Some guitar makers at a lutherie school in Belgium wondered if it would be possible to build acoustic guitars from local wood that would sound as good as guitars made from the well-known tropical woods. This straightforward query was the starting point for a broad and long-term international research project.

An Overview

LGRP Phase 1
Between 2012 and 2014, the Leonardo Guitar Research Project conducted comparative playing and listening tests, both blind and non-blind, on the sound perception and preferences of guitars made from tropical wood and guitars made from non-tropical wood. The only exception was the soundboard which used the same wood and bracing pattern for both tropical and non-tropical guitars. In total, 15 classical guitars were played by 3 experienced guitarists and assessed by themselves and by listeners both in room and under concert hall conditions in two locations (Belgium, Finland). / See Research Report #1

LGRP Phase 2
Between 2014 and 2017, we conducted blind and non-blind comparative playing and listening tests with 20 classical and 24 acoustic steel-string guitars made of either tropical or non-tropical woods (again, except for the soundboard which remained the same for tropical and non-tropical guitars), built by 22 luthiers and tested by 20 professional guitarists over 3 countries. The tests were first conducted in Belgium and then repeated in Finland and the United Kingdom, with different guitars built in each country using strictly the same guitar design and controlled building process, the same research design, methodology, and procedures of data recording and processing. / See Research Report #2
      At the end of 2017, we organized an additional blind online listening test with the same 16 classical guitars as in the previous test, built in Belgium, in which 226 assessors participated. / See Online Listening Test, Report#3

Collaboration with academic researchers
Apart from the comparisons of different tonewoods we conducted, at the end of 2017, tests to investigate the long-term stability of alternative wood species for classical guitar necks. This study inspired academic researchers at the University of Gent Laboratory for Wood Technology (UGent-Woodlab) to conduct a similar experiment using the same LGRP setup and measurement method plus a new setup method. These 2 setups were compared in terms of workability and reliability.
      The UGent paper, with Bingrui Chen as main author and LGRP as co-author among others, entitled “Monitoring the long-term stability of classical guitar necks: evaluating the potential of European-grown wood species in musical instrument manufacturing” was published in “Springer Nature, Wood Science and Technology”, June 2025 / More info, see Related Research.

LGRP comparative tests / Tropical versus non-tropical: some results
All comparative tests between guitars built with common tropical woods and alternative non-tropical woods, in which all comparisons were made with a same guitar model and with European spruce for the soundboards, show that, in blind conditions, players and listeners cannot distinguish between the two types of guitars at better than chance levels.
      The preferences for one type over another are equivalent and did not result in significant differences at a 5% significance level from the average, applying to all individual assessments compared with each other per country. All the results of the blind tests in the various studies combined show a preference ratio of 50/50 tropical versus non-tropical for both classical and steel-string guitars.

In the non-blind experiments, on the other hand, the prevailing bias in favour of tropical woods was strongly confirmed, especially in our first study (2013/2014) where results show a marked fallback in preference of some 50% for the non-tropical wood guitars, indicating that sound perception can be strongly influenced by visually transmitted information such as the 'look' of an instrument (dark versus blond woods !), or the preconceptions surrounding ‘good’ and ‘bad’ tonewoods that their recognition allows.
      A few years later, during the "Phase 2" studies, most assessors were aware of the LGR-Project and tended to sympathize with it. That is probably the reason for the less pronounced bias in favor of tropical guitars, although that preference was still present but to a lesser extent.

Comparative tests, inconsistencies and psychoacoustics.
To draw general conclusions about the studies and to understand the observations, we need to focus on some rather complex but intriguing psychoacoustic phenomena that occurred in the blind tests. We must then consider to what extent these phenomena can be taken into account when interpreting the raw results.

Test results, additional observations and player comments indicate a high degree of subjectivity and inconsistency in perceptions. These vary not only from one assessor to the other, but often within the same assessor as perceptions and preferences sometimes changed significantly within a single test session, from one assessment to the next, about physically the same guitar, while the player was under the illusion that two different guitars were being tested in a group of five while there were only four. This double test for the same guitar is further referred to as "test-retest".

    Depending on the calculation method and based on solely the rating data for the test-retest guitar, the degree of consistency (when the scores are corresponding for the supposed two guitars) varies from an average of 27% to 41% within the same blind session for the same player, with significantly less consistency across two different sessions with the same group of guitars.
    However, we need to look at these numbers in a broader context: From the comments that players could provide during the assessments, we could clearly deduce that when they gave similar and therefore consistent-looking scores for what they considered different guitars (but which was physically the same), they did so for different reasons regarding different perceived sound characteristics. This shows that similar scores do not necessarily correspond to similar perceptions; usually not. This finding puts the calculation of consistency based on preference ratings into perspective.
    Furthermore, we must take into account the fact that the scores of most players differed only within a small margin of 3 or 4 points on a scale of 1 to 10, which creates a high potential for fairly coincidental score similarities. As a result, the physically different guitars and the guitar used for the test-retest often indeed received the same scores (while the two test-retest scores themselves were identical in only 27% of cases). Either the perceptions were accurate and the preferences for the different guitars were relatively close, or the differences between the guitars were not very clear to the players, which may have led to a certain degree of randomness, which then seems to be confirmed by the inconsistent test-retest results.
    In general, we can conclude that consistency in perception was very low to almost non-existent (although this is difficult to quantify, given a certain degree of probability) and that this was most likely because the guitars were very difficult to distinguish from each other and not because the musicians had ‘bad hearing’.

This so-called 'bad hearing' phenomena requires more explanation: after all, when guitar sounds are very similar or even identical, as in the test-retest experiment, the mind of the blinded assessor has a strong tendency to fill in the supposed differences itself. As a result, through a cycle of self-confirming expectations and biases, it unconsciously creates new and sometimes very different soundscapes for the same guitar. Confirmation bias is a powerful and treacherous force that explains a lot within the context of our experiments. By selectively filtering facts and manipulating analysis to match expectations and beliefs, it enables people to mislead themselves.
      The brain doesn't just register sounds; to some extent and depending on several factors, it decodes, interprets and modulates the perceived audio material, with the ability to focus on certain sonic features and/or fade out others, comparable to what one could technically do with an analog or digital equalizer, but here using the ear-brain system, while one is not, and cannot even be, consciously aware of this process! ¹
      Paradoxically, professional musicians, due to their more developed skills, experience, knowledge and thus their advanced ability for 'mental imagery'¹ regarding sounds and music making, seem to be able to perform these processes 'better' and 'more creatively' than non-musicians or even amateur musicians (according to academic studies investigating auditory neurocognition in musicians ¹ ² ³ ).
    During a total of 110 blind test sessions that were conducted, none of the 20 players noticed that, per session, two guitars were actually identical, even when tested one after the other. The auditory illusion was complete. The clear presence of that illusion confirms the existence of cerebral mechanisms that influence the reception and perception of the same sound material during the decoding and assessment process, and in our tests, mostly with a different result for each assessment for the same instrument. Believing it to be a different guitar, the assessors' brain was reset to zero between the first evaluation and the second evaluation of that same guitar in a same session, allowing a different mental evaluation process to be started and completed. This was noticed in almost every player.
    These seemingly surprising brain mechanisms are well known in psychoacoustics and neuropsychology and obviously cannot simply be considered as 'abnormal', on the contrary. On the other hand, we must come to the conclusion that the human ear-brain system is not always a reliable measuring tool.

Indeed, "not always" : To gain more insight and to empirically investigate this phenomenon of shifting perceptions (which were already strongly evident in our 'Phase 1' experiments), and also to check our methodology and experimental design, we integrated (during Phase 2, Belgium) the so-called 'strange guitar experiment': an additional blind test where a guitar of a different model and with a different bracing system was placed between the project guitars. The research question was: 'can players distinguish that one guitar from the other guitars that previously were difficult to distinguish from each other?' All players spontaneously reported in their comments, which they were able to provide during testing, that they could now perceive clearly different and definable tonal qualities in this particular guitar compared to the others. Unlike our comparative experiments with guitars of the same design and with a same soundboard, the players' perceptions remained now consistent throughout the entire testing session.
      Identifying the sound did not require the activation of creative mental 'imagery' processes, precisely because, in this experiment, the tonal qualities, compared to the other guitars, were clearly provided by the nature of the instrument itself.

Regarding our comparative studies, we can conclude that the degree of (in)consistency in players' ratings and the underlying perceptions can also provide an indication of the extent to which the guitars are distinguishable from each other, rather than only looking at the raw numerical and statistically processed rating data.
The common occurring inconsistent perceptions indeed seem to confirm the indistinguishability of the different guitars and therefore the consistency of their tonal characteristics. If the sounds of the different guitars had varied more, the assessments would likely have been more consistent.

All results together lead to a twofold general conclusion:
1) As long as the guitar design and the material and quality of the soundboard remains the same for both types, alternative non-tropical woods can be used to build guitars that sound as good as those made from the commonly used exotic woods.

2) Since our blind playing and listening tests revealed no significant difference between the two guitar types, this suggests that it is primarily the model, the top wood, the bridge design and the bracing system (as the only constant factors in our experiments) that determine the sound, and much less the back and sides, contrary to popular belief. If this is the case, it obviously explains why it is difficult to distinguish between the two guitar types or between guitars with backs and sides made of a specific wood, regardless of whether that wood is tropical or non-tropical.
      This brings us to questions and conclusions that go beyond the tropical versus non-tropical issue; the two-type concept was the starting point of the LGRP investigation, but, since it now has been confirmed that in blind tests one type cannot be significantly distinguished from another, this classification, based on widely held but incorrect beliefs, appears to be purely artificial and essentially no longer valid. On the other hand, as long as the guitar community continues to believe that acoustic guitars sound better with exotic woods, this kind of research remains relevant.

Note: A similar study led by Lancaster University (UK), comparing "traditionally prized, expensive, and rare woods" with "cheaper and more readily available woods" came to a comparable conclusion: "The poor ability of guitarists to discriminate under blinded conditions between guitars with backs and sides made of different woods suggests that back wood has only a marginal impact on the sound of an acoustic guitar."  The study was published in 'The Journal of the Acoustical  Society of America', December 2018. (link to the paper).

Discussion and questions:
While preference scores consistently hover around a 50/50 ratio, creating the illusion that all guitars sound more or less the same, there are, in fact, subtle variations in sound, timbre, and volume, not only between tropical and non-tropical guitars but also within the same type, whether or not they are tropical.

The soundboards of all guitars are as similar as possible, but because they are composed of organic material, they are not "perfectly" identical. For instance, a soundboard may be slightly denser or less quarter-sawn than another, which can lead to variations in stiffness and other characteristics that affect the sound.

The questions we currently have no answers to, are as follows:
● Are the (small) differences in sound between the guitars used for the tests, apart from the fact that they belong to the tropical or non-tropical group, the result of using different types of wood for back and sides, or are they the result of inevitable differences in the different spruce tops? (or in both?)
● Which factors play a role, and to what extent? (this question also applies to other elements within the instrument)
● Could luthiers build differently (techniques, materials) and in a more targeted way if they knew the answer?
These and many related questions offer fascinating opportunities for further research.

Further research
One way to investigate the above questions could be to build a series of guitars with different types of wood for the back and sides and an identical top for all guitars made from a material that can be controlled, such as carbon-based composite materials or more ecological alternatives such as flax fiber, with which various builders, both for stringed instruments and guitars, have already achieved promising results.
      This is just one example of an experimental design using constant and variable elements for empirical comparative research to study acoustic properties of wood within the entirety of an instrument. Many variations are possible...