Saturday, May 16, 2020

’Seeing’ and the Camera

Enshrining Distortions - 'Forged in Fire'

Is what you ’see’ on a YouTube video even remotely what you would see if you were physically at the demonstration?

One of the significant problems in photographing high temperatures is that modern digital imaging systems simply do not respond to visible light the same way the human eye does. This is due primarily because of the way CCD or CMOS sensors, the heart of digital cameras, specifically react to light. This can be confounded by how light metering systems tend to average across an image. This second aspect particularly a problem inside the typical lower room lighting found in most workshops. (1)

For comparison, I looked at a large number of ‘heat by colour’ charts available via an internet search. If you did this, you would find a huge number of potential charts - many of which do not agree, at least in detail. The source web sites range from hobby blacksmith, through to industrial suppliers, and engineering data. I chose to include one from a blog site, primarily because it appeared fairly close to what I personally observe in the forge. (2)
I have chosen to alter that chart slightly, however. It appeared to me on the original source that they had flipped ‘Light Red’ and ‘Orange’ - which I reversed on the version shown below. Even with this adjustment, you can see on this chart that there is almost no difference indicated between ‘Bright Orange Red’ and ‘Light Red’.
On consultation with David Robertson, he suggested the second chart included here. I agree with him that this actually more closely matches what I perceive personally. I have attempted to adjust the two charts so the slots more or less align, but as you see, they do use different ‘slots’ in terms of numbers. The second chart also only provides temperatures in Fahrenheit.

I have also inserted a blue line at the ‘critical’ temperature. ( 3 )
The green line marks the reading off the pyrometer during these tests. Again, I found the ‘Basic Knifemaking’ chart was actually closer to the ‘bright orange’ I perceived for the metal as it was removed from the forge.
One of the big problems with these printed charts is the difference between colour and brightness. That last aspect can not be realistically reproduced with a printed chart.

The images used for this article were shot at the Wareham Forge in early May, using my ‘home built’ two burner gas forge as the heat source. ( 4 ) I freely admit that there are some problems with this equipment, most notably is that at best it will provide working temperatures into the ‘Yellow’ range (visually and as recorded at one point during during this exercise = 1150 C)). I do not consider this forge hot enough to allow forge welding. The images were made at the end of a working session lasting about 2 hours (the forge was fully heated).

Temperatures were recorded using a Omega HH12B digital pyrometer with a type K
thermocouple (rated to 1260 C, for short exposure). This is a fairly good quality piece of equipment, not likely in the hands of most blacksmiths.

I had done an initial set of measurements and photographs, which I was not entirely happy with. The pyrometer readings especially did not match what I was perceiving, typically running about 100 C higher than my experience suggested. The contributions of David and my regular iron smelting partner Neil Peterson were especially helpful here. David reminded me that the ‘burner temperature inside the forge would not be the same as the functional temperature that a piece of metal would climb to. Neil suggested placing the thermocouple inside a block of metal, using drilled out block or piece of pipe.
For the second round of measurements and images, I did just that. I used a piece of 3/8 ID mild steel (schedule 40) pipe, about 8 inches long, and plugged at the far end. This encased the thermocouple for it’s entire length inside the forge. Temperature readings were on going, there was less variation between door closed and open with the pipe covering.
The average reading over the test was 1040 C

I had four different pieces of digital camera equipment available. None of these are new - so should not be considered the current ‘state of the art’. They may however well represent the kind of equipment available by the average user.

1) Digital SLR : Olympus E-300 / 2008
2) Digital ‘Pocket’ (Viewscreen) : Panasonic Lumex DMC-T51 / 2009
3) Cell Phone : Aceatel One Touch A460T/ 2016
4) Digital Camcorder : HP t250 / 2012
- Although all these allow for video recording, the images were made using the ‘still
photograph’ method (including the camcorder).
- The cameras were set using ’basic photo’ setting, automatic focus and exposure. (This especially important for comparison from the Olympus, which had the widest range of possible adjustments.)
- All images taken with ‘available light’, with exposure times variable.
(This explains why many of the images are not very crisp!)

The images were shot on three different working days. I could not get a full series on day one, because I forgot to check the battery power on all of the units (!). Day one and day two used the thermocouple simply laying inside the forge. Day three, the comparison images seen here, used the pipe sheath over the thermocouple.

The first image shows the general set up, this image shot day one - with the Olympus SLR.

This image has exposure determined by the overall low light level in the workshop,
so you can see that the bright interior of the forge is brightly washed out. I have altered the contrast just around the pyrometer screen, so the readout is more visible. Otherwise there has been no post processing of the majority of these photographs, other than reducing the size of the original images. The exception is the cell phone image of the forge interior, which had the red and yellow colour values increased to correct an extremely ‘purple’ tone to the source image. (No idea why this occurs with the cell phone camera?)

The first set of images from the various cameras were framed to include just an interior view of the forge, with a pieces of mild steel bar inside.
What this illustrates is the colour reproduction and ‘washing out’ problems.

Olympus SLR
Lumex Pocket
Acetel Phone (altered to render orange tones)
HP Camcorder (still image)
The second image set are of a 2 x 2 bar cut to a 2 ½ long block. This was set on the anvil fromthe forge, and the image taken quickly. An attempt was made to frame the same view.
What this illustrates is the general ‘bright against dark background’ problem.

Olympus SLR
Lumex Pocket
Acetel Phone (unaltered)
HP Camcorder (still image)
A last set of images was taken of that same block, as it cooled on the anvil, all using theOlympus SLR. The thermocouple was set inside a 1/4 inch diameter hole that had been drilled about 1 inch into the block. These images are framed to include the pyrometer reading at that instant. An attempt was made to capture one image roughly every 50 degrees as it cooled.
You may also note the difference between the image above with this camera (shot at roughly 1030 C) and the first image seen below.

725 - Critical Temperature

The Conclusions :

- It is clear that the various digital photography equipment records the identical situation quite differently.
- It is obvious that both colour and brightness are not accurately recorded.
- Images are most likely to suggest higher temperatures than what were actually present at that instant.
- Video camcorders and most especially cell phones have by far the greatest distortion in both colour and brightness.
- The closer the equipment is to single purpose photography - the more realistic the image colour is.

It should be noted for this set is that I felt overall the Olympus SLR, as a ‘real’ camera, did make the most accurate duplication of what my eyes perceived.
- Complex SLR cameras also have the ability to correct ‘white balance’, which with some experimentation should allow for more accurate colour representation.
- Also individual still images could also be processed after the fact by some software such as Photoshop to better reflect reality.
- Clearly professional filmmakers would also have better skills at accurate representation.

The single most important warning is that generally, what you see on YouTube - is most certainly NOT what you would see if you had been physically present.

One important related caution comes from what I have seen over and over.
You will ‘see’ on video people working steels well down into the low reds. If you realize that the camera is typically showing colour at least one ‘temperature grade’ higher than what is actually the actual case? That means what you are seeing is people actually working their metals well down below the critical range!
Two important considerations :
A) As all the working blacksmiths well know, steel, when at the orange heat, is certainly far more plastic than it might be at a dull red. The same hammer stroke just does far more shaping. Why are you working so hard?
B) If you work the metal below critical? Effectively you are creating microscopic cracks. Fine, you can get away with this when making a fire poker from mild steel.
Bladesmithing? Almost certain to produce catastrophic failure when the finished blade is subjected to the stress of quenching. (How often do you see just this whole thing played out on ‘Forged in Fire’??)

( 1 ) Neil Peterson, who contributed some ideas to this project, has struggled for years with attempting to take accurate images of the inside of our iron smelting furnaces. Both of these problems have made accurate recording of the temperatures involved, into the 1250 C range, especially difficult. Over the last 15 years, he has invested considerable time and research into those problems.

( 2)

( 3 ) For mild steel, the critical temperature is 725 C. This is technically the correct
minimum forging temperature. The critical temperature does vary with carbon content, and with alloy steels.
Minimum recommended industrial forging temperature for 1018 mild steel is 900 C

( 4 ) I had purchased this forge at Quad State a number of years back. The concept was that the air supplied to the burners would be pre-heated by running through a rear mounted exhaust vent - so improving efficiency and thus operating temperatures. In practice, the sharp bend combined with the flattening of the pipes ended up sharply reducing the amount of air available. ‘I never could get it to work’ was the maker’s comment. (I ended up paying less for the forge than the cost of the Kwool board insulation!).

It did need a fair amount of modification to become effective. The biggest new
element was the inclusion of a small fan (scrounged from an air hockey table at the dump). This allowed the correct amount of air forced towards the burners.

Originally prepared for the 'Special Supplement' May 2020 of the Iron Trillium.

1 comment:

Henry Troup said...

Most digital sensors will pick up some amount of IR. I demonstrate this with any IR remote control. But, due to variant materials in lenses, and the fact that no one typically worries about IR sensitivity, you may get random distortion of colour rendition - might explain the purple of the cell phone camera.

IR cut filters are available for some devices, you could try one on the SLR.


February 15 - May 15, 2012 : Supported by a Crafts Projects - Creation and Development Grant

COPYRIGHT NOTICE - All posted text and images @ Darrell Markewitz.
No duplication, in whole or in part, is permitted without the author's expressed written permission.
For a detailed copyright statement : go HERE