Infrared 'faux colour' video from an IR-converted (720 nm) DSLR and Final Cut Pro X

Shortly after my new-to-me infrared-converted (720 nm filter) Nikon D7100 arrived the week before last, I realised that I could try to make 'faux colour' infrared full hd videos. However, I had no idea whether with the software available I could process the footage. A day or two later a spell of wall to wall sunshine arrived and so I wandered down to the Auld Brig o'Doon and stuck my camera over the wall just to get some footage to import in Final Cut Pro X. In the interests of health and safety I did stand well away from the carriageway just in case a drunk came by on a horse pursued by witches who gather at intervals at Alloway's Auld Kirk.

Safely back home, I imported the footage into FCPX and found that converting the footage was actually easier than doing the same job with still photographs using the widely-known techniques and which I have written about at length in previous posts.

I have had installed for some time the excellent and free colour effects plugins produced by Alex Gollner (Alex4d). Included in that bundle is Channel Mixer a4d and with that plugin the red and blue channels can be reversed to produce the blue sky effect. Very soon I had worked out a protocol that worked best for me, with the proviso, of course, that other settings and effects could be applied to produced a whole range of 'looks' in the final output.

To explain what I have been up to I have made a short video which explains the steps in Final Cut Pro X:

I shall not repeat the procedure here but there are a couple of things to point out. Whoever converted the camera for the previous owner managed to set a custom white balance. The raw footage has that white balance applied. Others have found it extremely difficult to set a custom white balance in an infrared-converted D7100, and I have not yet found a way. The 'Balance Colour' first step of the conversion process is, therefore, on footage being taken with the custom white balance. Without a custom white balance, (i.e. white balance on 'auto' on the camera) I cannot get reasonable colour using the Channel Mixer; in these circumastances 'Balance Colour' does a reasonable job to bring the white balance back in line but not quite a good enough one. Therefore, a custom white balance set as advised for the particular IR filter on the sensor would appear to be essential.

I should also point out that my experience is confined to infrared-converted cameras, and does not extend to those uncoverted cameras with an infrared filter attached to the lens.

Fisheye Lens for Suitable for Infrared Photography: Kelda 6.5 mm f/3.5

From ePhotozine

I have been looking for a fisheye lens for infrared photography. I found it difficult to find out which fisheye lenses, if any, show a ‘hotspot, and I was reluctant to buy a second-hand Nikon or one of the still relatively expensive ones from the second-line manufacturers.

A couple of weeks ago I saw in ePhotozine that Kauser International were offering a Kelda 6.5 mm f/3.5 fisheye for APS-sized DSLR sensors for £129. Now, I am usually very wary of buying a cheap lens. The reviews of most reveal poor optical performance. However, the earlier review in ePhotozine of this lens showed it to be remarkably good at f/8 and f/5.6. I then found that this lens is or has been sold by a number of companies, some in slightly different configurations, starting, it seems with Samyang and described either as 6.5 or 8 mm focal length. Although Samyang appear to make their lenses in Korea, the Kelda is described as made in China. 

I ordered one (with a Nikon mount) on the grounds that if it proved not to be suitable for infrared, then I had lost very little. It arrived from Kauser virtually by return. It looks like the early model Samyang. The fact that it is manual focus with a manual aperture ring is no hardship. I was impressed by the build quality: a solid feel, smooth focusing ring and half-stop clicked aperture ring. The supplied plastic front lens cap has a faux spatter-coated sheet metal look. On this model the petal lens hood seems to be fixed (well, I can find no way to remove it and the plastic flange extends below the front element of the lens) even though the box states that it is removable.

The first test was my usual shot out of the window lens test. The weather has been so foul and so dull that a meaningful test was out of the question. However, at first glance I could see no sign of a dreaded hotspot. Yesterday, however, the weather was fine, and on the way back from the post office I stuck the lens (attached to my 590 nm infrared converted Nikon D80) into the Auld Kirk to check that it really was suitable for infrared. As you can see from the following Lightroom, Photoshop and Viveza-processed shots, there was no sign of a hotspot.

Reviewers of this lens in it various guises have commented, hardly surprisingly, on the flare when shooting against the light. With this lens’s coverage of 180° on the diagonal of the sensor, it is hard to avoid getting the sun (or one’s own shadow) in the frame. Reviewers have also marked the lens down for its lack of a depth-of-field scale. With the enormous depth of field of this lens (whether it is nominally 6.5 or 8 mm focal length), that absence, together with the absence of an infrared focus mark, is completely irrelevant. At f/8, the hyperfocal distance is about 30 cm which means that when focused at that mark, everything is in focus from the end of a six inch rule to infinity.

The Samyang fisheye lens family to which this version clearly belongs is noted for the stereographic projection of the image. The advantages of this projection over others are shown here.

In conclusion, I have a fisheye lens that works for infrared with an optical performance (ePhotozine used Imatest) that hits excellent between f/5.6 and f/11 with minimal chromatic aberration at f/11. I would rate this lens at the price I paid as a bargain. So far, so very good.

Samsung withdraw from the camera market. Why I am not surprised

The withdrawal of Samsung from the camera and camcorder market is sad news for those who invested in the brand. Only recently were Samsung cameras being pushed by some magazines and the overall marketing was intense. I presume the company had not achieved the share of the market it had planned.

I was never remotely tempted to buy a Samsung camera, the reason being that I have never been satisfied with anything Samsung I have bought. The software for a pre-Android telephone was appalling and the interface the cause of much frustration. Worse was to follow: a 'smart' (i.e. thick as two short planks) television with video recorder/DVD BluRay player. The latter has the worst user interface I have ever known in any electronic device, needing multiple clicks and arrow shifts just to play back a recorded television programme; it is not very reliable and sends inane messages to the television screen when in operation. I loathe that machine. I have a non-Samsung brand large hard drive which registers as a Samsung drive on the Mac; it is the noisiest drive I have ever encountered. So, you do get the message that as far as I am concerned the Samsung brand is contaminated.

More generally, large manufacturers with different divisions must live in fear and dread of a dud product contaminating the brand. For much of the 1970s and 80s I used to go to a conference in New Hampshire. It was held in one of the liberal arts colleges that American parents waste their money on. Initially, the food was good, very good, in fact, if you can stand piling all the sweet and savoury items on one plate. In later years the food deteriorated. The catering had been contracted out to a major American hotel chain. 'I would never stay at one of their hotels', said a student who was working there for the summer. The premium brand was contaminated by the poor performance of one division which probably had nothing to do with the outside catering business. The heads of Canon, Nikon, Sony and Panasonic must tremble as new products are launched because a number are simply not best in their class. A manufacturer once told me that if they produced anything it had to be the best or second-best of its class in the world and, he added, certain of achieving a 40% margin! And no it wasn't Apple.

Sony FDR-AX33 Camcorder: Zoom Range and Steady-Shot

I ended my last post by saying I was puzzled by the zoom range of the AX33 and the effect of using Active Steady-Shot (Balanced Optical SteadyShot, BOSS). This is a snippet from the manual:

And that is all it says. I wondered if this affected both ends of the zoom range. While you might expect zoom magnification to be noticeable at the long focal length end of the range, is there also an effect at the short (wide angle) end?

The additional apparent focal length is very welcome indeed for wildlife. Sony’s 10x zoom range has been extremely limiting in the past and I have carried a 1.7x extender for previous models.

I have extracted some stills to show the effect of the three Steady-Shot settings: Active, Standard and Off at both ends of the zoom range. I used a Full HD setting, rather than 4K since the effect is more pronounced.

Zoom at it longest focal length with
different SteadyShot settings

Zoom at its shortest focal length with
different SteadyShot settings. The angle
of view is slightly narrower with Active; the
window frame is included to show the effect
more clearly

As you can see there is a huge effect at full zoom between Active and Standard or Off. There is an effect at wide-angle but it is not so pronounced. However, it is worth noting that if you seek to gain the longest apparent focal length then set Steady-Shot to Active. On the other hand if you want the widest angle shot possible, set SteadyShot to Standard or Off, depending on how the camera is held.

What Sony does not mention is whether there is any effect on video quality of using Active Steady-Shot to get a longer effective focal length. I have seen it said that there is a very small effect which is hardly noticeable but that it the only information I have found.

Having had the camcorder since the summer, I find it has already been superseded - by the newly-announced AX53 with a true optical zoom range of 20x. Still the same viewfinder though.

Sony FDR-AX33 Camcorder: On location

It is one thing testing and using a camcorder around the house and on tame holidays; it is another using it in the wild, in extremes of weather and on land, air and sea to video wild life varying in size from a Blue Whale to a small insect. Neither is it useful to read most reviews and forum comments from ‘film makers’; they are looking for something entirely different.

I have now put this camcorder through the wild test for the first time and so I can add to the comments I made in my post of 12 June.

The video quality was excellent in medium to good light; less so, with noticeable apparent compensation for noise, in poor light. I made use of infrared Nightshot as I thought I would (and a good reason to buy Sony). Active Steady-Shot (BOSS) proved to be superb.

However, the electronic viewfinder (EVF) although improved from the last Sony I had is still a weak point: too dark (with no control over brightness) and with a seemingly low refresh rate. It is still an essential in a camcorder but all I can say is that it is better than nothing. By comparison with the EVF on a modern CSC, it can only be rated as poor.

I missed the GPS built into my last Sony. I carried a separate battery charger (essential when travelling but not supplied with the camcorder) and four batteries.

One annoyance was that after being outside in the sleet in the mountains of Tasmania, condensation misted the lens when back in the warmth. The condensation was not simply on the outside of the lens. I eventually got rid of it by leaving the camera on (i.e. taking video) for about 10 minutes; as it heated up the internal condensation cleared. I though condensation problems disappeared with the last tape drive but this was clearly not the case. Incidentally, when cleaning the lens, it is a case of first catch your lens; the lens assembly moves on it gimbal as you try to clean the surface.

The USB lead which emerges from the side and is tucked into the strap was a small worry; it seems too vulnerable to damage.

I am, overall, very pleased with this camcorder. I am also pleased that that is the case because there are not many options in the fairly light and useable (for my interests) camcorder category. However, I do feel Sony could have done better, as I said originally, by including features like GPS and charging more.

One aspect has been puzzling me, and others too judging from the comments on forums. That is the non-digital zoom range with different Steady-Shot settings. That will be the subject of my next post.

‘Deep’ Infrared photography using a 850 nm filter over the lens of a 720 nm or other converted DSLR: Handheld Shots

The one drawback of using an 850 nm filter over the lens of a DSLR converted with a filter of a lower cut-off (e.g. 720 or 590 nm) is that the viewfinder is completely dark. On a tripod it is an easy matter to align the camera, screw the filter in place, and adjust the exposure.

With later Nikon DSLRs models, like the D90, the solution to using the camera off the tripod is easy: use Liveview, However, ready-converted Nikon D70s and D80s are popular in UK especially for those trying infrared photography for the first time. With no Liveview it would seem there is no way that they could be used with an 850 nm filter for hand-held photographs. But there is a way: put a cheap direct vision finder of the type used for rangefinder cameras in the accessory shoe and use that to line up the shot when the filter is in place—not to frame the shot necessarily but to align the camera to the central point already identified without the filter in place.

The procedure then is simple: Without the 850 nm filter, frame the shot through the normal reflex finder and remember where the central point of the photograph lies in the scene (easy with a central focus point showing in the finder); add the filter and then with the direct finder aim the centre of that finder at the spot identified earlier, and take the photograph.

Provided you aim for the centre it does not matter what focal length the direct vision finder is simulating, although it does help if the frame of a longer focal length lens is outlined. To reiterate the finder is being used as a direction indicator or gunsight, not as a means of framing the shot. This method works, of course, for all manner of zoom and fixed focal length lenses since the frame shown is of no concern.

The photographs show a brightline finder fitted to a Nikon D80. The central point is easy to see because of the marking for a 135 mm focal length lens (on what it was designed for, a full-frame 35 mm camera). Incidentally, I do not recommend the finder shown (Helios, made in Japan); a poorer article it would be hard to imagine since it was poor in both design and construction. I got this some time ago from an eBay seller but only got round to using it more recently. The first thing I noticed was the name ring seem misplaced but then I saw that the frame were moving around inside the finder. I took it apart. It had been crudely glued but the non-original as well as the original adhesive had come unstuck so that the frame (on a piece of film and glued between two bits of glass) was rattling about. Fastening it back in so that the centre of the view was at the centre of the frames was not easy since there was no fixed point for attachment. I assume the factory must have had some sort of jig to drop the film-glass sandwich onto drops of glue. I put the whole thing back together, aligning the frames by eye, and it does now align with the centre of the proper viewfinder (at what is effectively infinity) pretty well spot on.

‘Deep’ Infrared photography using a 850 nm filter over the lens of a 720 nm converted camera

I have repeated the trials I did with an 850 nm filter on my 590 nm-converted Nikon D80 with my new-to-me 720 nm-converted D90. I needed to increase the exposure by two stops compared with the value metered by the camera without the filter in place; as before I had to do this on manual because metering through the filter produced nonsensical results.