[baro-nite]

Here are two shots, shown at various crops. Compare and contrast. (If you have a wide enough monitor, make the window wide and the images will show side-by-side.)


25% crops


50% crops


100% crops

The birds aren't in exactly the same spot, but at similar distances and in similar light. Both shots use the DA*300 at f/5.6. One is on the Q, ISO 125, 1/200s. The other is on the 1.4x-S and 2x-S Pentax Rear Converters on the K-5, ISO 1000, 1/250s. Apologies for the different shutter speeds, despite which, unfortunately, I had to push the Q shot +2.25ev in RAW development, which added noise. I don't understand the difference in exposure given how similar the light was.

Curve adjustment, sharpening, and NR were done subjectively.

[crewl1]

They both look very good. I am going to guess the one that is looking right is the Q image?
Going by the depth of field differences.

[drougge]

The birds aren't in exactly the same spot, but at similar distances and in similar light. Both shots use the DA*300 at f/5.6. One is on the Q, ISO 125, 1/200s. The other is on the 1.4x-S and 2x-S Pentax Rear Converters on the K-5, ISO 1000, 1/250s. Apologies for the different shutter speeds, despite which, unfortunately, I had to push the Q shot +2.25ev in RAW development, which added noise. I don't understand the difference in exposure given how similar the light was.

Well, 2 stops light difference isn't something you notice much. Then there's the question of aperture accuracy. On both cameras. It's probably good to within 1/3 stop on the K5 before you add the TCs, but then there's some extra mechanical slop. And on the Q the adapters don't even try to give rotational invariance (I'd love to use a clearer term, sorry) on the aperture lever movement.

Let me try to clarify: The KA (and later) mounts specify how far to move the aperture level per stop, and at least the Pentax adapter should be following that part of the specification (and the third party ones probably not). But that movement is indexed from the stop position of the lever, not absolute relative to the mount. This is because the movement is small, and exactly where the lens stops when mounted isn't all that exact.

[baro-nite]

Going by the depth of field differences.

I wouldn't go by DOF, which should be at least roughly similar anyway (effective f/16 on the K-5 vs. f/5.6 on the Q), because the subject angles are at least somewhat different and the focus point may well have been different. I'll ID which is which later, to give more people a chance to make an unbiased comparison.

Well, 2 stops light difference isn't something you notice much.

More like 2.5 stops when you factor in the shutter speed difference, but yes, I take your point. However both shots have the same kind of bird with at least parts of the bird in full sun, from about the same angle, so I would not expect 2+ stops of difference in the light itself. As to the adapter, I once did a quick test to find the actual one-stop position on the ring. Since the ring doesn't click I can't be positive I didn't accidentally move it during the session, but in general that seems not to happen. I'll be doing more testing later so will be sure to confirm the aperture setting. Good point about possible mechanical slop -- I should also test the TCs on the K-5; quite possibly there's enough lag that the lens didn't stop down fully.

Frankly I was surprised at how close the performance was. Obviously this is not a controlled test and is not conclusive, but I wanted to try something practical, including a comparison of handling, about which I'll have more comments later.

[TenZ.NL]

Is it correct that you used 2 converters, a 1.4 and a 2.0, stacked on each other? IIRC a 1.4 TC looses 1 stop of light and a 2x TC looses 2 stops if used alone. I don`t know if stacking adds additional loss of light but a simple calculation with these figures makes 3 stops x iso 125 = iso1000. Or do I see this wrong?

[baro-nite]

Is it correct that you used 2 converters, a 1.4 and a 2.0, stacked on each other? IIRC a 1.4 TC looses 1 stop of light and a 2x TC looses 2 stops if used alone. I don`t know if stacking adds additional loss of light but a simple calculation with these figures makes 3 stops x iso 125 = iso1000. Or do I see this wrong?

Yes, 1.4x and 2x stacked, so you're absolutely correct it is 3 stops of light, which is why I used ISO 1000 -- I wanted to make things as equivalent as possible.

[crewl1]

A few Q's:

Why did you title the thread heresy?

Are the initial photos both uncropped as taken from the camera?

What is the resulting focal length from the TC stack on the DA300?

How long a plate did you need to balance the DA300 with TCs on your tripod head?

What is the best way to hard boil an egg so I don't crack the shell?

[baro-nite]

A few Q's:

A few A's:

I feel like a bit of a heretiQ for comparing our beloved teleQonverter to normal TCs. It's important to me to know the advantages and limitations of the tools we have available, so it's a comparison I feel I need to explore. Going into this first test I had a hunch the TCs would win. Then on first looking through the sets of shots my initial conclusion was that the Q won hands down. Then on looking closer I got to where I am now -- the two could be very close indeed. I will do more testing, but if this conclusion holds up then it will come down to handling differences and convenience.

The initial samples are 25% crops, 2800x2100 pixels resized to 700x525.

Effective FL of the DA*300 on the TCs is 840mm (actual value, not taking "crop factor" into account). This rig on the K-5 has a wider angle of view than the 300 on the Q. However, the 2.8x boost from the TCs is very close to the Q's boost in pixel pitch compared to the K-5, which is about 3.1x. I'm more interested in this latter difference, because it gives a better idea of how large you can take the image in final output (print or screen).

I use an Arca/Swiss plate on my DA*300 tripod foot; it's a bit over 3" long. Fortunately it is just long enough that if I mount the foot backwards I have just enough room in the clamp to get the whole thing to balance. I've been meaning to get another rail, though, to complete my pano/gimbal rig, and then I won't have to reverse the foot.

Some people like to poke a tiny hole in one end of the egg. Pointy end, I think. I can't be bothered.

[crewl1]

A few A's:

I feel like a bit of a heretiQ for comparing our beloved teleQonverter to normal TCs. It's important to me to know the advantages and limitations of the tools we have available, so it's a comparison I feel I need to explore. Going into this first test I had a hunch the TCs would win. Then on first looking through the sets of shots my initial conclusion was that the Q won hands down. Then on looking closer I got to where I am now -- the two could be very close indeed. I will do more testing, but if this conclusion holds up then it will come down to handling differences and convenience.

The initial samples are 25% crops, 2800x2100 pixels resized to 700x525.

Effective FL of the DA*300 on the TCs is 840mm (actual value, not taking "crop factor" into account). This rig on the K-5 has a wider angle of view than the 300 on the Q. However, the 2.8x boost from the TCs is very close to the Q's boost in pixel pitch compared to the K-5, which is about 3.1x. I'm more interested in this latter difference, because it gives a better idea of how large you can take the image in final output (print or screen).

I use an Arca/Swiss plate on my DA*300 tripod foot; it's a bit over 3" long. Fortunately it is just long enough that if I mount the foot backwards I have just enough room in the clamp to get the whole thing to balance. I've been meaning to get another rail, though, to complete my pano/gimbal rig, and then I won't have to reverse the foot.

Some people like to poke a tiny hole in one end of the egg. Pointy end, I think. I can't be bothered.

Thanks for the A's. BRB,, time for breakfast

[vonBaloney]

I wanna see the one where you put the two TCs on the lens AND use the Q!

[baro-nite]

I wanna see the one where you put the two TCs on the lens AND use the Q!

Actually, I've done that. (Different TCs, though.)

[baro-nite]

Minor update on the exposure question, based on some quick tests with an 18% gray card.

The OEM adapter's aperture ring is quite accurate with this lens, with not much difference between the 0 setting and where the ring stops a bit past that, and then pretty close to actual half-stop changes with each half-stop mark on the ring. This is a little different from what I had found in an earlier test, but this should be a better test target.

ISO 125 on the Q is really more like ISO 100 compared to the K-5. This accords with the results from DxO. So that explains a quarter-stop of the difference.

The K-5, on the other hand, does not give such accurate results with this lens. At the settings I tested, 1/3-stop increments from f/4.0 to f/8.0, compared to f/4.0 the other settings were all roughly a quarter-stop overexposed from the expected value, with the TCs and without.

On the other other hand, I seem to be losing a bit more than 3 stops with the two TCs.

So, combined with the Q's ISO difference, that only makes up for about a 1/3-stop of the overall difference I observed. I'll have to put the rest down to operator error, either an incorrect setting on the adapter's aperture ring, or lighting conditions that were more different than I had realized.

[kshapero]

What is a TC?

[baro-nite]

What is a TC?

Teleconverter. I'm using both the Pentax Rear Converter-A 1.4x-S and 2x-S in these tests.

The Q has been proven to be a good way to extend the reach of an excellent lens such as the DA*300. My question is whether or not we can accomplish the same thing with TCs.

[drougge]

On the other other hand, I seem to be losing a bit more than 3 stops with the two TCs.

That's to be expected, you're adding 12 elements to the lens. (The 3 stop difference is true for the f-stop, not for the t-stop. Which is generally unknown on still camera lenses anyway, but the change matters here. Also assuming they are actually sqrt(2) and 2 long, which might not be quite true. But I expect all those surfaces to be a bigger light stealer than any extra length they might have.)