Saturday, May 06, 2017

Takahashi Mewlon 210 Vignetting on Full Frame

The Takahashi Mewlon 210 Dall-Kirkham Cassegrain only has an 18mm image circle at prime focus, and 39mm with the dedicated reducer.

Here are some images of daytime scenes taken with a Canon 6D full-frame DSLR.

Native focal length (2415mm, f/11.5)


With a Televue TRF2008 0.8X reducer/flattener for the TV85 (designed for 600mm radius of curvature) (f/9.2)

and with an Altair Astro 0.6X reducer/flattener (no longer available, but also sold by some other Europe-based sellers) (f/6.9)
 I will try to get some images of a large globular like Omega Centauri when I get the chance, but the illuminated field is not bad at all (an f/4 newtonian with a Paracorr has a comparable field to the Mewlon with the 0.6X reducer!)


Thursday, May 04, 2017

Takahashi Mewlon 210 Maintenance

I've been looking for a Takahashi Mewlon 210 Dall-Kirkham Cassegrain since about 2010, but the new prices are pretty high and I didn't want to pay list price. I was able to find a used one (I am probably the third or fourth owner) on SingAstro for a pretty fair price. The downside was that it had some dings on the paint.

I had some enamel repair paint and some Canon touch-up paint for old grey L lenses (not a perfect color match for Takahashi hardware) but they suffice for now.

After a thorough cleaning (lots of black marks on the Takahashi hardware came off with the use of a rubber pencil eraser) I found several areas where there was paint loss. On the gloss-white tube, I used the enamel repair paint.

Before and after:

A large area of paint loss on the front of the tube:


A smaller area where the paint flaked off, also on the front of the tube:


 A ding on the finder bracket, and after patching with the Canon paint:



The Mewlon also came with a Vixen dovetail, but I wanted to mount it on my Mach1 mount and did not want to buy tube rings (which cost more, and add unwanted weight). I had a short Losmandy dovetail from FarPoint Astro, but it didn't have the holes in the right spots.

So I bolted the stock Vixen dovetail to the FarPoint dovetail and centered it so that I could use the Vixen dovetail as a drilling guide.


And after drilling. The second hole was not very well done; but it's not visible when clamped on the Losmandy saddle; and is definitely not visible at night!


Also, I found that the cheap reticle illuminator from a Seben reticle eyepiece actually fits in the reticle socket of the 50mm finder (not very well, however):


I did feel some grittiness when screwing in the illuminator plug afterwards; I think some metal came off the cheap Seben illuminator due to the not-quite-correct threading. So I probably have to buy the real Takahashi illuminator.

After maintenance, attaching the Losmandy dovetail, and mounting on the Mach1:


Sunday, March 26, 2017

Canon 28-105mm f/3.5-4.5 USM II

TL; DR - use it at f/8 and it has (almost) L-class sharpness.

Back in 2001 or 2002 this was one of the lenses I wanted. It was too steep for me and I ended up with a Sigma 28-105mm f/2.8-4 (the infamously soft and bulky lens) which I used for some time on an EOS3000N and EOS50 until it got damaged (diaphragm stuck wide-open).

Fast-forward fifteen years and I have one from KEH for about $120 in "bargain" condition.

I have two lenses that cover (parts of) the range of this 28-105mm: the 16-35mm f/4 L IS, and the 70-200mm f/2.8 L IS. I was planning to compare these lenses but it turns out that The Digital Picture already has a lens comparison service. So here's the summary (so far as I can tell) on a 5D Mk III (which has the same sensor as my 6D):

Compared to the 16-35mm f/4 L IS (an $800 lens):
  • at 28mm and f/8, the 28-105 almost matches the 16-35mm wide-open (at f/4)
  • this is also true at 35mm
Compared to the 70-200mm f/2.8 L (a $1200 lens):
  • at 70mm and f/8, the 28-105 matches the 70-200mm at 70mm wide-open (at f/2.8) and at f/4, but notably the 28-105 has better corners than the 70-200 (note we are comparing f/8 to f/2.8 and f/4)
  • at 105mm and f/8, the 28-105 matches the 70-200mm at 100mm wide-open (at f/2.8) and at f/4, but the 70-200 beats the zoom in the corners even at f/2.8
The long and the short of it: the 28-105mm can produce almost-L class sharpness so long as you stick to f/8. However, the 28-105 has one massive feature that trumps these L lenses:
It is tiny (about the same size as a 35mm f/1.4) and not much larger than the Nifty Fifty.

My wife and I have traveled a lot with the 6D and 16-35mm f/4 L IS, and it is quite a large and bulky setup, which is why we ended up also buying a Canon G5X (which has a 24-105mm equivalent lens, with IS, and a 1" sensor). The G5X is tiny, but slow (slow to auto-focus and take photos) and has an EVF instead of an optical viewfinder.

I was stuck with the 16-35mm in Monterey in October 2016 when we went whale-watching, and 35mm is much too short for whales.  The 28-105mm would have been useful to have at that time: 105mm long end, f/8 is useful as there was full sun, and the 28-105mm doesn't add much weight or bulk to the camera bag.

Ultimately that's what I see the 28-105mm as: a useful adjunct to an ultra-wide L lens for travel. The 70-200mm is simply too large and bulky to be convenient when traveling. Furthermore, on a 5D Classic, I believe the gap between the 28-105mm and the L zooms would be even less.

I believe there's one reason the 28-105mm f/3.5-4.5 is unpopular and cheap, in spite of its FTM focusing, ring USM, and focusing scale: the zoom range is not very useful on reduced-frame (APS-C) DSLR's.

Tuesday, February 07, 2017

Amazon Polly Just For Laughs

If you've ever had a burning desire to create custom dialog for Consuela the housekeeper in Family Guy..

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
#! /usr/bin/python
import boto3
import io
from subprocess import call

client = boto3.client('polly', region_name='us-east-1')

# aws polly describe-voices --region us-east-1|grep Name
response = client.synthesize_speech(
    OutputFormat='mp3',
    SampleRate='16000',
    Text='<speak><p>Hola. Doggie a puera!</p></speak>',
    TextType='ssml',
    VoiceId='Conchita'
)

# Access the audio stream from the response
if 'AudioStream' in response:
    sb = response.get('AudioStream')
    data = sb.read()
    fd = io.open('hello.mp3', mode='wb', closefd=True)
    fd.write(data)
    fd.close()

    call(["afplay", "hello.mp3"])

Saturday, January 21, 2017

Canon 70-200mm f/2.8L versus Pentax 200mm f/4 Super-Multi-Coated Takumar

I have compared two Canon lenses, the 10-20mm f/3.5-4.5 consumer zoom (on APS-C) and the 16-35mm f/4 L IS here. The results were predictable - the L zoom was consistently better than the consumer zoom, at about three times the cost (and increased weight).

Also, I used to collect old, Pentax screw mount manual focus lenses. These old lenses have a cult following, with various miraculous attributes being ascribed to them (glorious bokeh, an otherworldly quality of the image, etc. etc.) While some of these old lenses (particularly the old Leica lenses) do deserve the praise, I had always wondered how these old lenses compared to modern, non-consumer grade ones.

So in this post, I compare the older, non-IS Canon 70-200mm f/2.8L zoom with a decades-old, manual focus, Pentax screw mount 200mm f/4 Super-Multi-Coated Takumar. The comparison was done at f/4 (since the Pentax manual focus prime cannot open up to f/2.8) on a Canon 6D, which should have plenty of resolution. The target image was a neighboring building at (close to) infinity focus.  The Canon zoom was auto-focused, while the Pentax was manually focused using Live View (it ended up at the infinity stop anyway). Base ISO was used and a fairly high shutter speed (1/2000 second) on a tripod.

First the full image:


The highlighted areas are zoomed in (1:1) for comparison. They roughly cover the center of the frame, and the edge. The extreme edge was not used, because it turns out the Canon 70-200mm zoom is not really 200mm at the long end (it is a bit short).

Let's look at the center of the frame, where both lenses should be performing their best. The zoom is one stop down, so should show improved performance.. but it's a zoom, while the competition is a (wide-open) prime:

Canon 70-200mm f/2.8L at 200mm f/4, center:
Pentax 200mm SMC Takumar at f/4, center:
We can see that the Pentax is ever so slightly longer than the Canon, it has less contrast, and is less sharp. We're talking a $40 lens (at KEH) versus a $1300 lens. I wondered if the Pentax was not critically focused, but it was at the infinity stop, and Live View could not yield a better focus. In any case, "in the field" one would not have the luxury of tripod and Live View and would probably just use the prime at the infinity stop.

The difference is certainly obvious at pixel-peeping distances, but on a regular full-size image the two lenses would be indistinguishable.

Canon 70-200mm f/2.8L at 200mm f/4, edge:
Pentax 200mm SMC Takumar at f/4, edge:
again we can see better contrast with the Canon (better coatings perhaps) and also more detail in the fine vertical lines of the blinds. What is noteworthy is that both lenses don't really suffer much (if any) drop in resolution, going from the center of the frame to the (near) edge.

Is the $1300 Canon a better lens? of course it is - sharper, more contrast, autofocus - but is it thirty times better? that depends on the user.  If autofocus and zoom are important, then these are worth paying money for.

The Canon is quite bulky in comparison to the relatively compact Pentax prime, however.
I was whale-watching in Monterey in October 2016, and did not bring any long lenses (the Canon 70-200mm is very hard to travel with). Ended up with a lot of distant images of whales (using the 35mm end of the 16-35mm f/4L IS). At that time, I would have gladly used the small Pentax prime, in spite of all its shortcomings and non-L image quality.

In conclusion: yes, the Canon 70-200mm f/2.8L is a better lens at the 200mm end than the 1960's - 1970's Pentax 200mm f/4 Super-Multi-Coated Takumar.  But the SMC Takumar is more than good enough and I would say can more than hold its own image quality wise (so long as you don't pixel-peep at 100% magnification).