New guidance. In providing a client with a specification for historic building survey I had a reminder that Historic England has published new guidance on drawing so I took the time to read it through. It is a guide to reconstruction drawing “in support of a written description” it has no bearing on conservation mapping or condition recording other diagramming historic context. In coming to this conclusion I thought I’d share the reasons drawing is important in the conservation record and why we need drawings even though it has been suggested to me it’s time to ‘unlearn line’ and embrace other forms of information transmission.
A site sketch is a selective record of observations and reflects the personal ‘filters’ of the draughtsperson. Unless it is supported by measurement and photography its value to the conservation team is limited. As an amplifier of key details a sketch like this is a winner, it’s showing details that are hard to see in photographs and difficult to decipher from a point cloud or SFM model. Drawing by Alan T Adams.
Understanding differentiated and undifferentiated data is important here. A drawing is highly differentiated, a processs of imformation filtering or selection has taken place to define the subject by lines on the plot. The stage at which differentiation occurs in the recording process has an impact on the quality of the record. There are 2 ‘paths’ here: selection of line at capture (such as in preparation of the sketch or placing a line by TST) or by post capture analysis (tracing lines off a point cloud or photogrammetric product).The first is heavy on site time and the second requires near total converage to be be confident all the required details are captured. A choice has to be made and it’s not simply a matter of saying scanning or photogrammetry will do this or TST will do that: time, equipment, and software rental costs are tied up here.
Drawings have many functions. The conclusion of Historic England’s new guidance on building recording makes a strong case for drawing and measuring by hand. It’s a bit strange to single out a technique in this way as the effectiveness of any recording method is dependent on meeting a specific information need, not its method. How the data is captured in nowhere near as important as what is captured. Provided a repeatable method is used the process of information selection is the most significant driver in documentation outcome. Drawing is a good discipline for developing information selection skills but a field sketch has to work very hard to perform as a survey drawing for much more than a 3 cell plan, these days few clients can afford to pay a draughtsman the hours.
“Drawing and measuring by hand remain useful tools for recording historic buildings.They are versatile methods, requiring close observation and physical contact with the subject, which can lead to additional discoveries and deepen understanding while conducting site work. They can also be combined with digital survey to produce highly accurate measured drawings.
Used on their own, they can provide illustrative material ranging from basic block diagrams to full measured surveys with the accuracy necessary to support written accounts of building evolution. Drawings produced in this way can also be used as the basis for more accurate interpretive cutaway and reconstruction drawings.”
The method of information capture will infuence the outcome but this is not the biggest constraint on the utility of drawings as historic record by any means.
‘As proposed’ can be missleading: this suburban dwelling proposed in 1930 may never have been built even though permission was granted and a detailed drawing is on record.
Knowing the intent of a drawing will make the constraints on its precision and accuracy clear. The myriad of possible functions a drawing may serve can lead us to believe any number of possible performances- it may be interpretive, diagrammatic, diagnostic, analytical, or conjectural.
Whenever I am commissioned to undertake measured survey work I find it useful to recall the basics of the process:
Know your scale: be clear about the function of the survey in terms of the 3 key tasks of measurement, selection & presentation
No action without control. A framework of provably measured geometry (baseline, grid, traverse) must be in place before measuring detail.
Work from the whole to the part. Solve the big shapes first.
Don’t confuse accuracy and precision.
Understand the importance of scale.It seems obvious but a survey at 1:20 scale will cost you 3 to 5 times more than one at 1:50 scale because at the larger scale there is much more to be seen, measured and recorded.
Measuring up details and plotting them out at 1:5 is a slow and costly business, at this scale careful inspection is required and a sketch is often the best way taking notes. Agreeing on an appropriate scale for the work is often a starting point for both surveyor and client.
Knowing how things fit together and where to look for the evidence of pattern, type and style is something experience will bring.
There is a match between method and drawing type but the value of site drawing is in understanding diagnostic detail, Alan Adams proposes this distribution of technique:
But practice tells me the process of getting line drawings done from the various data sources looks more like this:
Preparing plans of anything but the most simple building is best done with a TST, I have failed to get measurements to work out so many times I rarely work up plans measured by rod and tape now. Similarly measuring elevations is best done by TST, getting ridge profiles and chimney stacks by other means is tedious.
Getting consistently square on measurement by rod and tape is the bane of the process.
Selecting the detail for measured drawing and avoiding missaplication of the technique should use a more ecconomic means to capture wider area data rather than stretching the limit of measured drawing performance by spending a lot of time capturing low precision data.
Drawing is costly, photography cheap. The value of the photograph cannot be underestimated but, by differentiation of the subject into lines, drawing provides a base for markup and direction of works; it shows location, form and design intent but nothing of condition or material.
Precision & accuracy. Precision is the measurable performance of data, accuracy is the correctness with which the data describes the object. A laser scan point cloud is a very precise data set but its accuracy to record a given feature may fail if the point density is too low or the feature is not in the line of sight of the scan. Clients often feel they want ‘the most accurate’ survey and insist on costly scanning in the belief the high precision of scan data will ensure great accuracy when in reality the accuracy may depend on where the scanner is placed more than its performance.
Image: Leica AG
Some methods are more ‘line friendly’ than others. Feature extraction from point clouds is a slow and software heavy exercise, even more so from SFM. Traditional deliverables have moved on (CAD is now the norm) but lines are still required and drawing sheet equivalence remains the quality benchmark.
A TST trace, a sketch and a handful of key measurements is all it takes to complete the hidden detail in this section.
The power of a drawing is its ability to communicate, preparing a site drawing is costly in terms of time but is uniquely valuable in conveying how things work, it’s weakness is its subjectivity, but this is also its strength: it will show key information based on the inspection and understanding of the draughtsperson. Unlike other methods of record it is reliant on the skills of the observer but, when combined with a metric framework, drawings define the quality of the record.