f# and DX/FX questions

1. Is an ƒ 2.8 DX lens as bright as an ƒ 2.8 full frame lens, i.e. is there any loss of low light performance beyond that expected by the smaller sensor size?

2. Will a lens like the AF-S DX 35mm f/1.8G have deeper depth of field at the same aperture on a DX camera than a 35mm f/1.8 made for full format sensors when used on a full frame camera?

3. I'm trying to understand if the aperture number on a lens like the AF-S DX 35mm f/1.8G means exactly what it says, or weather it needs correcting to a 35mm equivalent like the focal length (which is actually 35*1.5)?

Thanks!

skrulm8 said:

1. Is an ƒ 2.8 DX lens as bright as an ƒ 2.8 full frame lens, i.e. is there any loss of low light performance beyond that expected by the smaller sensor size?

Click to expand...

Light transmission will be identical. Obviously a bit less total light will be hitting the smaller sensor, but the f-stop is the same. A DX camera has the same field of view as is you cropped out the center portion of an FX image.

skrulm8 said:

2. Will a lens like the AF-S DX 35mm f/1.8G have deeper depth of field at the same aperture on a DX camera than a 35mm f/1.8 made for full format sensors when used on a full frame camera?

Click to expand...

Well, no. There is a relationship between sensor size and DOF. For macro shooters the "free" DOF on DX is good. For portraits, not so good. But it is not the same. Google DOF sensor size for more info.

You can start here: http://www.cambridgeincolour.com/tutorials/digital-camera-sensor-size.htm

skrulm8 said:

3. I'm trying to understand if the aperture number on a lens like the AF-S DX 35mm f/1.8G means exactly what it says, or weather it needs correcting to a 35mm equivalent like the focal length (which is actually 35*1.5)?

Click to expand...

It means exactly what it says. But the effects of the crop impact both effective focal length and aperture for DOF.

So a 100mm f2.8 lens on DX would have the same field of view and DOF as a 152mm f4.3 lens on FX.
--

See my plan (in my profile) for what I shoot with. See my gallery for images I find amusing.

Thanks, I found your answer very helpful!

skrulm8 said:

2. Will a lens like the AF-S DX 35mm f/1.8G have deeper depth of field at the same aperture on a DX camera than a 35mm f/1.8 made for full format sensors when used on a full frame camera?

Click to expand...

This bit depends.
DX and FX has different dof.

If you maintain the same angle of view, which means going closer with DX, for the same viewfinder crop DX has just over 1 stop more dof.

If you keep a constant focus distance the greater magnification effect with DX gives just under 1 stop dof.

On the other hand using a longer lens on FX to get the DX angle of view from the same focus distance reduces dof by nearly 2 stops.

--
Leonard Shepherd

Good photography is mainly about doing simple things well. The challenge is doing simple things well enough for good results.

skrulm8 said:

1. Is an ƒ 2.8 DX lens as bright as an ƒ 2.8 full frame lens, i.e. is there any loss of low light performance beyond that expected by the smaller sensor size?

Click to expand...

The light intensity (ie, the light per square millimeter) coming out of it is the same (assuming everything else being equal, eg, vignetting). But since the DX sensor is only about half the size of a FX sensor, the total number of photons hitting the sensor is naturally only about half (which matters if you think in terms of signal to noise ratio).

skrulm8 said:

2. Will a lens like the AF-S DX 35mm f/1.8G have deeper depth of field at the same aperture on a DX camera than a 35mm f/1.8 made for full format sensors when used on a full frame camera?

Click to expand...

With this you are talking about two different FOV. This is a bit like asking whether a 50 mm lens has more DOF than a 35 mm lens. You could compare things at the same subject distance or that the same framing (thus different subject distance for both lenses). But even with the same framing, the FOV is different for the (blurred) background, making a comparison difficult.

skrulm8 said:

3. I'm trying to understand if the aperture number on a lens like the AF-S DX 35mm f/1.8G means exactly what it says, or weather it needs correcting to a 35mm equivalent like the focal length (which is actually 35*1.5)?

Click to expand...

That depends on what you want to use the f-number for. Like the focal length, the f-number is a physical property of the lens, it stays the same regardless of the sensor size.

But once you start using 'equivalent' values, you must decide what you mean with equivalence. One US dollar is worth exactly $1 but it might currently be worth €0.70 via the currency exchange rates but in purchasing power it might be worth €0.90 (ie, you can buy the same with $1 in the U.S. than you can with €0.90 in Europe). So, what is the equivalent value, depends on your definition of equivalence.

If you want to compare DX to full frame, here's my take:

equivalent FX focal length = DX focal length x crop factor
equivalent FX Fstop = DX Fstop x crop factor
Equivalent FX sensor speed = DX speed x crop factor

For example, the equivalent of a 50/F1.4 full frame lens would be roughly 35/F1.0 (actually 33.3/F0.93).

So, now we get the following:

1. The 35/F1.0 creates an identical image on DX to the 50/1.4 full frame image (same FOV, DOF at equivalent enlargement)

2. The total light energy falling on the DX frame is the same as FX (change in Fstop compensates for change in sensor size, should have identical photon noise characteristics)

3. The DX sensor will need to operate at a lower ISO for the same shutter speed in FX (as it's now getting more energy per area due to the smaller sensor and larger aperture).

It's a very interesting trade at maximum aperture. The FX digital sensor I suspect is a lot more expensive than the DX, but a 35/F1.0 has to be harder to design than a 50/F1.4, especially since it would need to have 1.5*the resolution of the FF lens.

I'd be curious to find out how much an entry level camera (say a D3100) would cost if it had a FX sensor. Although it would be a marketing nightmare, it would be a great performer.

Steve

P.S. In the way of comparison, your FF photo at 50mm/F8 and 1/200 sec with ISO 50 film should look identical to a DX picture at 35mm/F5.6 and 1/200 sec with ISO 25 film.

sspoldi said:

So, now we get the following:

1. The 35/F1.0 creates an identical image on DX to the 50/1.4 full frame image (same FOV, DOF at equivalent enlargement)

Click to expand...

Shouldn't it be:
FX 50mm f1.4 = DX 35mm f2.1?

skrulm8 said:

1. Is an ƒ 2.8 DX lens as bright as an ƒ 2.8 full frame lens, i.e. is there any loss of low light performance beyond that expected by the smaller sensor size?

Click to expand...

f Stop is actually the ratio between your focal length and the size of the diaphragm that lets light enter your camera. No matter DX or FX, f-Stop numbers are the same. So for the same 35mm length and same f stop, the size of the hole letting light in is the same. Guess that answers your question.

skrulm8 said:

2. Will a lens like the AF-S DX 35mm f/1.8G have deeper depth of field at the same aperture on a DX camera than a 35mm f/1.8 made for full format sensors when used on a full frame camera?

Click to expand...

No, that is, if you compose the image the same. Lets say you're shooting at 85mm f/2.8. For a portrait with your subject at 10 feet, with DX, the DoF would be about 0.45'. Now, the same lens on FX would be 0.7'! However, to get the same composition on FX, you'd need to be at 6.66', not 10'. In this case, on FX, the DoF becomes 0.31'. So yes, FX sensors allow you to get a thinner DoF for the same image composition by about a factor of... gasp... 1.5!

skrulm8 said:

3. I'm trying to understand if the aperture number on a lens like the AF-S DX 35mm f/1.8G means exactly what it says, or weather it needs correcting to a 35mm equivalent like the focal length (which is actually 35*1.5)?

Click to expand...

Well, if you consider f-stop as a measure of DoF only then yes, one could correct it. But that is not true for what matters most out of f-stop, changing the amount of light hitting the sensor. Now, for the same ISO number (at that is the trick), using the same aperture, same shutter speed and same composition, you would get the same exposition from a D3S or a D90. The trick is that manufacturers set the ISO numbers accordingly in the camera (which really isn't a perfect science, DXOmark for one shows real measured iso numbers) so that these things match. This is why it is calle ISO for International Standard. It really is an adjusted value by the manufacturer to obtain a certain sensitivity standard.

However, the pixels in full frame sensors being much bigger, the gain applied to the CMOS electronics is lower (If you know electronics, this is like an OP-AMP gain). What is then improved is the noise performance for any given ISO because less noise is amplified in the process.

Well that's just my understanding...

skrulm8 said:

Thanks!

Click to expand...

Theov said:

sspoldi said:

So, now we get the following:

1. The 35/F1.0 creates an identical image on DX to the 50/1.4 full frame image (same FOV, DOF at equivalent enlargement)

Click to expand...

Shouldn't it be:
FX 50mm f1.4 = DX 35mm f2.1?

Click to expand...

I wish it was. The physics works out that DOF is a function of absolute aperture. So, 50/1.4 is approximately 35/1.0, or about 35 mm aperture diameter.

Interestingly enough, the blur spot due to diffraction is proportional to relative aperture (Fstop), but the smaller sensor needs a smaller blur spot so for diffraction effects you need a larger aperture in DX. No matter how you slice it, it's harder to design a DX lens than an equivalent FX lens. Let's take a 4x5 view camera and compare it to FF. At a crop factor of 3.6 (I made this up based on equivalent image area), a FF picture at 50/1.4 is equivalent to a 4x5 picture at 180/5.0. Nikon makes a 105/5.6 large format lens with shutter for less than $400, I'd bet it would make a way better image wide open in 4x5 than a 50/1.4 on FF 35mm. But I am speculating, I have no experience with large format so If I'm in error I do apologize.

The slower DX lens may be functionally equivalent if you're compareing FX flim to DX digital, but apples to apples the DX lens should be harder to design.

If anyone can comment from real experience please do, I'm just a dope with a calculator...

Cheers,
Steve

If you scroll to Leonard's reply, he gives you a cut to the chase answer. He's right on, not suggesting that others are wrong, just that his correct answer is easier to understand.
glo

sspoldi said:

Theov said:

sspoldi said:

So, now we get the following:

1. The 35/F1.0 creates an identical image on DX to the 50/1.4 full frame image (same FOV, DOF at equivalent enlargement)

Click to expand...

Shouldn't it be:
FX 50mm f1.4 = DX 35mm f2.1?

Click to expand...

I wish it was. The physics works out that DOF is a function of absolute aperture. So, 50/1.4 is approximately 35/1.0, or about 35 mm aperture diameter.

Click to expand...

DoF is also a function of focus distance and focal length which is a big reason why FX cameras have a DoF advantage. For the same composition, DoF will be thinner on a Full frame camera because you'll have to be closer to shoot it. In fact, for the same focus distance, DX cameras have thinner DoF than a FX camera but then you end up with a very different picture altogether. For any given camera the DoF increases to the square of the focusing distance. DoF will be four times larger if your focus distance is twice farther.

The Brenizer method is actually a good example of that, where one shoots at closer focus distance with a longer focal length to have a very thin DoF, but recomposes the multiple images in photoshop for that wide but very shallow DoF obtained in large format cameras. (Which I believe was also your point...)

Hehe, cool Sunday morning conversation!

Sylvain Larive said:

sspoldi said:

Theov said:

sspoldi said:

So, now we get the following:

1. The 35/F1.0 creates an identical image on DX to the 50/1.4 full frame image (same FOV, DOF at equivalent enlargement)

Click to expand...

Shouldn't it be:
FX 50mm f1.4 = DX 35mm f2.1?

Click to expand...

I wish it was. The physics works out that DOF is a function of absolute aperture. So, 50/1.4 is approximately 35/1.0, or about 35 mm aperture diameter.

Click to expand...

DoF is also a function of focus distance and focal length which is a big reason why FX cameras have a DoF advantage. For the same composition, DoF will be thinner on a Full frame camera because you'll have to be closer to shoot it. In fact, for the same focus distance, DX cameras have thinner DoF than a FX camera but then you end up with a very different picture altogether. For any given camera the DoF increases to the square of the focusing distance. DoF will be four times larger if your focus distance is twice farther.

The Brenizer method is actually a good example of that, where one shoots at closer focus distance with a longer focal length to have a very thin DoF, but recomposes the multiple images in photoshop for that wide but very shallow DoF obtained in large format cameras. (Which I believe was also your point...)

Hehe, cool Sunday morning conversation!

Click to expand...

It is isn't it! I had a hard time comparing FX and DX, since the lenses are really a bit different. I settled on the comparison in my first post, since it has identical apparent depth of field, field of view, and diffraction performance between FX and DX. If you don't compare like that it rapidly gets really hard to get a meaningful comparison.

There's a lot of rules of thumb (DOF proportional to aperture, inversely proportional to focal length squared, etc) but they all change when you change format sizes, and you really have to think about it (My brain starts to melt at that point).

Great discussion,
--
Steve

sspoldi said:

The physics works out that DOF is a function of absolute aperture. So, 50/1.4 is approximately 35/1.0, or about 35 mm aperture diameter.

Click to expand...

Part no

Changing from 50mm to 35mm and maintaining the same crop (which means going closer with the 35mm) is worth 2 f stops not 1.

Focal length (and magnification which includes focus distance) has a greater effect on depth of field than aperture.
Going from 25mm to 50mm is worth 4 stops dof, not 2.
--
Leonard Shepherd

Good photography is mainly about doing simple things well. The challenge is doing simple things well enough for good results.

Leonard Shepherd said:

sspoldi said:

The physics works out that DOF is a function of absolute aperture. So, 50/1.4 is approximately 35/1.0, or about 35 mm aperture diameter.

Click to expand...

Part no

Changing from 50mm to 35mm and maintaining the same crop (which means going closer with the 35mm) is worth 2 f stops not 1.

Click to expand...

Yes, but we also change the sensor size at the same moment and because you enlarge an image from a DX sensor more than one coming from a FX sensor, you need smaller circles of confusion which works in the opposite direction, thus the DOF of a 50 mm f/1.4 on a FX sensor really is equivalent to a 35 mm f/1 on a DX sensor (with a little bit of rounding).

sspoldi said:

equivalent FX focal length = DX focal length x crop factor
equivalent FX Fstop = DX Fstop x crop factor
Equivalent FX sensor speed = DX speed x crop factor

Click to expand...

Sorry, I made a mistake...
Equivalent FX sensor speed = DX speed x (crop factor)^2, my apologies.

I took 2 pictures by way of an illustration for the above relationships (well, the first 2 anyway).. the first picture is taken at roughly 70mm and F10, the second picture is taken at 35 mm and F5, cropped with a crop factor of about 2 (to simulate a smaller frame camera).

You can see that the DOF and FOV is quite close, matching the relationships above. I did not match the correct ISO, but if I did I would assume that the noise would also follow suit, that's an experiment for tomorrow.









Sorry for the poor quality photos, just wanted to illustrate my point. Note that since the position I took the photos didn't change, perspective is identical.
Cheers,
--
Steve

I don't think it makes alot of sense to compare a 35mm lens on DX to a 35mm lens on FX. It will give you a different FOV, or if you move forward to get the same framing you will have a different perspective.

However, if you use a lens that would give you pretty much the same framing and perspective, i.e. a 50mm f1.8, the FX option will give you about 1 stop narrower DoF....

FX camera, 50mm lens @ f1.8, distance 15 feet, ..... DoF is 2.93 feet
DX camera, 35mm lens @ f1.8 distance 15 feet,..... DoF is 4.03 feet

--
http://www.andrewsandersphotography.co.uk