Wow.... focus distance reported in EXIF!

[Romy Ocon]

Just posted this at DPR a short while ago:
forums.dpreview.com/forums/read.asp?forum=1029&message=33786590


I noticed it first minutes ago when looking at the EXIF data in my recent Pbase uploads, the focus distance is reported at the bottom! As I recall the shooting scenes, the reported focus distance (FD) appears to be quite right.

1. 5D2 + 500 f4 IS + Canon 1.4x, FD = 79.9 m
www.pbase.com/liquidstone/image/119545118/original

2. 7D + 500 f4 IS + Canon 1.4x, FD = 13.5 m
www.pbase.com/liquidstone/image/119567542/original

3. 7D + 400 2.8 IS + Canon 1.4x TC, FD = 49.7 m
www.pbase.com/liquidstone/image/119633466/original

I checked the File Info for No. 3 in PS CS3, and sure enough the focus distance is there:

File Info/Advanced/EXIF Properties
exif:SubjectDistance: 497/10

It appears that for the focus distance to be reported in my particular shots, the following is the common denominator:

1. Camera used is 7D or 5D2

2. Raw conversion is by DNG 5.6, opened/converted in PS CS3 ACR 4.6 (a conversion of a 7D RAW in LR3 beta sent to me by a friend at DPR also reports the focus distance in the EXIF).

3. All lenses and TCs used were Canon. Not sure if focus distance will be reported in other lens brands, or even in all Canon lenses.

This is a very useful info in birding, as one can form a mental data bank of shooting ranges and the focal length needed to deliver frame filling subjects.

[Edu Lorenzo Jr]

wow indeed! for exif peepers like me (this is where i learn the most), i can only estimate the distance from the lens through the DOF that I see, but of course that is not accurate specially with cropped images.

[Neon Rosell II]

Wow, never checked this before, but when I did it showed in exif data on

1. 40D + 600 mm... processed with ACR5.6 / PS CS4, the previous shots I processed with the previous version of ACR and CS4/CS3 did not show the distance.
www.flickr.com/photos/neon2rosell/4127877158/meta/

2. 5DM2 + 600 mm + 2x Kenko TC ( non-reporting). Processed with ACR 5.6/CS4
www.flickr.com/photos/neon2rosell/4123652363/meta/
it showed 13.7m which was about right.

The 7D + 600mm processed with DPP/CS4 did not show the distance, but with the KF that was posted by Ka Mastah processed with ACR5.6/CS4 showed the data.
www.flickr.com/photos/neon2rosell/4123987530/meta/ 13.5m

Is the data embedded in the photo and would depend on the software whether it shows it or not???
I'll try to process old photos with this new work flow and see if the distance would come out ;D

[Romy Ocon]

Nov 24, 2009 11:44:23 GMT Neon Rosell II said:

Is the data embedded in the photo and would depend on the software whether it shows it or not???

Yes, if the EXIF is left intact, most EXIF readers will show the distance data.

I've been playing with a lot of my captures, and here's what I know from my end and from report of others so far:

1. Only Adobe converters (DNG 5.6, ACR 5.6, and LR 2.6 - 3.0) are passing on the distance information. None in DPP, and in straight from the camera jpegs.

2. Lenses with distance info:

800 L IS
600 f4 IS
500 f4 IS
400 2.8 IS
100-400 IS
400 5.6L

3. Lenses without distance info:

Sigmonster
100 2.8 USM macro
EFS 10-22

4. It appears that the distance info is accurate up to 79.9 m only. Distances a little beyond that will initially appear as still 79.9 m, and once it is far enough, as infinity.

5. Closer distances yield more accurate results.

6. Cameras verified to record distance info in EXIF:

5D2
7D
40D

[Romy Ocon]

Neon, I can't seem to generate the distance info in my 40D RAWs, but I see that it shows in your flickr pic?


EDIT:

Ok, I was able to sort the issue. I opened the CR2 in ACR 4.6 (which can open the 40D's RAW), and the distance info wasn't recorded.

But when I converted the 40D RAW with DNG 5.6 before converting with ACR 4.6, the distance info was generated, wohoo!!!

Now, let me check my other cameras.

[ppaaoolloo]

Yet another reason for me to move to Light Room... I always wanted to know how far I was focusing.

[Neon Rosell II]

Ka Mastah,

I checked all my previous 40D + 600 mm also with 2x Kenko TC captures using PS Bridge and it shows the distance information. So, depending on the software you are using will the distance info come out.

Also found out that the maximum distance it could generate is 98.3m farther than this the it shows 4294967295.0m

The 300 mm L f4 IS can only show the distance between 1.5m and 7.7m farther shots shows the same 4294967295.0m

[Romy Ocon]

No distance info in my 20D and 1Dinosaur2.

[Romy Ocon]

Nov 24, 2009 12:40:10 GMT Neon Rosell II said:

Ka Mastah,

I checked all my previous 40D + 600 mm also with 2x Kenko TC captures using PS Bridge and it shows the distance information. So, depending on the software you are using will the distance info come out.

Also found out that the maximum distance it could generate is 98.3m farther than this the it shows 4294967295.0m

Ok, looks like the maximum distance reliably recorded before it rolls into infinity (the big number you listed) is dependent on the lens used.

[ppaaoolloo]

I always wondered why they never added subject distance to the EXIF.

[Eric Patdu]

Wow! This is great! Wonder is the data will also show on my gear. Will check it out.

UPDATE: The data shows on my "beginner's" XSi (450D) + EFS 55-250mm & EFS 18-55mm! Cool!

When focusing on a very distant object, the "Subject Distance" field is always 4294967295.0 m.

[Ralf Nabong]

I also saw it in LR when I upgraded to LR 2.6. I was wondering what it meant, didn't pay much attention and forgot to ask. My bad!

[Teddy Regpala]

Great find mastah! being an EXIF-peeper myslef, I never noticed this before, though the information is right there front and center!

Fo Nikon NEF images, converted by the Nikon software (ViewNX or CaptureNX / NX2), I realized that the information is not preserved in the Subject Distance Range (Tag-ID A40C, under the Camera tab). But it can be found under the Focus Distance (Tag-ID 0098, under the Manufacturer Notes tab). Not sure though if this information is used by Photoshop.

PhotoMe displays the information correctly, but Opanda can't (displays coded information).

[Mark Itol]

Hi guys,

Sorry to resurrect an old thread, but has anyone of you noticed some discrepancies with the subject distance reported? On my captures with 40D and 400mm f/5.6, converted using ACR 5.6, I notice I'm getting only 6.5m, 11.5m, or 76.6m (maximum I got so far), though I haven't checked all of my captures with varying distances yet.

Also, at one instance (during the Taiga Bean Goose shoot), it reported 76.6m when the subject was only around 25-30m as per the lens' focus distance scale which I got to check right after the capture. 25-30m also seemed to be the more correct figure if based on my own estimate of the distance. I've had two cases where it reported 76.6m instead of 25-30m:

1) AI Servo, center AF point
2) One Shot AF, AF point above center

I haven't done controlled testing yet and I will still have to check other files, but 76.6m vs. 25-30m seems too much a difference even if we say the system isn't 100% accurate.

Any ideas? Thanks.

[Romy Ocon]

Mark,

After a few months of looking at the distance reported with various lenses, I believe the computation is based on DOF. This means that it reports more accurate information for longer lenses because of the inherent thinner DOF of such, especially at closer distances. Most of my far shots has a maximum distance reported at 79.9 m, and if longer than that it's reported as infinity.

Try it on shorter lenses and the infinity distance is reached much sooner because of the deeper DOF. The possible explanation to your bean goose shot is the DOF is similar at 30 m - 76.6 m for the focal length and Av you used.

Bottom line is: the focus distance reported becomes more accurate as the focal length increases and the distance gets nearer because of the thinner DOF.

[Mark Itol]

Thanks, Ka Mastah. I knew you will have an insight on this.

Using some online DOF calculators, I checked the DOF of 1.6x CF body, 400mm at f/8 (Av used in Bean Goose shots) for distances of 30m and 76.6m, but the results were not equal. Perhaps the FD computation is bracketing the DOF figures and the DOF figures for 30m and 76.6m fall on the same bracket, and are being interpreted as 76.6m.

I'll probably just try testing on various distances and see what figures will appear.