Take pictures at real-world size without a 1:1 lens

[TCampbell]

You may want to clarify the instructions with your instructor. Normally, when we discuss "1:1" image scale in photography, we're talking about the size of the object on the CAMERA SENSOR and not the object size on the computer screen. That's a huge difference.

Let's just assume your instructions were correct and your instructor did mean the computer screen (which would be odd, but we'll go with this idea for now.) The resolution accepted for monitors is generally 72 dpi (dots per inch). HOWEVER... most modern monitors have surpassed that by quite a bit. My own monitor is doing just slightly over 100 dpi. You'd need to use an editing program (e.g. photoshop) to resize the photo based on this. For example... suppose you snapped a product photo of your mobile phone. Let's suppose the mobile phone is 5" tall. Based on a 100 dpi monitor resolution (or whatever dpi you're working with) you'd resize the image so that the area phone itself occupies 500 lines from top to bottom (not the full image size... just the phone. So if the full image had an extra 1" margin above and below the phone you'd need to account for that.)

HOWEVER... I suspect you mis-heard your instructor and when he/she asks you to get an image at "1:1" scale, he/she is really talking about macro photography. In macro photography, the image size on the CAMERA SENSOR is the same size as the object in real life.

Let's use a €0.01 coin as an example:

The coin has a diameter which measures 16.3mm. Your Nikon D7000 has an APS-C size sensor which measures 23.6mm x 15.6mm.

That means that if you were to take a photo of this coin at 1:1 scale, the count would fit horizontally (with room to spare) but it wouldn't actually fit in the image vertically... the coin is .7mm larger than the height of the sensor -- so you'd slightly clip off the top & bottom of the coin.

After taking such a photo, when you display it on the computer screen, the coin will be huge!

There are numerous ways to take close-up photos. The highest quality method is to use a true 1:1 scale macro lens (be careful, because lots of lenses advertised as "macro" can't really do 1:1 scale... generally if it's a "zoom" lens that also claims to be a "macro" lens then it's not really a 1:1 scale lens. But usually the prime (non-zoom) lenses that claim to be "macro" are 1:1 scale (I do know of a few exceptions so you still have to be careful.)

You can buy close-up diopters. These are generally single-element lenses that thread onto the end of the camera lens. While they do let you get closer, the single-element design has drawbacks. If you've ever used a simple magnifying glass to look at the text on the pages of a book, you might notice that the center is magnified, but the edges show color fringing around the edges of the letters (even though the print is supposed to be black print on a white page.) The color "fringing" is caused by the dispersion (chromatic aberration) of the edges of the lens behaving like a prism and splitting light into it's constituent wavelengths. Close-up diopters are generally very inexpensive.

You can use "extension tubes". These are simple hollow barrels. You attach the extension tube to the camera body, then attach the lens to the other end of the extension tube. By moving the entire lens farther from the camera you are effectively reducing the closest focusing distance allowed by the lens and creating a larger image (actually they reduce the entire focusing range -- the lens will generally not be able to focus out to "infinity" when attached to an extension tube). There is no glass or lens element of any kind in an extension tube -- they are completely hollow (nothing but air inside). Since there are no lenses, there's no degradation of optical quality (well... that's not entirely true... lenses are technically optimized for a specific back-focus distance.) Let's just say the degradation would be minimal -- especially when compared to "close-up" diopters. Extension tubes also tend to be very inexpensive because of their simplicity (they usually do have electronic contacts so that they do pass the camera's communication pins through to the lens.)

You can also use a "reversing ring". This simple little gadget is a ring that has the camera's bayonet style mount on one side of the ring, and camera lens "filter" threads on the other side. So screw it on to the FRONT side of your lens (as though you were attaching a filter.) You now have a bayonet mount on the FRONT of your lens. Turn the lens around and connect it to your camera body. The downside is that since the lens is mounted backwards, you have no control over it. You manually focus it. You can't control the f-stop.

Back up at the top, I mentioned using a true "1:1" macro lens. With that option there are no compromises. You have full control of everything. The auto-focus works. The f-stop works. You don't have optical compromises. You get to use the full focal range of the lens. BUT... while this is certainly the most fully-functional / least compromises path, it's also the most expensive.

If you were really doing macro work seriously, then get the true macro lens. For a class project, use one of the other methods (especially since students are supposed to be poor & starving and not enough money to buy gear.)

 

[Bedo]

Thanks for all the replies.

First of all I apologize because I didn't write the question very well.

I actually don't need to take photos in 1:1 scale (in the sense that my image shouldn't be at real size on the CMOS sensor).



What I really need is to use the well-know relation:


1 pixel = 0.264583333 millimeters


So if I place a 100mm ruler on the photo, it must be saved with ( 100 / 0.264583333 ) pixels on length.


How can I obtain this result?


By the way I need to take pictures of human-ear in order to perform a psychoacoustic experiment.

 

[unpopular]

If the exercise is not about pixel relationship to physical space, then I'd just use a scale.

I had a project that required 1:1 reproduction. I quickly learned that using a physical scale is really the only way to go. I had all sorts of complicated means of doing this, but in the end it just wasn't reliable. There are so many factors involved with magnification, it's much easier just to photograph a ruler or some other known size in the scene and use it as a reference.

This is why when you see scientific and forensic photographs there is often a ruler or square in the frame.

 

[Helen B]

Assuming that you mean what you say - ie 1:1 on your computer screen - it is quite easy to set up to do repeatedly and quickly. First display any photo you have taken with the D7100 on your screen in the exact way you intend to show the future images. Then try to take a focused picture of that picture, filling the whole frame, or if the aspect ratio doesn't match, one side of the frame. Try all your lenses. It doesn't matter if you can't get an exact match - just select the picture that gets closest. Note the distance from lens to screen, and make yourself a little rig to hold the camera that distance from the object. It can be a piece of string, a short chain, a simple wire frame - use your ingenuity.

Now, whatever standard process you use to resize the newly taken image so that it fits the original screen image (ie the new screen image is identical to the original screen you took the picture of) can be applied to the pictures you take using your rig.

This is a minor adaptation of old, proven techniques used to churn out images at an exact, repeatable magnification without having to do individual measurements or calculations. I've been brief, so ask if anything is unclear. The principles are really simple.

Edit: You wrote your clarification while I was typing that. The above will work for you, with a slight modification to account for any difference between the screen you use for the original test and your 'standard' pixel pitch.

You might want to get a simple diopter. It sounds like you won't need much more, if anything.

 

[amolitor]

Yes Mully. But what is 'it'?

 

[480sparky]

Yes Mully. But what is 'it'?

...............

........By the way I need to take pictures of human-ear.............

 

[The_Traveler]

First of all I apologize because I didn't write the question very well.
I actually don't need to take photos in 1:1 scale (in the sense that my image shouldn't be at real size on the CMOS sensor).
What I really need is to use the well-know relation:

1 pixel = 0.264583333 millimeters


So if I place a 100mm ruler on the photo, it must be saved with ( 100 / 0.264583333 ) pixels on length.
How can I obtain this result?

By the way I need to take pictures of human-ear in order to perform a psychoacoustic experiment.

This is all proving quite illuminating to me.
I can't, however, concede that a single pixel is .264 etc mm.
If that were so than a D800 sensor would be several times larger than a D700 sensor but we know that the physical dimensions are very close to the same.
My D700 sensor is 36 mm and 4000 pixels long, thus a single pixel is .008456.

If you save something so that the dimensions in pixels is the same size as the object in millimeters that would mean that a 100 mm ruler (approximately 4 inches) would
look like this
but the actual display in size on your screen would be dependent on the screen size and resolution.
Thus a human ear (about 70 mm) would be
but also dependent on screen size and resolution.

Instead of trying to put this in terms of resolution, why not just tell us what you are trying to accomplish and that will give us a better hint of how to respond.

 

[Bedo]

We are using this conversion value: Meter Pixel Conversion -- EndMemo

Instead of trying to put this in terms of resolution, why not just tell us what you are trying to accomplish and that will give us a better hint of how to respond.

We need to calculate the time that the sound needs to go from the external pinna to the ear focus point. We calculate it as distance / sound_speed.

So we need to calculate the exact ear size (width, height, ... ) starting from a photo.

We could place a ruler but we need to take a lot of photos so I would like to know if there is a way to calculate the correct focal lenght and subject distance with a particular lens to obtain a conversion value of 1 pixel = 0.264583333 millimeters


By the way I'm also curious to understand how the relation 1 pixel = 0.264583333 millimeters has been calculated. I searched for it on Google but I didn't find an answer...

 

[unpopular]

We are using this conversion value: Meter Pixel Conversion -- EndMemo

I am not sure what to make of this calculator, but I am pretty sure it's not doing what you think it is.

 

[Bedo]

Yes, I would really like to know how the conversion 1 pixel = 0.264583333 millimeters has been calculated.

I'm going to ask to [FONT=Helvetica, Tahoma, Arial, sans-serif]my postdoctoral student and let you know![/FONT]

 

[Helen B]

The standard method I already suggested will work for you. Now that you have explained more (why didn't you do that to begin with?) it is even simpler. You make a wire frame that holds the camera (with pre-focused lens) a fixed distance from the ear. First you use it to take a picture of a ruler. This tells you your real pixel dimension to use for later image pixel-distance conversion (ie your pixel to mm conversion). (Forget that equivalence you found somewhere - it isn't appropriate for you and you are simply confusing yourself with it.) Then you take pictures of the ears, framing them in the wires. Use the longest lens you have to limit errors caused by distance. This is all tried and tested practice.

 

[amolitor]

The pixel size of 0.264mm is, I think, a pretty standard dot pitch for monitors at some standard resolutions. I assume from this that the OP is interested in rendering things in "real life" size on a monitor.

 

[Helen B]

The pixel size of 0.264mm is, I think, a pretty standard dot pitch for monitors at some standard resolutions. I assume from this that the OP is interested in rendering things in "real life" size on a monitor.

That's what I thought the first post sounded like (1:1 on the monitor). Now it sounds like all that is required is a simple way of quickly converting from image pixels to distance. The 0.26 mm thing is a red herring, and the sooner the OP realizes that the better. The quoted 9 sf precision is kinda worrying, however. Also strange is the absence of a response to the various methods already suggested.

 

[amolitor]

The pixel size of 0.264mm is, I think, a pretty standard dot pitch for monitors at some standard resolutions. I assume from this that the OP is interested in rendering things in "real life" size on a monitor.

That's what I thought the first post sounded like (1:1 on the monitor). Now it sounds like all that is required is a simple way of quickly converting from image pixels to distance. The 0.26 mm thing is a red herring, and the sooner the OP realizes that the better. The quoted 9 sf precision is kinda worrying, however. Also strange is the absence of a response to the various methods already suggested.

Yes, to all of this

 

[The_Traveler]

The standard method I already suggested will work for you. Now that you have explained more (why didn't you do that to begin with?) it is even simpler. You make a wire frame that holds the camera (with pre-focused lens) a fixed distance from the ear. First you use it to take a picture of a ruler. This tells you your real pixel dimension to use for later image pixel-distance conversion (ie your pixel to mm conversion). (Forget that equivalence you found somewhere - it isn't appropriate for you and you are simply confusing yourself with it.) Then you take pictures of the ears, framing them in the wires. Use the longest lens you have to limit errors caused by distance. This is all tried and tested practice.

What Helen said is totally right on the money.
I have done this exact thing (with teeth rather than ears)

Fix the camera to object distance (preferably as close as possible to filling the frame but at some distance to minimize placement error but maximize size, so using a 105 macro or a macro adaptor on available lens )
Take a picture of a high accuracy rule that is on the same plane as the subject.
ABFO No. 2 Photomacrographic Scale - Arrowhead Forensics or Motion Control Systems and Components
Working at known size of viewed image, measure with the PS built-in ruler (record x and y values) and record

(Ideally you could place the ruler over the ear and take a image of the rule with each subject. That would maximize accuracy and minimize effects of subject distance error, etc. but require extra measurements. This could be an advantage by writing the subject number on the rule and thus identifying each image.)

Record x and y distance so that rotation and displacement problems are minimized. Then calculate hypotenuse to get real distance and normalize using the original measurements of the ruler.

All of the calculations can be done just by entering the x and y values in a spreadsheet and using formulae to calculate it all easily.