Canon 16-35mm f2.8 Mk3

Canon 16-35mm f2.8 Mk3

D8E6995 1 400x400 Canon 16 35mm f2.8 Mk3

Wow, it’s a bit big!  That was the first thought I had when getting hold of this lens for the first time – I thought for a second the lovely Leanne at Calumet Manchester had given me 24-70 by mistake.

It’s longer, fatter in the barrel and somewhat heavier than its Mk2 predecessor – but is it any better?

I suppose I can be a bit more objective than most reviewers of this lens when it comes to Canon wide glass because I never use it!

Canon has always seemed to have a different ethos to that of Nikon and TPMs such as Zeiss when it comes to wide glass design.

For sports/photojournalism they have always functioned perfectly well because they are usually quite light, fast to use, versatile, cheap(ish) and adequately sharp for the job -and they’ve sold millions over the years…and rightly so.

But if you wanted a high resolution wide angle with good micro contrast and superb sharpness then, as a landscape photographer for example, you’d be struggling.

Low resolution, poor contrast, vignetting, axial and lateral chromatic aberration, extreme corner distortion and coma are typical faults with wide angle lenses across the board, but Canon wide glass has had more exemplars of these faults than most.

Don’t get me wrong, Nikon have produced some real ‘dogs’ too – just not quite as many!

Let’s face it, no Canon wide could slip into a line up of of Zeiss glass and go optically unnoticed.

When Nikon brought out the 14-24 f2.8 why did Novoflex start flogging mount adapters to Canon shooters?

The only folk who will argue with me are those that have never tried Nikon or Zeiss.

As Canon WA glass goes, the 16-35mm f2.8 Mk2 does an OK job with landscapes, and for the most part produces results very much like the Nikon 24-70mm f2.8 when both lenses are shot at around 26mm to 32mm, but it leaves more than a bit to be desired when being shot wide open.

Its worst fault for me, shooting wide open, is the vile level of COMA.  There’s been many a wide-field astro shot consigned to the bin because if it.

D8E7003 400x400 Canon 16 35mm f2.8 Mk3

Canon 16-35mm f2.8 Mk3 with the Mk2 on the right.

So, is the new Canon 16-35mm f2.8 Mk3 an improvement – it needs to be for the price.

D8E6997 1 400x400 Canon 16 35mm f2.8 Mk3

Canon 16-35mm f2.8 Mk3 with the Mk2 on the right.

The Canon 16-35mm f2.8 Mk3 takes the same 82mm screw mount filters.

There is a newly re-designed lens hood.

Typical of this style of wide zoom, even though its an internal FOCUS lens, like its predecessor, it’s not strictly an internal ZOOM.  The front element moves in and out as the zoom ring is rotated, being furthest forward at 16mm, furthest back at around 26mm and then creeping forward again as we go to 35mm.

Designs like this have ‘compromise’ stamped all over them.  The legendary Nikon 14-24mm does the same sort of external zooming with its front element group, but is furthest forward at 14mm and furthest back at 24mm.  The Nikon is a super wide zoom while the Canon is a super wide to standard wide zoom.  Standard wide angle requires a different element design and layout – so COMPROMISE!

This moving front element makes all lenses designed this way ‘suckers and blowers’ so the cautious among you might want to put one of those lens protect filters on the front.

If you do, then PLEASE, don’t pay thousands for a lens and then be a cheapskate.  You lose light with every air/glass surface you place in the optical pathway.  And a lot of these filters SAY they are optically correct when they are most definitely NOT.  The finest lens in the world turns into a turd if you stick a cheap filter on it.

D8E7011 1 400x400 Canon 16 35mm f2.8 Mk3

Canon 16-35mm f2.8 Mk3 with the Mk2 on the right.

So let’s take a look at vignetting.  We’ll do that in two ways.

Firstly, let’s see how the vignetting at f2.8 changes with focal length, with the Mk2 on the left, and the Mk3 on the right:

Mk3f2.8FLVig 246x400 Canon 16 35mm f2.8 Mk3

Vignetting of the Canon 16-35mm f2.8 Mk II & III at various focal lengths at a constant f2.8

Next, let’s stay at 16mm focal length on both lenses and look at the vignetting through the aperture range:

16mmVgTest 1 407x900 Canon 16 35mm f2.8 Mk3

Vignetting of the Canon 16-35mm f2.8 Mk II & III at various apertures and a constant 16mm focal length.

Now these vignette results didn’t leave me in a state of shock and awe in the slightest.

You need to view the images at 100% to see the subtle improvements in the Canon 16-35mm f2.8 Mk3.

In the first test – maximum aperture vs focal length, the new variant looks equal to or slightly worse than the Mk 2 at 16mm.

But things begin to improve a bit once we are getting towards 24mm.

On the second test – 16mm vs aperture range, again we see the awful maximum aperture vignetting compared to its predecessor.

From f5.6 to f16 it’s perhaps a sliver better than the Mk2. But, notice that the images are a bit brighter.  This is most likely to do with the improvements made to the multi-coatings.

Canon 16-35mm f2.8 Mk3 vs Mk2 Image Comparisons

Let me begin by saying this – 16mm on the Mk3 is NOT the same 16mm that the Mk2 has!

1D9A6348 Canon 16 35mm f2.8 Mk3 1D9A6347 Canon 16 35mm f2.8 Mk3

Mouse over the slider – see what I mean?

Both shots are 16mm @ f11 on the Canon 5DMk4.  The camera was locked down on my heavy Gitzo, and the camera was triggered with a Canon TC-80N3 – in other words NOTHING moved!

The images have not been adjusted in any way – no lens correction profiles – as shot.

Notice the Mk3 image has greater ‘contrast’ and is less flat-looking?

Okay, so let’s look at the ubiquitous ‘brick wall’ test.

We are doing shots on the 5DMk4 using both Mk2 and Mk3 lens variants.

  • at 200ISO
  • at f2.8, f5.6 and f11
  • at 16mm, 25mm and 35mm
  • at a fixed ‘Cloudy B1’ manual camera white balance
  • manual focusing
  • the camera has been re-focused using x10 live view between each frame.
Screen Shot 2016 12 30 at 13.03.17 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window.

Screen Shot 2016 12 30 at 14.24.56 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 16mm @ f5.6

Screen Shot 2016 12 30 at 14.27.56 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 16mm @ f11

The above screen grabs give you a great ‘feel’ for all the differences in contrast and lens colour cast between the Mk2 and new Mk3 variants – these are quite significant.  Even more so when when you look at the vignetting, distortion and AoV differences.

Moving on to the full resolution comparisons:

Again, no adjustments at all other than Lightroom standard profile sharpening, and we are looking at the frame centers at 1:1 ratio:

Screen Shot 2016 12 30 at 13.19.28 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 16mm @ f2.8

At 16mm @ f2.8 (above) the Canon 16-35mm f2.8 Mk3 is noticeably sharper than its predecessor.

Screen Shot 2016 12 30 at 13.31.01 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 16mm @ f5.6

Stopping down to f5.6 @ 16mm yields a better sharpness on the older Mk2 variant.  Is there a tiny bit of improved sharpness on the new Mk3 – perhaps.

Screen Shot 2016 12 30 at 13.42.08 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 16mm @ f11

Now at 16mm @ f11 both lenses seem ever so slightly less sharp.  But that is not down to diffraction as you’ll see later with the 25mm and 35mm tests.  I could be an error on my part when focusing, but for me to make the same mistake on two different lenses is a bit of a long shot.  I’ve re-shot and got the same result – methinks it might have something to do with that ‘compromise’ I mentioned earlier on….or, it could be me!

Moving from 16mm to 25mm and 35mm:

Screen Shot 2016 12 30 at 14.32.48 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 25mm @ f2.8

Screen Shot 2016 12 30 at 14.33.04 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 25mm @ f2.8

Screen Shot 2016 12 30 at 14.34.28 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 25mm @ f5.6

Screen Shot 2016 12 30 at 14.34.47 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 25mm @ f5.6

Screen Shot 2016 12 30 at 14.35.32 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 25mm @ f11

Screen Shot 2016 12 30 at 14.35.46 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 25mm @ f11

Screen Shot 2016 12 30 at 14.36.28 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 35mm @ f2.8

Screen Shot 2016 12 30 at 14.36.44 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 35mm @ f2.8

Screen Shot 2016 12 30 at 14.37.24 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 35mm @ f5.6

Screen Shot 2016 12 30 at 14.37.38 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 35mm @ f5.6

Screen Shot 2016 12 30 at 14.38.21 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 35mm @ f11

Screen Shot 2016 12 30 at 14.38.36 600x375 Canon 16 35mm f2.8 Mk3

IMPORTANT – CLICK IMAGE to view at full size in new window. 35mm @ f11

Make sure you have viewed all the above screen shots at full resolution.

Okay, so we have visually covered iteration comparisons for the Canon 16-35mm f2.8 Mk3 and its predecessor in terms of distortion, vignetting, field/angle of view and sharpness.

In terms of stopped-down sharpness, on the Canon 5DMk4 at least, I’d expect to get into the realms of aperture diffraction around f14 to f16.

Wide open the Mk3 version stomps all over the Mk2, and I think it stays ahead through to at least f11 across the entire focal length zoom range.

Chromatic Aberration

The Mk2 16-35 f2.8 has a somewhat noticeable chromatic aberration problem, so how does the new Mk3 version measure up in comparison – both shots are 16mm @ f11:

1D9A6347 Edit 2 600x400 Canon 16 35mm f2.8 Mk3

Click the image and a full size jpeg (80% quality) will open in a new window.

1D9A6348 Edit 2 600x400 Canon 16 35mm f2.8 Mk3

Click the image and a full size jpeg (80% quality) will open in a new window.

Compare the lamp post on the right and the window and alarm box on the left of the shots.

Though still present, chromatic aberration is much reduced on the new Canon 16-35mm f2.8 Mk3.  Along the middle axes of the image – especially the horizontal – there have been big improvements.

The Lightroom ‘remove chromatic aberration’ function cleans the raw file up beautifully without having to go anywhere near the manual corrections – just tick the checkbox. But doing the same to a Mk2 image usually leaves vestiges of both red and green fringing at the frame edges at 16mm.

Coma Test:

What’s Coma? It’s a lens design flaw which renders ‘tails’ and ‘wings’ on off-axis points of light.

And here is a shining example, courtesy of the Mk2 16-35:

J6Q7329 Canon 16 35mm f2.8 Mk3

Coma on the Canon 16-35mm f2.8 Mk2 variant.

Mmmm…yummy!

It’s not exactly the best time of year for Milky Way astro shots here in the UK – New Year as it is.  But we ventured out at midnight the other night just to test this Mk3 version of the lens.

The area is fairly local and surrounded on all sides by huge light polution but it served the purpose of the test.

Shooting wide open f2.8 @ 6400ISO, stacking 8 shots done in quick succession here’s the truth about the coma on the new Mk3 16-35mm variant:

1D9A6327 18 600x400 Canon 16 35mm f2.8 Mk3

Click the image and a full size jpeg (80% quality) will open in a new window.

It’s not the best astro you’ll ever see, but it does show that the coma is still there, but it’s a lot less intrusive.

In Conclusion

So there we have it – the new Canon 16-35mm f2.8 Mk3.

Is it better than it’s Mk2 predecessor?  Well yes, it is – and in pretty much every aspect I’d say.

The vignetting at 16mm f2.8 is quite strong – nearly 4 stops darker than the image center.  This WILL cause you problems if you have peripheral deep shadow areas, as even on the 5DMk4, pulling 4 stops will make the shadow areas go a bit noisy.

I also think that 16mm is now more like 18mm, but what’s a couple of mills between friends ehh!

Would I buy one? Well, that depends.

If I had a Mk2 variant and needed the lens format then I would be looking to trade in immediately.

Wedding, street, sports/photojournalism and events photographers would be mad if they didn’t have one of these in their bag. And I think wildlife photographers would benefit as well – I reckon it would be perfect on the 1DX Mk2 for just about anything.

Not being a Canon shooter for anything below a 200-400 I won’t be putting it on my ‘wants’ list at all, but if you are ‘Canon-only’ then I strongly recommend you have a look at this lens.

As for landscapes and wide field astro, erm…..let’s just say there’s more than one way to skin a cat, and some are better than others.  Having said that, if you are a landscape shooter with a Mk2 variant and you can’t afford/ just don’t want a plethora of glass for specific tasks then it’s a big improvement on what you’ve already got.

Where to buy this lens in the UK – buy it here Calumet Photographic

Many thanks to Reece Piper, Leanne and Richard from Calumet for loaning this lens for the purposes of review.

And a big thanks to June Lown for the loan of the Mk2 to make the comparison.

17.5 hours that’s taken – Jesus, it’s like having a full-time job!  If this review has been useful to you then please consider chucking me a small donation – or a big one if you are that way inclined!

Many thanks to the handful of readers who contributed over the last week or so – you’ve done your bit and I’m eternally grateful to you.

Happy New Year everyone!

Prospec USB 3.0 card reader

Prospec USB 3.0 card reader.

A few weeks ago I decided that my Mac Pro4.1 early 2009 needed to be upgraded to USB 3, seeing as I’d not long since fitted it with an SSD for the OS – yes, I found myself in the mood to spend money; obviously I was unwell!

So I bought a 4-port expansion card from Mac Upgrades, installed it in the Mac and connected my 8TB G-drive – and all has worked perfectly ever since.

Being a mainly Nikon user, the XQD reader showed an instant improvement in transfer speed too, but for all my CF cards I was still stuck with my ancient Sandisk USB 2 reader.

I perused a few USB 3 reader prices in was left gasping – obviously I was now feeling a little better!

But then I spied this:

D4D3688 Edit Prospec USB 3.0 card reader

Prospec USB 3.0 Multi-Card Reader

Twenty quid! That’s right, just £20 – bargain!

Real world read-speed testing using Black Magic, and 32 GB Prospec 1010x CF cards yield the following results:

  1. Sandisk USB 2 reader – 36.9 MB/sec
  2. Prospec USB 3.0 reader – 112.7 MB/sec

and just as a comparison, the Sony XQD reader and H-series card averages 139.2 MB/sec.

Previous blog posts have shown you that I’m a fan of Calumet Prospec CF cards; mainly because they are re-badged Delkins, and in my experience simply bomb-proof and good VFM.

I can’t say for sure without checking, but this Prospec USB 3 reader looks VERY much like a re-brand of the Delkin USB 3.0 multi-card reader, but is basically £10 cheaper.

It certainly sucks up uncompressed D800E 14bit RAW files at an impressive rate of knots I can tell you.

In my earlier Prospec CF card post I did allude to the slightly odd fact that the larger the CF card capacity was, the faster its read speed became.  I also bought a 64Gb Prospec 1010x, just to give the D800E more capacity for shooting HQ time-lapse – this card clocks a read-speed average of 119.8 MB/sec – basically 7MB/sec faster than its 32GB cousin.

Yes, there’s cards and readers out there that might yield faster results; but at what cost to your pocket?

But this level of “REAL WORLD” performance is plenty fast enough for yours truly – especially if, like me, you have short arms and long pockets!

You can view the product HERE – BUY one!

 

Prospec 32Gb UDMA 7 CF

 The Prospec 32Gb UDMA 7 Compact Flash Card from Calumet – review

Prospec Prospec 32Gb UDMA 7 CF

The Prospec 32Gb UDMA 7 Compact Flash Card from Calumet – click image to visit the product order page.

Because I’m something of a photographic ‘old fart’ and have been taking images for a living since before the venerable A1 was a glimmer in some Canon engineers eye, I treat everything new with suspicion!

Back in those days when we shot on film, suspicious gits like me had our favourite films we used.  These were whittled down from the vast array on offer, not only on the basis of their performance, but also on reliability.

Did the sprocket holes tear in a particular cameraif they did then to avoid imminent disaster you’d never put that film in that camera.

Were the ‘tails’ always taped to the cassette spool? Christ, that one nearly cost me a boat-load of money at a wedding once – both bride and groom were barristers!

Thank the Lord we don’t have problems like that any more.

But CF cards come with their equivalent problem – card failure.

CF cards are basically Flash RAM, just like SSDs in the latest computers.  And as such they are prone to some degree of instant failure over time.

 The Way I Use Camera Storage Media

When I’ve finished shooting something I move it from the camera storage to my main imaging machine as soon as possible.  Once the shots are on the computer, the card goes back in the camera and is immediately reformatted.

Back when the D3 was Nikon’s flagship I’d had some clonking failures of Lexar CF cards and Sandisk Ultras, when Kevin Treadwell at TFC Birmingham put me on to Delkin cards – and I have never had a problem with them………(see footnote)…

And the two original 16Gb Delkins I bought for the D3 are still going strong to this day!

So when I eventually moved to the D4 I wanted a 32Gb Delkin UDMA 7 to go with the 32Gb Sony XQD, but I couldn’t locate one anywhere.

I was moaning about this to Stuart Tudor-Wood at Calumet Birmingham and he suggested I tried their own Prospec 32Gb UDMA 7 instead.

Can you imagine what the “suspicious old photography fart” in me was saying; I was used to possibly the most reliable memory on the planet, and the priciest; and here was something I’d never heard of, and costing pretty much the same..

I did manage to walk out of the door without paying for it, so that was a small victory – but it only lasted 3 weeks until I was doing another workshop there – he nailed me for it the minute I walked through the door!

So, here we are, two years later.  The other day I realised how long I’d had this card in regular use.  It’s a mighty rare occurrence for me to have a day shooting long lens action and not generate some use of the CF card.

Even if I’ve only shot to the XQD card, force of habit leads me to format both cards one after the other.

And if I’m going to shoot video I shoot that to the XQD and then remove it if I want to shoot stills, which consequently go to the CF card.

When I’m testing lenses and shooting for workshop slides etc I shoot to the CF card, then stick that in my steam-powered laptop – it can never see my XQD reader and I daren’t even think about asking it to accommodate tethered shooting..

I bought the card a new brother so I had two of them in the 1DX when I went to Norway in February this year.

All in all, looking back through all the image batches I’ve shot I reckoned that this one card had been reformatted 981 times in a selection of camera bodies from a 1DX and 5DMk3 to a D4S, D4, D800, D800E and D810.

In the D4 this card allows me to shoot 41 14bit uncompressed raw files before I hit the end of the camera buffer; which let’s face it, is plenty.

So I promptly set about shooting 20 bursts of 41 frames, wiping the card between bursts, just so that I could say with some degree of certainty that the card had just hit its 1000th reformat.

Now THAT is a good, solid and highly reliable card that has put up with more abuse in the last two years than some of my clients would give their CF cards in a lifetime.

What Do The ‘Hieroglyphics’ All Mean?

I must admit to being somewhat ambivalent to the majority of these speed numbers – at the end of the day, as a photographer you are more interested in sustained write speed than you are in read speed – or at least you should be.

The faster the write speed of the media the longer it takes to hit the end of the camera internal buffer – this means you shoot more sequential frames in a burst.

CF6 Prospec 32Gb UDMA 7 CF

This logo means the media is compliant with the CFA’s CF6 specification, which is supposed to provide data transfer rates of 167Mb/sec.

But that quoted transfer rate could be either read or write – they never seem to commit!

CF6 specification means UDMA 7 compliance; whereas UDMA 6 equated to CF 5 specification and transfer speeds of 133Mb/sec.

CF cards have to communicate with a host device – in our case our camera; and UMDA is the most efficient and fastest mode of communication. But not ALL cameras are UDMA compliant.  And those that are might not be UDMA 7 compliant.

If a camera is not UDMA compatible then believe it or not, a slower non-UDMA card might make the camera work faster. If you put a UDMA 7 card in a camera that is only UDMA 6 compliant then that’s fine, but communication between the two will be at UDMA 6 speeds.

The moral here is to check your camera specifications, and available firmware upgrades.

What does 1010x mean? Known as the card Commercial X Rating it’s an indication of read speed more than anything else, and basically relates the speed of the card as a multiple of the old CD-ROM standard of 150KBs.  So 1010x equates to 151.15Mbs.

But here’s the thing; none of these speeds, theoretical or otherwise, are derived via a camera – they are all acquired on a test-bench computer and a variety of card interfaces.

There is a simple if slightly ‘rough ‘n ready’ test that you can do to check the camera/media combo write speed:

  1. Set the camera to its fastest RAW shooting frame rate (Canon 1DX users note, that’s NOT 14fps!).
  2. Set the burst length to 30 frames.
  3. Put the camera in manual mode, auto iso, and set the shortest shutter speed.

Get a stop-watch and be prepared to start it when the ‘data write’ indicator lights up.

Press and hold the shutter button to start the burst of exposures, take your finger off the shutter button when the camera stops shooting.  The indicator light will still be on, and the stop watch should still be running.  Stop the clock when the light goes out!

Multiply the number of frames by the size of your RAW file, then divide the result by the time in seconds and you’ll have a rough value for your data write speed in Mb/sec.

In a Canon 1DX the Prospec 32Gb UDMA 7 CF card chugs away at a highly respectable 69Mbs write speed.

32Gb of storage with great reliability and ‘real world’ write speed like that is great – especially for the price.

Nikon Users

If, like me, you mainly shoot uncompressed 14 bit RAW (not an option for Canon shooters) the write speed of the Prospec 32Gb UDMA 7 CF card in a D4/4S will drop to around 41Mbs due to the much increased file size of each RAW frame – 35.9Mb average RAW size, as opposed to the 1DX average file size of around 26.4Mb.

But sensible burst shooting in conjunction with the huge Nikon D4/4S buffer means you will never suffer from ‘buffer lock-out’.

Conclusion

Bearing in mind that this is just my opinion, the Prospec 32Gb UDMA 7 CF card represents exceptionally good value for money.  Yes, there are faster cards out there; but I’ve been bitten by both S****** and L**** in the past, and ‘once bitten twice shy’ and all that……

For me the write/format longevity of this card is what strikes me the most, and for the price this card is well worthy of anyone’s consideration.

D4D2589 900x599 Prospec 32Gb UDMA 7 CF

Two years and over 1000 reformats, the Prospec 32Gb UDMA 7 CF card is still going strong – have a Kite!

Footnote

I’ve just had it confirmed……..the Prospec is actually Delkin; so bearing in mind what I said earlier, the reliability comes as no surprise!

Please consider supporting this blog.

This blog really does need your support. All the information I put on these pages I do freely, but it does involve costs in both time and money.

If you find this post useful and informative please could you help by making a small donation – it would really help me out a lot – whatever you can afford would be gratefully received.

Donations would help offset the costs of running this blog and so help me to bring you lots more useful and informative content.

Many thanks in advance.