My Lenses: A Focal Length Comparison

Yesterday, I received my Voigtlander 15mm (e mount) lens. This with my recently added Sony 55mm should finally complete my lens kit for the foreseeable future. To celebrate I created a lens focal length comparison of all my lenses in my admittedly unexciting backyard. Note that this on a full frame camera (a7ii), but the relative differences would be same for a crop sensor camera. The results and full list of lenses are below:

Full Size Focal Length Comparison Image

My Lenses

  • Voigtlander 15mm f4.5 iii
  • Canon nFD 20mm f2.8
  • Canon nFD 24mm f2
  • Sony Zeiss 35mm f2.8
  • Canon nFD 50mm f1.4
  • Sony Zeiss 55mm f1.8
  • Canon nFD 85mm f1.8
  • Minolta MC 100mm f2.5
  • Minolta MD 135mm f2.8
  • Sony 55-210mm f4.5-6.3 (rear baffle removed)

The Cannon 50mm will be sold now that I have the Sony 55mm, but I was curious just how much that 5mm changed the field of view. I’ll likely sell the Cannon 20mm too as I haven’t been very impressed with it and adding Voigtlander 15mm removes its one advantage of being my widest lens. The Minolta 135mm is also potentially on the chopping block as noticeably inferior to the Minolta 100mm.

I rarely use longer lenses and already owned the Sony zoom from my NEX days, so the easy hack of removing the rear baffle to get better sensor coverage worked well for me. Due to vignetting from this hack I basically consider this a 210mm prime for the few times I need more reach. One day I may pick up the Sony 70-300mm f4.5-5.6 to replace this if I start needing longer lens more often.

How Far We’ve Come: 40 Years Of Processing Power

It’s amazing how far we’ve come in the speed of our computing. While the number of instructions a CPU can process in a second isn’t a perfect benchmark, as it doesn’t account for the efficiency of the instructions, it will do for our simple comparison. To put the difference the years have made into perspective: 1 second of processing by Intel’s (2011) i7 3960X would have taken the best 1985 personal computer over 4 hours 29 minutes!

Year Chip Millions of Instructions per Second
1985 Intel 386DX 11 MIPS at 33 MHz
1992 Intel 486DX 54 MIPS at 66 MHz
1996 Intel Pentium Pro 541 MIPS at 200 MHz
1999 Intel Pentium III 2,054 MIPS at 600 MHz
iPhone 4S ~5,000 MIPS
2003 Intel Pentium 4 9,726 MIPS at 3.2 GHz
iPhone 5S ~20,500 MIPS
iPhone 6 ~25,000 MIPS
2006 Intel Core 2 X6800 (2 core) 27,079 MIPS at 2.93 GHz
2006 Intel Core 2 QX6700 (4 core) 49,161 MIPS at 2.66 GHz
2008 Intel Core i7 920 (4 core) 82,300 MIPS at 2.66 GHz
2011 Intel Core i7 3960X (6 core) 177,730 MIPS at 3.33 GHz
2013 Intel Core i7 4770K (4 core) 133,740 MIPS at 3.9 GHz
2014 Intel Core i7 5960X (8 core) 238,310 MIPS at 3.0 GHz
2015 Intel Core i7 6700K (4 core) ~161,173 MIPS at 4.0 GHz

Note: Due to Apple’s obsession with secrecy the iPhone numbers are approximations based on custom utilities and not official. The i7 6700K was just announced last week so benchmarks aren’t available yet, but Intel claims it is 20% faster than the i7 4770K.

Drafts Recursive Email Actions For Appigo’s ToDo App

Drafts is a great app for automating tasks on iOS.  That is, saving a few screen taps and/or loading seconds here and there on tasks you do a lot to make yourself more efficient.  Drafts does this through “actions”.  More complicated recursive (looping) actions are harder to create but do more than simple actions.  That said here’s what my problem was:

emailaction

I love Appigo’s ToDo as my task manager.  It’s ability to share a (grocery) list with someone is worth price of the premium service alone ($20/yr).  However it only has very basic x callback urls for use in Drafts actions and while the email importing of tasks are intelligently parsed – the ones passed from Drafts are not.  Solution: use an email action. Just add your Appigo import email and any task options you want as the default.  In this case I have it set to use the HEB list and set the priority to none. Note: HEB is the name of my local grocer.

This works great for single tasks, but I often wander around the kitchen adding several food items at a time.  Solution: a recursive email action on each line of a draft.  Problem: no one tells you how to do this online. There are several examples of recursive actions with another app action and a draft action, but none using two drafts actions.  The reason is you can’t do an email action inside a custom action.  So first create email action to only use the first line and then add this code as a custom URL action. Note: make sure the name in the code matches the name of the email action you create.  This will email off each line separately to be parsed and added to your ToDo.

drafts://x-callback-url/create?text=[[title]]&action=HEB&x-success={{drafts://x-callback-url/create?text=[[body]]&action=ToDo%20HEB&afterSuccess=Delete&allowEmpty=NO}}

 

Many thanks to geekswithjuniors.com and theaxx.net for posting about their actions so I could figure out mine. UPDATE: theaxx has moved here: unapologetic.io

Fixing My Desktop’s Noise & Heat


When Intel released it’s iCore CPUs a few years ago, it was a huge step forward in processing power.  Reading up on how overclockable the the entry level i7 was, I ordered my own and built a new desktop around a new i7-920 and GeForce 285.  And it was awesome and powerful, but loud and hot in my office.  However, in 2009, noise and heat were expected if you wanted game on your PC, so a warm office was the norm.

Fast forward to 6 months ago when I replace my Tivo with a HTPC.  Because the HTPC would reside in the living room, I researched and purchased components that got good reviews for being quiet. That said, I never expected it to be dead silent.  Thanks to an ivy-bridge i5-3570K processor, a SSD, a quiet case, and a silent power supply it makes negligible noise.  One can hear the CPU fan if you put your head behind the HTPC but not otherwise, so for practical purposes it is a silent PC.  Which got me thinking about the aging beast in my office.

Armed with a Kill-A-Watt and a dream, I set out to fix my desktop’s heat and noise problem…CPU Cooler

 

At the end of 2010 I had upgraded my graphics card to the ATI HD6950 for two reasons.  First, it let me game on across 3 screens thanks to eyefinity support.  Second, it was quieter and had a lower power draw than the vacuum-cleaner-sounding GeForce 285 (OK that might be a slight overstatement).  That and the addition of an SSD as my main drive were the only changes to my still potent rig. But the silence of my HTPC raised my expectations and the constant whirring of fans became painfully more noticeable.

First, I attacked the CPU problem by replacing the stock cooler with a quieter Zalman one.  Playing around with my CPU overclock settings I discovered that the i7-920 overclocking to 3.2Ghz used 40 more watts of power and ran 11C hotter at idle than the stock 2.6Ghz. However, reducing the OC to 3.0Ghz only used 1 additional watt and only ran 1C hotter at idle. The lower OC was also 15C cooler at full load (84C vs 99C). This was a step in the right direction, but even with these savings my desktop was still pulling 189W idling with the monitors asleep (best case).  Compared to my HTPC that used only 43W idling and <60W while in use, I still had a way to go.

Unplugging all my case fans one by one help me located my next biggest offender: my power supply.  This was an easy fix since I had been very impressed with the KingWin Lazer Platinum Series 550 watt power supply that I had used in my HTPC.  It has two settings for it’s fan: extremely quiet and off.  Seriously.  It acts as a passive PSU in my HTPC.  I keep the fan on in my desktop since the thermal requirements are higher but wow what a difference.  More amazing than that is the efficiency of the thing.  Its 80 PLUS Platinum High Efficiency is no joke and it dropped my idle power draw to 168W.  That’s 21W of pure heat waste removed.

Next up was my graphics card.  In my research on GPU power draws I discovered an oddity: The HD6000 series didn’t truly idle with more than 1 monitor.  I verified this on my machine and found it was running at 50% of max clock with 2 or 3 monitors attached.  This meant it was running 10C hotter at idle (58C vs 48C) and using 40 extra watts to be attached to an extra monitor.  This was even if the monitor and signal were asleep.  Dropping back to a single monitor dropped my idle load down to 128W, but wasn’t a practical solution as extra monitors are a must for productivity.

Enter the ASUS GTX660 DC2O. The GTX660 was a sweet spot for power usage as it is the fully implemented version of NVidia’s mid-grade chip and ASUS made the quietest version of it according to reviews.  Also this card also offered a 20% boost in gaming performance even with that less power and noise.  The new GPU reduced my idle power draw to just 125W (with multiple monitors) another 41W savings.  The card also idles at just 39C and is inaudible (because of the the CPU fan) unless intensely gaming.  Even then the CPU fan is usually louder than it due it’s added load.  Gaming power draw fluctuates a lot but the GTX660 seems to use 15-35W less at the same setttings.

Results: So 3 upgrades and some tweaking later and my desktop is considerably quieter and it’s idle power draw has dropped from 229W to 125W.  Since at idle my computer is doing the same amount of work, all of that power difference was simply heat waste.  It takes approximately 39W of to raise 1,000 cubic feet of air 1 degree Fahrenheit.  My office is closer to 1300 cubic feet so that 104W translates into 2F cooler in my office.  That is substantial because that’s a 2F difference while my computer is sitting there doing nothing.  So when I walk into the office to start working I’m already better off and those differences continue through my now much cooler and quieter day.

 

How Far We’ve Come: 35 Years Of Processing Power

It’s amazing how far we’ve come in the speed of our computing. While MIPs isn’t a perfect benchmark, as it doesn’t account for the efficiency of the instructions, it will do for our simple comparison.   To put the difference the years have made into perspective: 1 second of processing by the current i7 champ would have taken the best 1985 computer over 4 hours 29 minutes!  Another fun comparison is that the new iPhone 4S clocks in at about 5,000 MIPs which is faster than a computer from 1999.

Year Chip Millions of Instructions Per Second
1985 Intel 386DX 11 MIPS at 33 MHz
1992 Intel 486DX 54 MIPS at 66 MHz
1996 Intel Pentium Pro 541 MIPS at 200 MHz
1999 Intel Pentium III 2,054 MIPS at 600 MHz
2003 Intel Pentium 4 9,726 MIPS at 3.2 GHz
2008 Intel Core i7 920 82,300 MIPS at 2.66 GHz
2011 Intel Core i7 3960X 177,730 MIPS at 3.33 GHz

Here’s another fun comparison: an old 8MB vs a new 8GB (8000MB) memory card:

.