why do i have "funky" numbers in ss and f stop?

[annmpacheco]

I updated latest version earlier today and I have noticed unusual numbers in "info" example: Aperture: 5.7
Shutter: 1/790?  What is this, not familiar with these numbers....  thanks

[Luiz Muzzi]

The same thing is happening with me (PM build 15011 and OS 10.8.5).
Weird numbers for f/ and shutter speed. If I am not mistaken it started with the new build.
TIA,

-Luiz Muzzi

[Kevin M. Cox]

Photo Mechanic recently started using the APEX values for shutter speed and aperture.
http://en.wikipedia.org/wiki/APEX_system#Use_of_APEX_values_in_Exif

There was talk of making an option to see the "real" APEX value or the value your camera was set to, but I'm not sure if that was implemented.

[jwvaughn]

Win 7, PM 15011

I had not noticed this and checked.  What I found is that the values are "weird" on my original files but fine on files saved as PSD or jpg from Photoshop.

I would think they should at least be consistent!

Jerry

[dennis]

I had not noticed this and checked.  What I found is that the values are "weird" on my original files but fine on files saved as PSD or jpg from Photoshop.

I would think they should at least be consistent!

As Luiz pointed out, some cameras now write newer Exif tags for Shutter Value and Aperture Value.  These are APEX numbers and are the ACTUAL values used by the camera and not the displayed settings.  Photoshop (depending on the version) is consistently only writing the old Exif values.  I thought that it would be useful for people to see the actual values when they are available with the new Exif tags, but this is apparently causing a lot of confusion and grief.  Of course the actual values aren't always consistent either based on the settings (the camera may vary them slightly).

For example, an aperture setting of f/1.4 is really supposed to be sqrt(2) = 1.414213562 etc.  The camera rounds this to 1.4.

An aperture setting of f/5.6 is really supposed to be sqrt(2^5) = 5.656854249 etc which should round up to 5.7 but is instead truncated to 5.6 for some reason.

Interestingly, an aperture setting of f/13 is ambiguous.  It could mean a 1/3 stop adjustment past f/11 of sqrt(2^(22/3)) (actual value of 12.699208416).  It could also mean a 1/2 stop adjustment past f/11 of sqrt(2^(15/2)) (actual value of 13.454342644).  A user would not be able to tell the difference using the rounded camera setting value but they are actually different by 1/6 stop.

Since this is causing a lot of confusion I will likely change PM so that it will show the setting (old Exif) when present and ignore the actual (new Exif) value if present.  However, some new cameras only write the new Exif value and so they will show the actual value (unless I were to try and do the same rounding that the cameras do for their setting).  Another option is to have a user preference to "show actual values for aperture and shutter speed when available" but this involves a UI change.  Another option is to have separate variables for actual versus setting.  The easiest of these options is to continue to show the old Exif setting if it is present, and to show the new Exif APEX value if the old Exif value is not available.

Similar issue with shutter (1/32 -> 1/30, 1/64 -> 1/60, 1/128 -> 1/120, 1/512 -> 1/500, etc).

HTH...

--dennis

[Luiz Muzzi]

I had not noticed this and checked.  What I found is that the values are "weird" on my original files but fine on files saved as PSD or jpg from Photoshop.

I would think they should at least be consistent!

As Luiz pointed out, some cameras now write newer Exif tags for Shutter Value and Aperture Value.  These are APEX numbers and are the ACTUAL values used by the camera and not the displayed settings.  Photoshop (depending on the version) is consistently only writing the old Exif values.  I thought that it would be useful for people to see the actual values when they are available with the new Exif tags, but this is apparently causing a lot of confusion and grief.  Of course the actual values aren't always consistent either based on the settings (the camera may vary them slightly).

For example, an aperture setting of f/1.4 is really supposed to be sqrt(2) = 1.414213562 etc.  The camera rounds this to 1.4.

An aperture setting of f/5.6 is really supposed to be sqrt(2^5) = 5.656854249 etc which should round up to 5.7 but is instead truncated to 5.6 for some reason.

Interestingly, an aperture setting of f/13 is ambiguous.  It could mean a 1/3 stop adjustment past f/11 of sqrt(2^(22/3)) (actual value of 12.699208416).  It could also mean a 1/2 stop adjustment past f/11 of sqrt(2^(15/2)) (actual value of 13.454342644).  A user would not be able to tell the difference using the rounded camera setting value but they are actually different by 1/6 stop.

Since this is causing a lot of confusion I will likely change PM so that it will show the setting (old Exif) when present and ignore the actual (new Exif) value if present.  However, some new cameras only write the new Exif value and so they will show the actual value (unless I were to try and do the same rounding that the cameras do for their setting).  Another option is to have a user preference to "show actual values for aperture and shutter speed when available" but this involves a UI change.  Another option is to have separate variables for actual versus setting.  The easiest of these options is to continue to show the old Exif setting if it is present, and to show the new Exif APEX value if the old Exif value is not available.

Similar issue with shutter (1/32 -> 1/30, 1/64 -> 1/60, 1/128 -> 1/120, 1/512 -> 1/500, etc).

HTH...

--dennis

Hi, Dennis
It is really kind of interesting. I can see the old numbers when using Canon's proprietary software (DPP) and when I use PM I see the perhaps real but nonetheless weird numbers. See pictures attached.
Definitely I would rather see the old numbers or at least have the option to switch back.
Regards,

-Luiz Muzzi

[attachment deleted by admin]

[jwvaughn]

I'm with Luiz and prefer the old style so thanks!

Have you considered the possibility of adding a switch to allow users to choose?


Jerry

[mklass]

Ditto


Mick
Sent from my iPhone using Tapatalk

[dennis]

I am modifying PM to record both old and new Exif tags so they are preserved "going forward" (not going to discuss what this means).

So there are two ways to expose this.  One is a preference like "Show actual values for Aperture and Shutter Speed when available".  If you like the old values then you would leave this unchecked (by default).  The other option is to have separate variables for "setting" (old Exif) versus "actual" (new Exif).

There are 3 situations that will occur with a particular photo (other than having no Exif):

a) old Exif tags only (specifically 0x829a, 0x829d)
b) new Exif tags only (specifically 0x9201, 0x9202)
c) both old and new Exif tags

What is happening now with Canon models like 70D is that we have (c) and so the way the preference would work IF we had one is to simply choose old versus new Exif tags (unchecked we use (a), checked we use (b)).  But it gets more complicated for cases (a) and (b).

So IF we had a preference as above and it was unchecked, then PM would still show the actual (new Exif) value for case (b) unless I were to try to figure out how to round like the camera does.  This can get ugly real fast because the actual values could be the result of some weird combination of whether the value is dialed in as a setting, or is computed based on the other parameters having been set (and of course exposure compensation and differences between camera makes/models and how lenses report their actual aperture and 1/3 stops for this and 1/2 stops for that etc).

And IF we had the preference and it was checked, then (a) would still show the old "setting" (although I suppose the {f} variable could display something like "~5.6" for aperture).

If I were to create two new variables for actual shutter speed and actual aperture, then we have a similar conundrum but perhaps it would be more useful.  For example, for case (b) then the variable {f} for the old Exif tag for aperture could show "~5.6" (rather than being blank) assuming I can determine this, and the new variable for the new Exif tag would display the actual value like "5.7" rather than being blank (which is more accurate than 5.6 anyway).  For case (a) it would be "5.6" for the old {f} variable and "~5.6" for the new variable.

It sure is hard trying to make everyone happy!

Right now I am leaning for new variables.

Thoughts?

--dennis

[jr_ed]

I would definitely like to have the ability to still have values reported as they used to, meaning that my "old" (existing) variables still report the "old" number (report 1/100 instead of 1/99). I can see why the exact values could be useful, but I simply find it easier to read 1/100 than 1/99.

I'm flexible on the implementation, and could probably live with either the new variables or the preference setting. But I often find it useful to use the same variables across multiple cameras, so if we have one set of variables that refer only to the "old" values, and a second set for only the "new" values, then if I get cameras over time that have only one or the other, I could end up needing different variable sets.

So if you really want to make everyone happy, I could suggest:

1)  Use the existing (old) variables with a preference box something like "Show actual values for Aperture and Shutter Speed when available" as you suggested. These variables could then follow the preference if both are available, but would use the only available value if only one is available. I think this would keep those of us happy who are used to the older versions of PM.
2)  Add two new sets of variables: one that displays only the "old" values, and a second set that displays only new values. For anyone who uses these, they should then be totally unambiguous.

That may complicate your code more than you want, but I thought I'd at least throw that out as a suggestion. I'll personally be happy as long as I can still get the values older PM versions displayed, but have some option for getting the new ones.

[Luiz Muzzi]

I am modifying PM to record both old and new Exif tags so they are preserved "going forward" (not going to discuss what this means).

So there are two ways to expose this.  One is a preference like "Show actual values for Aperture and Shutter Speed when available".  If you like the old values then you would leave this unchecked (by default).  The other option is to have separate variables for "setting" (old Exif) versus "actual" (new Exif).

There are 3 situations that will occur with a particular photo (other than having no Exif):

a) old Exif tags only (specifically 0x829a, 0x829d)
b) new Exif tags only (specifically 0x9201, 0x9202)
c) both old and new Exif tags

What is happening now with Canon models like 70D is that we have (c) and so the way the preference would work IF we had one is to simply choose old versus new Exif tags (unchecked we use (a), checked we use (b)).  But it gets more complicated for cases (a) and (b).

So IF we had a preference as above and it was unchecked, then PM would still show the actual (new Exif) value for case (b) unless I were to try to figure out how to round like the camera does.  This can get ugly real fast because the actual values could be the result of some weird combination of whether the value is dialed in as a setting, or is computed based on the other parameters having been set (and of course exposure compensation and differences between camera makes/models and how lenses report their actual aperture and 1/3 stops for this and 1/2 stops for that etc).

And IF we had the preference and it was checked, then (a) would still show the old "setting" (although I suppose the {f} variable could display something like "~5.6" for aperture).

If I were to create two new variables for actual shutter speed and actual aperture, then we have a similar conundrum but perhaps it would be more useful.  For example, for case (b) then the variable {f} for the old Exif tag for aperture could show "~5.6" (rather than being blank) assuming I can determine this, and the new variable for the new Exif tag would display the actual value like "5.7" rather than being blank (which is more accurate than 5.6 anyway).  For case (a) it would be "5.6" for the old {f} variable and "~5.6" for the new variable.

It sure is hard trying to make everyone happy!

Right now I am leaning for new variables.

Thoughts?

--dennis

Hi, Dennis
I would rather have a preference like "Show actual values for Aperture and Shutter Speed when available".
As I prefer the old values I  would leave it unchecked. Now, if Canon files begin to show only the new exif values, well, let it be so!
TIA

-Luiz Muzzi

[jwvaughn]

I think showing both would be my preference.  That would mean either could display N/A if the data were not available. No calculation required of you.  Additionally I would suggest labeling the "new" values as APEX xxx where the x's are the variable being displayed.

I am used to the old system but it looks like I may be dragged kicking and screaming into using APEX (-;  Having both displayed would help me get used to seeing it and help make the transition easier -also my reasoning behind the labeling.

Jerry

[Luiz Muzzi]

I think showing both would be my preference.  That would mean either could display N/A if the data were not available. No calculation required of you.  Additionally I would suggest labeling the "new" values as APEX xxx where the x's are the variable being displayed.

I am used to the old system but it looks like I may be dragged kicking and screaming into using APEX (-;  Having both displayed would help me get used to seeing it and help make the transition easier -also my reasoning behind the labeling.

Jerry

Hi,
That would also be fine for me.
As to the use of APEX values, I would like to ask if there is a real tendency in the industry to use them. I had never seen those values before.
TIA

-Luiz Muzzi

[dennis]

Okay just to throw a monkey wrench into this...

The newer Exif tags are APEX values and stored as such.  So if the aperture is f/5.6 then the corresponding APEX value that is stored is 5.0.  If the shutter speed is 1/60 (or really 1/64) then the corresponding APEX value is 6.0.

So, another option is to create new variables (I would choose {av} and {tv} mirroring the APEX definition) which show the actual APEX values, not translated to what a camera setting would be.

This is what some applications like Max OSX Preview displays.  This certainly distinguishes a "setting" versus the "actual APEX" value (well except for aperture of f/4 and APEX of 4 which are the same  ).

If these new Exif values aren't present in a photo (case (a) above) then the {av} and {tv} variables would be empty (or I could convert old settings to APEX).  If the old Exif values aren't present (very rare case (b) above), then I would convert the APEX values from the new Exif fields to the equivalent "setting" value for the old variables.  And in this case you will get strange numbers like 1/99 for {ss}.

How does that sound?

And another thing to chew on: if you have a lens that has an aperture size larger than f/1.0 (there are some out there!) then apparently the APEX/Exif spec wouldn't be able to store the APEX value because it would be < 0.0 and the Exif tag for aperture value is supposed to be an "unsigned" value (i.e. zero or positive).  Guess we will worry about that when Canon and Nikon start making these lenses...

One more thing: if the Exif tag for APEX Brightness Value (0x9203) is present, then the {brightness} variable will show this directly rather than trying to compute it based on aperture, shutter, and ISO.  This variable has always been displayed as an APEX value.  I haven't seen this tag on Canon cameras but some Sony cameras (e.g. Alpha 7R) do indeed write this tag.

--dennis

[jwvaughn]

Dennis,

As long as the meaning of the values is clearly distinguishable my sanity should remain intact   If the industry is moving to APEX, we had better start getting comfortable with it

I keep arguing with myself when it comes to calculating “missing” values. It would appear that Photoshop may already be doing it so there is precedence for it. My personal bias would be to not do it and it really is a bias as I cannot articulate a good argument against calculating them.

As for the large aperture issue, that sounds like a problem in the standard that at some point may have to be fixed.

On the brightness issue - is it really a problem? If I understand it, a measure value may be displayed instead of a calculated one which is a good thing to me.

Jerry