A study on the suitability of 3D-printed models to analyse sounds of the whistling arrows

This paper examines the subject of whistling arrows, a type of military equipment that can produce a whistling sound through the use of horn or bone whistles. Despite their fascinating nature, whistling arrows have been a largely understudied topic within the medieval archery legacy due to the difficulty involved in their manufacture and testing. However, the growing availability of 3D printers has made it increasingly feasible for researchers to test and analyse archaeological artifacts. Accordingly, this study tests the hypothesis of using 3D models produced by 3D printers to overcome the challenges associated with manufacturing whistling arrows. Ten 3D whistle models based on three different typologies dating to the Turkic Khaganate period were printed and subsequently tested, with the resulting sounds recorded and analysed. Through this approach, the study seeks to contribute to a better understanding of the acoustic and mechanical properties of whistling arrows.

Liao Wanzhen has asserted that the majority of researchers consider arrow whistles and whistling arrows to be one and the same. However, careful analysis of archaeological findings reveals a distinction between these two types of arrows. While whistling arrows feature a pointed arrowhead designed for penetrating armour or injuring enemies, whistle arrows lack this feature. It is possible that arrow whistles were only used to produce signals, whereas whistling arrows had a dual purpose of signalling and causing harm (Wanzhen 2000). Peter Dekker, a leading expert in Chinese archery, also acknowledges this differentiation, although he suggests using the term 'whistle arrow' rather than 'arrow whistle' (Dekker 2013). We concur with Dekker's recommendation and will adopt this terminology in our article (Fig. 1).
The origins of whistling arrows can be traced back to the Shih Chi, a book written by Ssu-ma Chi'en, also known as the 'Great Historian of China'. This historical text includes a narrative describing the use of whistling arrows by Mo-tun, the leader of the Hsiung-nu people: Mo-tun had some arrows made that whistled in flight and used to drill his troops in shooting from horseback. 'Shoot whenever you see my whistling arrow strike!' he ordered, 'and anyone who fails to shoot will be cut down!' Then he went out hunting for birds and animals, and if any of his men failed to shoot at what he himself had shot at, he cut them down on the spot. After this, he shot a whistling arrow at one of his best horses. Some of his men hung back and did not dare shoot at the horse, whereupon Mo-tun at once executed them. A little later he took an arrow and shot at his favourite wife. Again, some of his men shrank back in terror and failed to discharge their arrows, and again he executed them on the spot. Finally, he went out hunting with his men and shot a whistling arrow at one of his father's finest horses. All his followers promptly discharged their arrows in the same direction, and Mo-tun knew that at last they could be trusted. Accompanying his father, the Shan-yü T'ou-man, on a hunting expedition, he shot a whistling arrow at his father and every one of his followers aimed their arrows in the same direction and shot the Shan-yü dead. Then Mo-tun executed his stepmother, his younger brother, and all the high officials of the nation who refused to take orders from him and set himself up as the new Shan-yü (Watson 1968, p. 161;Onat, Orsoy, & Ercilasun, 2015, p. 6).
Whistling arrows, like other types of arrowheads, can be made from a variety of materials such as horn, bone, wood, and iron. These arrows' function based on the same principles as wind instruments or whistles (Wanzhen 2000;McEwen & Elmy 1970, pp. 24−25). When air passes through the holes of the whistle, it produces a whistling sound. During the firing of an arrow, air friction occurs around it. As the air becomes trapped in the holes of the whistle, it enters the sound chamber and circulates, creating a loud whistling sound. Each whistling arrow is capable of producing a unique musical note due to the movement of air within the sound chamber. It is worth noting that this approach to the production of whistling arrows is supported by various scholars (McEwen & Elmy 1970, pp. 24-25).
Insufficient research has been conducted on whistling arrows as mentioned earlier. The primary factors contributing to this limitation are financial requirements and production processes similar to those of experimental archaeology (Eren et al. 2016, p. 122). In this regard, 3D printers have become a highly viable option since they are capable of producing any object at an affordable cost and in a short time (Eren et al. 2016, p. 122;Fragkos et al. 2018, p. 6). Furthermore, there are encouraging scientific studies on the production of custom wind instruments (Dabin et al. 2016, p. 288;Kantaros & Diegel 2018, p. 1516Savan & Simian 2014, pp. 540-541;Lorenzoni et al. 2013, p. 423). In an article, Amit Zoran compared a metal flute with one produced by a 3D printer and demonstrated that the frequency spectra of the B1 note of both instruments were very similar ( Fig. 2) (Zoran 2011, p. 385). Consequently, it is possible to predict that whistles for arrows produced through 3D printing will produce their own unique sound in musical notes.
The present study aims to investigate the hypothesis that 3D-printed whistle replicas for arrows will produce sounds equivalent in terms of musical notes. In order to test this hypothesis, we will conduct experiments by manufacturing 3D-printed whistles inspired by artifacts dating back to the period of Turkic Khaganate (Map 1), and subsequently, record and analyse the resulting sounds.

Modelling and printing the 3D whistle models for arrows
As documented by several studies, the majority of whistling arrow artifacts have been discovered in kurgans associated with Eurasian steppe cultures. Notably, a considerable number of such artifacts have been recovered from kurgans dating back to the Turkic Khaganate 1 period, which is regarded as a significant culture of the time. Specifically, over 65 whistle artifacts have been excavated from Turkic Khaganate kurgans in the Altai region (Kubarev 2005;Kubarev 2002;Kubarev 1995;Larin & Surazakov 1994;Mogil'nikov, 1990;Khudyakov et al. 1990;Kiryushin et al. 1998). These finds have preserved whistles in various forms, allowing their reproduction through 3D printing. For this study, whistle replicas have been selected from the Barburgazy-I, Kara-Koba-I, and Oroktoj excavations, all belonging to the Turkic Khaganate period (Fig. 3). The whistles from each excavation exhibit distinct forms; to avoid any confusion, the whistles from Barburgazy-I are referred to as 'egg type,' those from Kara-Koba-I as 'vase type,' and those from Oroktoj as 'flattened type' (Fig. 4) (Kubarev 2005;Mogil'nikov, 1990;Khudyakov et al. 1990).
In order to test the hypothesis that 3D-printed whistle replicas (Figs. 5 and 6)for arrows produce the same musical notes, we prepared ten different 3D digital models of whistles by varying the diameter and number of the whistling holes (Figs. 7,8,9,10,11,12,13,14,15,16,17,18,19,and 20). These models were drawn using FreeCAD 0.19 software. Ultimaker's Cura software was used as a slicer to prepare the G-code file. ABG brand's 1.75 mm PLA type filament was selected for printing, and the printing process Map 1 Turkic Khaganate at the end of the sixteenth century was carried out using an Ender 3 v2 3D printer, with a printing time of 28-35 min per whistle model. The printing settings for Cura were configured based on Table 1.

Testing setup and recording of the sounds of the 3D-printed whistles for arrows
In accordance with various studies, it has been noted that experimental archaeology tests are not always entirely reliable, and experiments that involve human participants may affect the reliability of results that are derived from archaeological data (Eren et al. 2016, pp. 108-109). In order to address this issue, we endeavoured to establish a testing setup that would minimize the possibility of errors that may be attributable to human involvement.
To generate the necessary airflow for producing the sound of the whistles, we employed an air-pumping engine. A cylindrical pipe that was long enough to accommodate the arrow shaft and whistles was connected to the air engine (Fig. 5). Subsequently, the engine was activated and the air suction force in the pipe was quantified using the UNI-T's Arrow whistle types, which were printed as 3D models to analyse UT363 anemometer, which yielded a measurement of 29.8 m/s. Consequently, a standardized testing setup was implemented to ensure consistency across all the different whistle designs that were to be evaluated.
The attachment of each whistle to the arrow shaft was conducted in a fixed manner, after which they were inserted into the air intake tube. Sound recordings were then obtained with a Logitech H340 microphone. Each recording was subsequently analysed using Cubase 5 software, and detailed analyses of frequency and note values were conducted, resulting in the creation of spectra (as illustrated in Figs. 21,22,23,24,25,26,27,28,29,and 30).

Results
As anticipated, each tested whistle had a distinct acoustic profile. However, in the tests conducted, samples No. 3 and No. 4 did not provide sufficient sound data for analysis ( Fig. 23 and 24). The sound frequencies were measured in the range of 2800 to 4000 Hz for the whistles that were successfully analysed. The corresponding note values for these frequencies ranged from F6 to B6. Among the analysed samples, No. 10 had the lowest pitch, producing the note F6, while the highest notes were produced by No. 2 and No. 8, which gave the note B6 (Table 2).
When an evaluation was made on the whistling holes and their counts, it was seen that the whistles with two holes had a lower sound than the whistles with three holes. The overall effect of the diameter of the whistling holes on sound could not be determined. This was because the sound became higher in the larger holes in the vase-type whistles, but lower in the flattened whistles (Table 2).

Discussions
The utilization of animal-based materials, such as bone, antler, and horn to produce whistles used in whistling arrows poses a significant challenge. According to Wanzhen (2000) and McEwen and Elmy (1970, pp. 24-25), these natural materials are challenging to acquire and necessitate a high level of expertise in processing. Researchers must dedicate extensive periods to master these materials, which are scarce and difficult to obtain. However, a slight mistake can cause them to break and become unsuitable for testing, as highlighted by Eren et al. (2016, p. 106). These limitations associated with the manufacturing and testing process could hinder the investigation of whistling arrows and whistles.
One of the issues that should be taken into consideration is the formation processes of the design differences in the whistle finds. As is well known, there is a rich body of literature on how and why technology developed (Basalla 2013;Bronowski;Dafoe 2015;Smith 1977;McClellan and Dorn 2016;Edinborough 2005). Since this subject is not directly related to our study, we will not go into detail. However, it is likely that the design features of some of the whistles we have discussed directly affect the phenomenon of sound frequency, which is the focus of our article. The most striking among these features are the hole diameters of the whistles. As can be seen in Table 2, the hole diameters of the whistles emerge as factors that affect the resulting notes. However, there is no certainty as to whether these holes are deliberately produced in certain dimensions. The person who produced the whistle may have created these hole diameters consciously or unconsciously. On the other hand, the fact that the whistles were produced in typologically different shapes seems to be the result of a special effort. In this case, the possibility that the effect of typological differences on the whistle sound is a conscious selection comes to mind. However, there is an issue that negatively affects this argument. As is known, the kurgans associated with the steppe peoples of Eurasia belong to the ruling class and nobility, which we can define as the upper stratum   of the society (Khudyakov 1986;Kubarev G. V., 2005;Gorbunov 2003). Based on this fact, it can be assumed that the whistles produced in different designs obtained from the kurgans were produced as an indicator of social status within a patronage relationship. In this case, the frequency differences arising from typological differences would not go beyond a coincidental set of unique values. However, as mentioned above, this issue does not directly address the purpose of our study and the comments we will make may remain superficial. Therefore, we believe that researchers, especially those who specialize in the functioning of technology, should re-evaluate the data we have obtained within the scope of the article, thus creating a different perspective on the technological evolution of the cultures in the region.
The use of 3D printers is a viable option for researchers interested in studying whistling arrows due to their low cost and high precision in production. As previously mentioned, whistles used in medieval arrows were primarily crafted from organic materials, such as horn or bone (Wanzhen 2000;McEwen & Elmy 1970, pp. 24-25). Utilizing a reusable material, such as polylactic acid (PLA), instead of these organic materials is advantageous in various ways. The researcher can produce whistle designs tailored to their research needs in a computer environment and print them in large quantities and on demand. This efficiency in the production process facilitates faster testing and allows researchers to concentrate on testing procedures. Furthermore, horns or antlers from endangered animals, such as ibexes, are required to produce these materials (Rovero, et al. 2020, p. 118;Joshi, et al. 2020, p. 1;Han et al. 2021, p. 1). Nonetheless, 3D-printed whistles eliminate the need for horns and antlers obtained from these animals, particularly for sound comparison tests.
In the present study, it has been demonstrated that 3D printers can be utilized to produce whistles of a desired design and size. This capability provides an opportunity to produce and compare more complex and numerous samples. It is widely recognized that increasing the number of tests enhances the reliability of the results obtained. Therefore, the use of 3D printers can contribute to the improvement of research outcomes by allowing the production of a greater number of whistle samples for testing purposes.
In this preliminary investigation, intriguing findings have been revealed concerning various aspects of whistling arrows and whistles. It was discovered that whistles containing two holes produce lower sounds in comparison to those with three holes. This observation indicates that each design elicits a distinct sound, and these designs are distinguishable from one another. Thus, it is plausible that whistling arrows could have been used to issue commands to soldiers, as described in Motun's narrative (Watson 1968, p. 161;Onat et al. 2015, p. 6). Furthermore, it may be possible for horses to be trained to recognize various commands, such as attack and retreat, based on different whistle sounds. Investigation into this topic would enhance understanding of the battle tactics employed by nomadic armies.

Conclusions
This article stems from the premise that instruments created with 3D printers can replicate the pitch of their original models. It has conclusively demonstrated that the whistles    utilized in arrows can be produced with 3D printers and produce sound in specific pitches. Nevertheless, this study did not compare whistles of identical designs made from varying materials. This comparative analysis should be undertaken in subsequent research endeavours. Whistles and whistling arrows, like other military equipment, are a part of technological development. For this reason, the data we obtained in this study should be re-evaluated in the context of the evolution of technology. Since, as was mentioned in the discussion section, it is not known whether some of the features that cause the whistles to resonate with different musical frequencies were designed randomly or deliberately. This situation should be addressed by researchers who have studied directly related to the evolution of technology and should be analysed in further studies.
The present study sheds light on the effect of specific physical properties of whistles on the resulting sound. Although the number and diameter of the holes have been identified as factors affecting the whistling sound, further research is required to thoroughly examine this issue. The dimensions, design, number of holes, and diameter of whistles produced via 3D printing need to be meticulously compared in future research. Through this comparison, the impact of design changes on the sound produced can be better formulated. However, conducting such a comparison requires a large-scale scientific project since hundreds of 3D whistle models would need to be produced and thousands of hours of testing conducted.
As previously mentioned, the 3D models utilized in this study were printed using the Ender 3 v2 printer by Creality. In future studies, it would be prudent to employ diverse 3D printer models and technologies. By doing so, the findings of subsequent research can be rendered more dependable and consistent.
Author contribution R.E.C. and M.A. wrote the main manuscript. R.E.C. prepared test setups and recorded audio profiles of whistles, and prepared result tables and figures. M.A. read Russian literature and contributed literature view. All authors reviewed the manuscript.
Funding This study is supported by Ege University Scientific Research Projects Coordination Unit. Project Number: 22507.

Data Availability
The data obtained in the present study have been illustrated in Figs. 21-30 and Table 2, and are available for use in scientific research upon reasonable request from the corresponding author with appropriate citation. The data will be provided in a usable format for analysis and will be accompanied by a detailed data dictionary to facilitate interpretation. Requests for access to the data will be assessed on a case-by-case basis, considering ethical, legal, and privacy issues related to the data. Certain limitations may restrict data availability, and approval by relevant institutional review boards or data access committees may be necessary. Interested parties can contact the corresponding author at recep.efe.coban@ege.edu.tr to request access to the data.

Competing interests
The authors declare no competing interests.

Conflict of interest
We confirm that neither we nor any of our relatives nor any business with which we are associated have any personal or business interest in or potential for personal gain from any of the organizations or projects linked to FreeCAD, Ultimaker, Creality, Logitech, ABG, UNI-T, and Cubase firms.