Let’s sort this out… Ever since an article went live over at www.newsshooter.com I’ve had an email or a phone call almost once a day asking me if the new Sigma Cine Zooms were parfocal. Or as many of the calls/emails referred to it: “parafocal, paranormal, parafocusing, paralegal”, etc. etc. It’s parfocal (par·fo·cal), okay?! It seems like the entire industry suddenly became hyper aware of this term which has been around for a very long time and overnight, everyone’s a lens expert. Let’s make sure we’re all on the same page…
What does it mean for a lens to be parfocal? Basically, it means that the lens holds it’s focus while zooming. If you set your zoom at the wide end, focus on your subject, then zoom in – the image should stay focused throughout the entire zoom range. This is something that cinema lenses take into consideration during the design phase and will compensate for and sometimes even have adjustments for after they’re assembled. Still photography lenses are generally NOT parfocal, despite the occasional “well I got a good copy that IS parfocal”… Manufacturers can save a lot of money and assembly time by not including measures to keep a lens parfocal. Besides, with a still photo lens, you press the shutter button half way and the lens focuses instantly, regardless of where you have the zoom set. So it really doesn’t matter in the still photo world. This brings us to the new Sigma Cine zooms. They ARE based on still photo lenses which would imply that they were not designed to be parfocal.
First we’ll take a look at how we test for this. We use a simple chart to illustrate what we call zoom curve. Zoom curve is something that every zoom lens exhibits to a degree. Observing this curve will depend on how discriminating we are in the evaluation process. We use different tolerances for different lenses. For example, wide angle lenses have a much tighter tolerance than telephoto lenses, simply because of the sensitivity of flange depth on wide angle zooms. The unit of measurement is usually hundredths of a millimeter or thousandths of an inch and is obtained on our auto-collimator bench. This allows us to evaluate and record the discrepancy in zoom curve and adjust accordingly. Every lens that comes through Duclos Lenses for sale is subjected to this test, among many others, to ensure the lens is performing the best it can. Performing quality assurance tests at this stage gives us the opportunity to make any adjustments that are needed before delivering a final product from Sigma, or any other brand for that matter, to our customer. So let’s see how the Sigma stacks up!
Pretty damn good! I picked a random 18-35mm from our stock and another popular $20,000+ zoom lens which will remain nameless as to not bash any brand or model (I also truncated the zoom range values to further hide it’s identity). As I mentioned earlier, depending on the specific zoom range, we can alter the parameters to be more or less discerning. In this case, we’re using a range of +/- 0.1mm. At wide focal lengths, if a lens exceeds the 0.1mm range, something is really wrong… Let’s look at the Zoom Curve Charts below:
If you still aren’t quite clear as to how Zoom Curve works: flat=good, curvy=bad. So you can immediately see that YES, the Sigma Cine 18-35mm exhibits some Zoom Curve. This is why the phrase “nearly parfocal” has spread like wildfire. If all lens manufacturers were being honest, they would label every lens as “nearly parfocal”. While this particular Mystery $20k zoom is obviously performing much worse than the Sigma, it’s still not bad and would most likely go unnoticed, especially at smaller apertures. Moving on to some more details: our measurements give the Sigma a curve of +0.019 to – 0.009, a range of +/- 0.028mm. Not bad at all – in fact, quite respectable.
So why did Sigma call their new cine zooms “nearly parfocal”? I have no idea! My best guess is that Sigma, being new to the cinema lens market, is being honest to a fault. I’ll continue to test additional lenses to monitor copy variation, but so far, the results shown above are consistent, even from the earliest models we received from the factory. I really don’t have any motivation to prove that Sigma Cine zooms are or are not parfocal, but I felt obligated as a result of the influx of questions in regards to this particular feature, on these particular zooms. I could spend days comparing lens A to lens B and see who comes out on top, but then I’d have a bunch of enemies at the factory. Kinda like the T-stop Mega-Test that I scrapped, or the moral dilemma I’m currently facing in regards to the Celere tests… To answer the question “are the Sigma Cine zooms parfocal?”… No. No they’re not. And neither are any other cine lenses if you look hard enough.
Thank you, Matt for clearing that up.
Please correct me if I overstate what you wrote, but that means that there is no such thing as a true 100% parfocal lens (in the cinematography and photography business) !? Which is actually pretty intriguing 😉 So what we non-lens-tech guys accept as parfocal is what Sigma coined as “nearly parfocal” and all “parfocal” lenses exhibit a zoom curve to some extent or another.
That pretty much sums it up. If you look close enough, you’ll find zoom curve in just about any lens.
And since you started on the Sigmas, there’s an ongoing discussion about the 50-100 and it’s characteristics. So a little insight on breathing, image shift, focus shift and the terminology (because I believe this is also sometimes all over the place) and how it is connected to lens design would be a great resource, I believe.
This particular post was focusing (pun intended) on zoom curve and if the Sigma zooms are parfocal or not. The 50-100mm does breath. A good amount.
Matthew,
Thanks for taking the time to sort out the parfocal question…only to raise more. You said the zoom curve on the Mystery lens would most likely go unnoticed. And, that other cine lenses are not perfectly parfocal. Is it worth your time and effort to compare these? And perhaps give a subjective answer to when does a zoom lens fall into questionable territory?
I did mention this to a degree… Once a lens leaves the +/- 0.1 tolerance, all bets are off. It’s not acceptable. Depending on the lens, especially wide angle lenses, even +/- 0.05 is encroaching in the unacceptable range. You’ll find that most non-cine lenses exceed this parameter.
Thanks for this!
I’m curious about your moral delima with the Celere lenses…
First off, I believe they have evolved over time, and the copies you got from Braincell Camera are very early versions. Being such a new company, I hear they’ve made numerous improvements in the new models. Also, I’m not sure anybody is claiming these lenses are optical performance kings… It’s all about the character. There aren’t many modern spherical cine lenses (In the Celere price range) that embrace optical character. I think that’s what intrigues most people (Including myself). That along with the price, and physical design (Small, compact, unified size and weight).
I guess I’m curious if the delima you face is telling people that they don’t perform as good on the test charts as other lenses, and therefore aren’t worth the hype. If that’s the case, I don’t think you have to worry. That’s probably not why people want them, so any criticism should only help the company in the long run.
Apparently I should have left that part out altogether. I have no qualms at all with the performance of the Celere primes. We’ll leave it at that.
Sweet! I hope to hear more of your findings about them soon.
On a more related note, I also hope you do a similar tests with the Sigma 24-35mm when it finally drops.
Thanks for the explanation Matthew. Is the resulting graph similar for the 50-100mm T2?
Yes. We tested several 50-100mm zooms as well and had similar results with the exception of one unit which was well outside our tolerance. We performed a few minor adjustments and were able to bring the curve down to something similar to what you see here with the 18-35mm.
Thanks for this Matt. Super informative. Every lens has its weaknesses, being informed just allows you to choose the right tool. I sat in on a barrel distortion test between 2 expensive zooms. They both had it but were just different in where they chose to “put it”.
So much good info in this post, I appreciate it. Personally I don’t feel like parfocal/varifocal is all that much of a big deal when it comes to real world production (for me at least) I’m not snapping zoom around DURING a take.
I think it’s good that you’re pointing out the finer details because it gets us familiar with the products. I think it’s OK that some of these lenses are better than others. Because they fill a need either way!
A very nice article. Thanks very much indeed because I’m trying to work out what lens/lenses to get for a panasonic GH5 and no-one ever seems to test how close any are to holding focus while zooming. I’ve heard that the Panasonic Leica 10-25mm f1.7 works quite well with the latest firmware update but can’t find any tests. {hint!} Thanks again