Shallow Review: Tamron 150-600 f/5-6.3 VC vs 300mm/2.8 II +2xTC III

[Pit123]

Have you the Tammy? Can you contribute for anything here at all?

You said it. You don't own it, and you've rejected the postings that are contrary to your opinions from owners of the lens.

In the case of the 1D X, you have 6.95µm pixels, or an area of 6.95^2µm: 48.3µm^2. The 7D has 4.3µm pixels, or an area of 4.3^2µm: 18.49µm^2. Again, because were working in two dimensions here, it's not a linear scale, you can fit 48.3/18.5 7D pixels into one 1D X pixel. That comes out to 2.61x 7D pixels per 1D X pixel. You would need a 2.6x TC in in order to completely normalize the crop difference between the 7D and 1D X, all else being equal.

Wouldn't that be a sqrt(2.6)x teleconverter? Every teleconverter I've seen lists their focal length multiplier, meaning a 2x teleconverter puts 4x the pixels on a target.

You are correct. That should actually read:

"That comes out to 2.61x 7D pixels per 1D X pixel. You would need a 1.6x TC in in order to completely normalize the crop difference between the 7D and 1D X, all else being equal."

Nice. I see you change your mind. You now admit to not have use a 2.6x tc to compensate, as you said before,, but a 1.6x.
If I have a 1dx and want the same pixels on the subject as the 7d, I have two choices. Put a 1,6x TC on the lens or upscale the image by 1.6x. When talk about upscaling you always talk about upscaling in both direction. Everything is upscaled 1.6x. Ok? Same with lenses. If I want a 400mm to act as a 600mm I put a 1.5x TC on it. Normally an upscaled image will not be as good as putting a tc on, or having a longer FL. But sometimes it does. And that not a claim for me that I can upscale the C400mm to give same IQ as the soft tammy @6.3. Its a question and a guess..Thats it!.
And now I'm out.
Thanks!

First off, my bad. It was a typo. Not a change of mind. But still, my bad.

Second, your still not understanding. Things are magnified 2.61x. The LENS FOCAL LENGTH, and ONLY the lens focal length, is scaled by 1.6! It's the same deal in the end. If you square 1.615 (the actual focal ratio), what do you get? 2.61! The 7D resolves 2.61x more detail than the 1D X. To compensate for that difference, you need a focal length 1.6x as long. A 646mm lens results in 2.61x more 1D X pixels falling on the same subject area as a 400mm lens on the 7D. Don't get too caught up on the simple and scalar ratio of focal lengths...that doesn't tell you enough about the actual differences involved.

The difference in focal length is linear (1.615x), the difference in magnification is squared (2.61x). The value that really matters, from a detail perspective, is the magnification factor. If you use a 400mm lens on both a 1D X and a 7D, the 7D will resolve 2.61x MORE DETAIL. It sounds like a lot. It really IS a lot!

If you still don't believe me, maybe you will believe a well-known professional in the field of bird photography:

Size Does Matter; The Power of the Square of the Focal Length

[quote author=Art Morris]"In the original “The Art of Bird Photography” I wrote something to this effect: the size of the bird in the frame is not a factor of the focal length but rather a factor of the square of the focal length. In other words, if you go from a 400mm lens to an 800 mm lens, the bird will be four times bigger in the frame (not twice as big). "

The focal ratio between 800mm and 400mm is 2x, but the subject is enlarged in the frame by 4x! Subject size in the frame is related to the SQUARE of the Focal Ratio, it is not a linear relationship. You can clearly see that in Art's animated image of the bird...one was taken at 700mm, the other at 1120mm. The square of that ratio, 1120/700^2 or 1.6^2 is, yup, you've got it, 2.6x! The animated bird jumped in size in the frame by 2.6x, not by 1.6x. You can even download the animated image and do some area measurements yourself if you want. Draw a box around the bird when it's smaller, draw a box around the bird when it's larger, and compare the areas.

WOW! If you really believe that double the focal length will enlarge the subject by 4x, then I have lost all respect for your skills! Sorry man!
It will enlarge the subject by 2x of course, nothing more, nothing less. Do you even own a camera? Or a lens?

Of course it does, the subject becomes twice as wide and twice as tall, that makes it cover four times the area it did before. You are forgetting you are going from a one dimensional measurement to a two dimensional representation of the subject. If it was a three dimensional model it would be doubled again for depth.

Here is a diagram, the image is a 3 x 2 format image, the blue square a subject, the green square is what happens if you either double the focal length or halve your subject distance, both edges are double the blue squares edges in length, the green square covers four times the area the blue square does.

Time to eat some humble pie
[/quote]
LOL
Lets say the subject is an arrow that goes from lower left corner to upper right corner on the blue area. How much bigger will the subject (arrow) be on the green area? Tell me!!! Obvious you cant. Exactly twice the length. So that gives me 2x magnification. Not?
What is wrong with you guys? Really!

 

[Pit123]

Do you even own a camera? Or a lens?

lol.. i guess you don´t.

Based on?

 

[jrista]

WOW! If you really believe that double the focal length will enlarge the subject by 4x, then I have lost all respect for your skills! Sorry man!
It will enlarge the subject by 2x of course, nothing more, nothing less. Do you even own a camera? Or a lens?

LOL

Well, to be quite frank, I'm not really concerned by the loss of your "respect." I guess I honestly care more about truth and facts than the respect of someone who refuses to acknowledge both when their hit in the face by them.

This is not complicated stuff. It is really basic geometry. YOU SHOULD KNOW THIS STUFF from your high school level math classes! Your still thinking one-dimensionally, as in "An 800mm lens is twice the length of a 400mm lens." Yes, that's true, an 800mm lens is just twice the length of a 400mm lens. But it has 1/4 the FoV!!! Project that 1/4 size FoV onto the same physical sensor area that a 400mm lens projects its 1/1 size FoV onto, and you've enlarged your subject four-fold.

Here is another pretty picture:

Hopefully these three diagrams will completely demonstrate the concept. The first shows the 35mm full frames that a 400mm (orange), 600mm (green) and 800mm (blue) lens produce. It also shows how those three frames appear when magnified to the same size (i.e. the sensor size). The 800mm frame has a 1:4 size ratio relative to the 400mm frame which, for the purposes of this debate, has a 1:1 size ratio. You can clearly see that the prairie dog is much more than twice as large in the frame in the magnified 800mm image...it is four times as large. To put it in different terms, the prairie dog is about 25% of the frame in the 400mm image, but it is 100% of the frame in the 800mm image. A factor of four difference.

The second diagram demonstrates the one dimensional linear relationship between lens focal lengths. The 400mm and 800mm lenses have a 1:2 ratio relationship, since a 400mm diagonal AoV is 6°11', where as an 800mm diagonal AoV is 3°05'.

The third diagram plainly lays out WHY an 800mm lens enlarges the subject four fold: You can fit four 800mm frames into a single 400mm frame!! Full-frame scaling is two dimensional, not one dimensional like angle of view. You have to account for both dimensions when determining the difference in total detail on your subject. An 800mm lens may only have 1/2 the AoV, but it resolves 4x as much detail.

If this doesn't clear the problem up for you, then nothing will. This is about as plain and obvious as I can possibly make it. This is very basic, high school level geometry. It should be very obvious.

 

[privatebydesign]

LOL
Lets say the subject is an arrow that goes from lower left corner to upper right corner on the blue area. How much bigger will the subject (arrow) be on the green area? Tell me!!! Obvious you cant. Exactly twice the length. So that gives me 2x magnification. Not?
What is wrong with you guys? Really!

Oh wow, you are not getting this at all. the arrow is two dimensional, if you double the length you double the width, if you do that then the arrow covers four times the area. See below, the yellow arrow is a scaled copy of the red arrow, it is twice as long, it covers four times the area. When you are measuring magnification of a two dimensional object you have to consider the area it occupies.

As an aside, pixel density has absolutely nothing to do with magnification, nothing, zero, nada. Initially magnification refers to how big the subject is on the sensor compared to its actual size in real life, subsequently it refers to the reproduction ratio, how big is it on screen or in print compared to how big it is in real life, neither of these have anything to do with pixel density.

 

[dcm]

This discussion seems to be about the term magnification and whether it is a 1 or 2 dimensional concept. Did anybody look up the definition of magnification? This might be helpful - http://en.wikipedia.org/wiki/Magnification

 

[jrista]

As an aside, pixel density has absolutely nothing to do with magnification, nothing, zero, nada. Initially magnification refers to how big the subject is on the sensor compared to its actual size in real life, subsequently it refers to the reproduction ratio, how big is it on screen or in print compared to how big it is in real life, neither of these have anything to do with pixel density.

That is one meaning of magnification in one context. There are multiple contexts within which the term magnification is valid. The magnification you've defined is optical magnification, but not the same kind of magnification we've been talking about so far. For your definition, "magnification" of the lens would be something like 0.15x, where an object a hundred feet away that may be five feet tall is reproduced at 12mm in size on the sensor. This is the kind of magnification that is sometimes published with lens specifications. That is indeed a valid definition of magnification, but it comes with a specific context.

The magnification that we've been talking about so far is relative magnification in terms of subject size to frame size. It doesn't really matter how big it is in real life in this context. In this context, magnification is referring to the change in relative subject size with a change in resolving power. The change in resolving power can either come from an increase or reduction in focal length, or an increase or reduction in pixel size. Either way, the subject is either magnified or reduced RELATIVE to some chosen reference point. Technically speaking, one need not necessarily involve a frame size...the frame could be infinite. The detail on the subject increases either way...the final images, when observed at 100% size, result in the subject imaged with an increased focal length or smaller pixels will be larger (this, magnified.)

Both definitions of magnification are valid. They just have different contexts within which they are valid.

This discussion seems to be about the term magnification and whether it is a 1 or 2 dimensional concept. Did anybody look up the definition of magnification? This might be helpful - http://en.wikipedia.org/wiki/Magnification

See my response above.

 

[redfield]

"Linear or transverse magnification — For real images, such as images projected on a screen, size means a linear dimension (measured, for example, in millimeters or inches)."

It's pretty clear that in this case magnification is linear ... 8)

 

[jrista]

"Linear or transverse magnification — For real images, such as images projected on a screen, size means a linear dimension (measured, for example, in millimeters or inches)."

It's pretty clear that in this case magnification is linear ... 8)

Perhaps people are assuming different contexts. Your talking about a linear change in angular size. I don't dispute the definition of that.

I'm talking about a change in area, "total size", because it is the change in subject size relative to the AREA of the sensor that results in an increase in detail. Not a 2x increase in detail, a 4x increase in detail. In my previous example, the 400mm lens resolves the prairie dog at about 25% of the frame. The 800mm lens resolves the prairie dog at 100% of the frame. That is a 2x change in angular size, but a 4x change in total subject size (width and height).

 

[privatebydesign]

This discussion seems to be about the term magnification and whether it is a 1 or 2 dimensional concept. Did anybody look up the definition of magnification? This might be helpful - http://en.wikipedia.org/wiki/Magnification

1:1 in macro shooting means the subject is the same size in every dimension in real life as it is on the sensor, 1:2 means it is 1/2 the size on each dimension, so the subject takes 1/4 the area it did in 1:1. Link.

This continues as the ratios get to normal sized shooting, shoot a 6' tall person that takes 1" of linear sensor real estate from their toes to their head and they are reproduced at 1:72, but they take 5184 times less sensor real estate. The magnification is 1:72, the area is 1/5184, enlarge the image by 72 on each edge and the paper is 5184 times bigger, it becomes life sized.

 

[privatebydesign]

The magnification that we've been talking about so far is relative magnification in terms of subject size to frame size. It doesn't really matter how big it is in real life in this context. In this context, magnification is referring to the change in relative subject size with a change in resolving power. The change in resolving power can either come from an increase or reduction in focal length, or an increase or reduction in pixel size. Either way, the subject is either magnified or reduced RELATIVE to some chosen reference point. Technically speaking, one need not necessarily involve a frame size...the frame could be infinite. The detail on the subject increases either way...the final images, when observed at 100% size, result in the subject imaged with an increased focal length or smaller pixels will be larger (this, magnified.)

Both definitions of magnification are valid. They just have different contexts within which they are valid.

If you enlarge a square more than another square you are changing the relative magnification. If you say my smaller square is as good scaled up as your bigger square we are not talking magnification, we are talking IQ.

The lens produces the magnification, pixels do not. If you are saying, we consider two different sized things as equal then again, that is not a question of magnification, it is a case of proving the two things are equal, I have proven to myself, and many others, the smaller thing is nowhere near as good as the bigger thing so the entire premise of the discussion is flawed.

All pixels are not equal, for a variety of reasons, it does not further the cause of anything to keep persisting in the crusade that they are.

The latest super tele comparisons are cropping. People are finding that cropping is a viable alternative to TC's and lesser quality longer lenses, something I have been saying against the grain for years. Straight pixels on the subject, the "pixels on duck" meme would have no rebutal for a 400 f2,8 IS MkII getting a sharper shot when cropped as a Tamron 150-600 shot at 600, but I'd bet a lot of money that it would.

Pixel density is a totally spurious, discredited, and fallacious argument. Pixel quality is, and always has been, king.

 

[jrista]

The magnification that we've been talking about so far is relative magnification in terms of subject size to frame size. It doesn't really matter how big it is in real life in this context. In this context, magnification is referring to the change in relative subject size with a change in resolving power. The change in resolving power can either come from an increase or reduction in focal length, or an increase or reduction in pixel size. Either way, the subject is either magnified or reduced RELATIVE to some chosen reference point. Technically speaking, one need not necessarily involve a frame size...the frame could be infinite. The detail on the subject increases either way...the final images, when observed at 100% size, result in the subject imaged with an increased focal length or smaller pixels will be larger (this, magnified.)

Both definitions of magnification are valid. They just have different contexts within which they are valid.

If you enlarge a square more than another square you are changing the relative magnification. If you say my smaller square is as good scaled up as your bigger square we are not talking magnification, we are talking IQ.

The lens produces the magnification, pixels do not. If you are saying, we consider two different sized things as equal then again, that is not a question of magnification, it is a case of proving the two things are equal, I have proven to myself, and many others, the smaller thing is nowhere near as good as the bigger thing so the entire premise of the discussion is flawed.

All pixels are not equal, for a variety of reasons, it does not further the cause of anything to keep persisting in the crusade that they are.

The latest super tele comparisons are cropping. People are finding that cropping is a viable alternative to TC's and lesser quality longer lenses, something I have been saying against the grain for years. Straight pixels on the subject, the "pixels on duck" meme would have no rebutal for a 400 f2,8 IS MkII getting a sharper shot when cropped as a Tamron 150-600 shot at 600, but I'd bet a lot of money that it would.

Pixel density is a totally spurious, discredited, and fallacious argument. Pixel quality is, and always has been, king.

When you throw non-equal pixel quality into the mix, I totally agree. I think I stated a few times before "all else being equal", but you are indeed correct, smaller pixels are generally NOT equal.

In which case, I totally agree, a 600mm lens on bigger pixels trounces a 400mm lens on smaller pixels cropped.

(I was trying to keep noise out of the discussion though, as that would have just complicated the whole deal...however, seeing as I think the magnification vs. area argument is pretty much settled now, bring on the noise! )

Speaking of bigger pixels and less noise...it seems PhaseOne and Hassy finally did it. They put out new MF backs with Sony 50mp 44x33mm MF sensors, that still have 14 stops of photographic dynamic range...plus the increase in signal DR thanks to the larger pixels (which I think are around 5.3-5.5 microns...not as large as a 5D III or 1D X, but larger than any APS-C sensor, and certainly larger than the D800 pixels.) What I would give to have that kind of sensor with a 10fps frame rate, ISO up to 25600, and Canon's 61pt AF sensor!

 

[privatebydesign]

I wasn't just thinking noise, I was thinking much better technique, a bigger magnification needs more support to have the same camera shake (even on the best tripods), any system aberrations are magnified more, AF becomes more critical etc etc.

The crop factor or additional magnification of smaller pixels compared to bigger ones entails at least the linear factor improvements in everything, AF, aberrations, noise, shutter speed, support etc etc.

I believe, very strongly, that is why the 1Dx has been so wholeheartedly embraced over the 1.3 crop 1D MkIV, a camera that was pretty much universally loved by everybody and had that all important sports "crop factor", even when the 1D MkIV has a much higher pixel density.

 

[AlanF]

My first bird photos with the Tammy! Here is the very first shot taken. 5DIII, f/8 iso640, and 600mm, a greylag goose at Fowlmere in Cambridgeshire.

The AF is very reproducible (A1 servo), much, much better than the 100-400L on 7D and better than on the 5DIII. The full-frame of the goose is reduced to 1200x800. The head is 100% crop.

 

[AlanF]

Some shots to show how well the "soft Tammy" at 600mm f/6.3 compares with 400mm f/6.3 upscaled 1.5x. The full-frames are reduced to 1200x800, the crop at 600mm is 100%, the one at 400mm is 100% upscaled by 1.5x. These are at the limits of resolution. the 600mm shows much extra detail, contrary to Pitbull's assertions.

Shot after shot was equally sharp, showing how good the AF is.

 

[jrista]

I wasn't just thinking noise, I was thinking much better technique, a bigger magnification needs more support to have the same camera shake (even on the best tripods), any system aberrations are magnified more, AF becomes more critical etc etc.

The crop factor or additional magnification of smaller pixels compared to bigger ones entails at least the linear factor improvements in everything, AF, aberrations, noise, shutter speed, support etc etc.

I believe, very strongly, that is why the 1Dx has been so wholeheartedly embraced over the 1.3 crop 1D MkIV, a camera that was pretty much universally loved by everybody and had that all important sports "crop factor", even when the 1D MkIV has a much higher pixel density.

Oh I don't doubt it. I drool over 1D X IQ all the time. Despite how good the 1D IV was, there is still a clear difference between it's IQ and the 1D X IQ. The dynamic range of the 1D X (I don't mean the kind that gives you more editing latitude, I mean the reduction in photon shot noise at all levels) really can't be beat. Not unless you reduce pixel size even more.

Besides, if your spending $5000 to $7000 on a camera body, you probably already have or are easily able to get a $10,000+ supertelephoto lens and teleconverters to go along with it.

 

[AlanF]

My very first bird in flight. The greylag is too far away to give a sharp image, and it was underexposed and I had to boost the shadows. I am showing it just to say that the AF locked on very quickly and well. It was iso640, f/8, 600mm.

 

[jrista]

My first bird photos with the Tammy! Here is the very first shot taken. 5DIII, f/8 iso640, and 600mm, a greylag goose at Fowlmere in Cambridgeshire.

The AF is very reproducible (A1 servo), much, much better than the 100-400L on 7D and better than on the 5DIII. The full-frame of the goose is reduced to 1200x800. The head is 100% crop.

Congrats! Detail looks great!

Have you tried any tracking for BIF yet? That was always a weak spot with the 100-400L and 7D IMO. It is so much better with the 600mm, but still not as good as the 5D III.

I love the Greylag Goose. Sometimes a few of them hang out here in Colorado. They tend to mate with Canada Goose, and you get some very interesting offspring. I actually need to head out to Duck Lake in City Park near Denver. This is about the time that the Cormorants and Greylags show up. The Cormorants nest every year on this island in Duck Lake...to the tune of a couple hundred (and plenty more when all the hatchlings come). It's still pretty cold, but I wonder if they are already here...

 

[Old Sarge]

Some shots to show how well the "soft Tammy" at 600mm f/6.3 compares with 400mm f/6.3 upscaled 1.5x. The full-frames are reduced to 1200x800, the crop at 600mm is 100%, the one at 400mm is 100% upscaled by 1.5x. These are at the limits of resolution. the 600mm shows much extra detail, contrary to Pitbull's assertions.

Shot after shot was equally sharp, showing how good the AF is.

Thank you for posting these pictures. The Tammy looks pretty good and the price seems attractive. May have to get one later this year.

 

[mustang]

Some shots to show how well the "soft Tammy" at 600mm f/6.3 compares with 400mm f/6.3 upscaled 1.5x. The full-frames are reduced to 1200x800, the crop at 600mm is 100%, the one at 400mm is 100% upscaled by 1.5x. These are at the limits of resolution. the 600mm shows much extra detail, contrary to Pitbull's assertions.

Shot after shot was equally sharp, showing how good the AF is.

The lens is very interesting. I have one question. I seems that the photos are edited, how much sharpening did you apply?

 

[TWI by Dustin Abbott]

Some good looking shots, Alan. I'm glad the lens is working out for you.