EV) suffered completely blown sections of the white wall. A linear conversion (-r 1 1 1 1 -m) showed the problem: While the red and blue channel contained good details, the green channel was completely blown. I was not able to produce any acceptable conversion - even not using the -H [2..9] option, which is supposed to reconstruct the highlights. All it did, was to introduce a massive pink cast. Looking at the channels, it turned out, that the green channel was simply filled with a way too dark shade of grey (R had an average of 168, B of 117 and G was filled with plain 109). So, what should be white wall with fine detail (rough plaster) got a dirty pink wall with greenish shadows.

ACR does a way, way better job recovering highlights: The channels of the same +3.0 EV overexposed shot were evenly filled with equal shades of grey, containing all detail. Since we know from the above linear conversion, that the green channel is absolutely blown, ACR reconstructs the green channel with detail from the two others, and balances the result that way, that i's a neutral grey. This way, ACR is able to reconstruct even the highlight's in a +3.3 EV overexposed shot. Even the +3.7 EV exposure is ACR able to handle pretty well, not blowing out all the white wall, but leaving some detail. Only the +4.0 EV image is as bad, as the +3.0 EV was with dcraw.

So, recovering highlights, ACR wins hands down. It is possible, to get all information out of the Raw file with dcraw using a linear conversion, but the image has a strong color cast then, and I wasn't able to balance the colors.

In terms of resolving of fine detail, dcraw wins by a small margin. If you have some very very fine detail in the picture, as you find it with fabrics, dcraw has an edge, a small one.
« Last Edit: May 31, 2006, 12:10:01 PM by Dennis » Logged

Best Regards

Dennis.
Dennis
Jr. Member
**
Offline Offline

Posts: 82


« Reply #56 on: May 31, 2006, 12:44:51 PM »
ReplyReply

Here I add a sample:
The first row shows the image converted linear with dcraw, the second row converted with highlight recovery of dcraw and the third row the conversion done with ACR. This is a crop of the white wall, exposed at +3.0 EV. One clearly can see, that the green channel is completely blown, and how dcraw and ACR do different things to reconstruct the highlights.
Logged

Best Regards

Dennis.
Ray
Sr. Member
****
Offline Offline

Posts: 8943


« Reply #57 on: June 02, 2006, 12:10:52 AM »
ReplyReply

Quote
So, recovering highlights, ACR wins hands down. It is possible, to get all information out of the Raw file with dcraw using a linear conversion, but the image has a strong color cast then, and I wasn't able to balance the colors.

In terms of resolving of fine detail, dcraw wins by a small margin. If you have some very very fine detail in the picture, as you find it with fabrics, dcraw has an edge, a small one.
[a href=\"index.php?act=findpost&pid=67023\"][{POST_SNAPBACK}][/a]

Thanks, bjanes and Dennis for your response. I tend to agree with Dennis that ACR is (some how) able to extract or reconstruct more color detail in blown highlights. I'm not sure if Capture One is able to better these results, but ACR's reconstruction is mostly credible, although the cyan shift in a blown blue sky could be better. (Or maybe not. What do I know!)

I was recently looking at some rain forest shots with sunlight streaming through. This is a very high DR situation. Tree trunks with direct sunlight are likely to be blown, or the shadows are likely to be very noisy, depending on choice of shutter speed.

In terms of detail from the luminous perspective, there's little difference between RSP and ACR. Howver, with a 'daylight' WB, the ACR conversion brings out the green of the moss on the trunk, which I know was there. I've failed to get RSP to produce that credible green.

Sorry, I'm on my 64 system and wouldn't attempt to post images demonstrating this.

Oops! Forgot to address the second part of the quote. How does DCRaw compare with RSP redarding fine detail extraction?

My impression is, that RSP produces more detail than ACR.
« Last Edit: June 02, 2006, 12:17:40 AM by Ray » Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2880



« Reply #58 on: June 02, 2006, 10:10:35 AM »
ReplyReply

Quote
Thanks, bjanes and Dennis for your response. I tend to agree with Dennis that ACR is (some how) able to extract or reconstruct more color detail in blown highlights. I'm not sure if Capture One is able to better these results, but ACR's reconstruction is mostly credible, although the cyan shift in a blown blue sky could be better. (Or maybe not. What do I know!)

I was recently looking at some rain forest shots with sunlight streaming through. This is a very high DR situation. Tree trunks with direct sunlight are likely to be blown, or the shadows are likely to be very noisy, depending on choice of shutter speed.

In terms of detail from the luminous perspective, there's little difference between RSP and ACR. Howver, with a 'daylight' WB, the ACR conversion brings out the green of the moss on the trunk, which I know was there. I've failed to get RSP to produce that credible green.

My impression is, that RSP produces more detail than ACR.
[{POST_SNAPBACK}][/a]

When dealing with very high DR situations, one should not overlook the possibility of doing a double conversion, one for highlights and another for shadows, and then blending them digitally. Jeff Schewe describes this process in a white paper on the Adobe web site (he has reportedly submitted an updated version, but it is not yet posted) and Steve Bingham also describes a slightly different method.

[a href=\"http://www.adobe.com/digitalimag/pdfs/highlight_recovery.pdf]http://www.adobe.com/digitalimag/pdfs/highlight_recovery.pdf[/url]

http://dustylens.com/luminosity_mask.htm

If the situation is static and several exposures from a tripod mounted camera are possible, then HDR in Photoshop is an even better option.
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8943


« Reply #59 on: June 03, 2006, 10:53:21 AM »
ReplyReply

Quote
When dealing with very high DR situations, one should not overlook the possibility of doing a double conversion

Bill,
I use this technique quite often but not to recover highlight detail, which one can do with a single conversion and a negative EC, but to reduce noise and improve image quality in the shadows. For example, I'll often blend 2 conversions, one of which has a -2EC setting and the other a +2EC. However, the improvements are fairly marginal and this technique is no substitute for 2 separate exposures.

I'd like to find a good algorithm that automatically aligns images which are slightly out of register, which always happens when attempting to blend two handheld shots. I'm surprised that Photoshop's HDR attempt at aligning images is so poor. I can always do a better job manually, although I have some trouble with precise rotation of the top layer.

I would expect that the 1D MK2 with its fast frame rate would be a better tool than the 5D for providing good shots for blending. With the 5D, whatever the exposure, I'm not able to take 2 shots within a shorter period than about 2/3rds of a sec, which is a long time for things to get out of register, whether it's due to movement of the subject or camera shake.

As you've gathered, I don't like carrying a tripod around.
Logged
Pages: « 1 2 [3] 4 5 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad