In a 2017 blind listening test of three new violins against three by Stradivari, audiences in Paris and New York guessed old versus new correctly in just 122 of 273 tries (44.7%), and they rated the new instruments as projecting better. The wood tells a similar story of overlap: spruce top plates from three countries ranged from 0.32 to 0.55 g/cm3 in density, a spread wide enough to swallow any “Cremonese” signature. Even so, chemistry finds real differences, such as aluminium levels of 1,200 and 2,900 ppm in wood from Andrea Guarneri’s grandson against under 50 ppm in Stradivari’s, and a 2026 tree-ring study traced Stradivari’s spruce to the Eastern Alps using 314 ring series from 284 instruments. We aggregated data from three PNAS blind tests, the Journal of the Acoustical Society of America, Acta Acustica, Angewandte Chemie, Dendrochronologia, PLOS ONE and Scientific Reports. For the business side of the instrument trade, see our separate musical instruments market statistics; this roundup stays with the wood and the sound.
TL;DR
- Spruce at 0.44 g/cm3 vs maple at 0.64 g/cm3 gives a calculated radiation ratio of about 12.5 vs 7.5 m4/(kg s) (from Viala et al. 2024 nominal values).
- Spruce density across 49 top plates: 0.32 to 0.55 g/cm3 (Carlier et al., WoodMusICK COST Action 2018).
- Geometry carries 46.3% to 56.8% of the modeled influence on violin vibration vs 38.9% for material (Viala et al., Scientific Reports 2024).
- 6 of 10 elite soloists chose a new violin over Old Italians, and listeners guessed old vs new right 44.7% of the time (Fritz et al., PNAS 2014 and 2017).
- 90 of 180+ listeners rated a fungus-treated violin best vs 39 for a 1711 Stradivarius (Empa, 2009 blind test).
- Only robust quality marker in 17 violins: A0 air mode radiation near 280 Hz (Bissinger, JASA 2008).
- Stradivari violins sound like a shorter vocal tract: 16.15 cm vs 17.02 cm for other old Italians (Tai et al., PNAS 2018).
- Wood from Andrea Guarneri’s grandson held 1,200 and 2,900 ppm aluminium vs under 50 ppm for Stradivari (Su et al., Angewandte Chemie 2021).
- Thermal treatment at 230 °C cut spruce density by 16% (Merhar et al., BioResources 2025).
- Stradivari soundboards average 122 growth rings, mean ring width 0.95 mm (Bernabei et al., Dendrochronologia 2026).
1. Spruce vs Maple: The Core Tonewood Properties
A violin uses two woods for two jobs. Spruce, light and stiff along the grain, makes the top plate (the soundboard) that turns string vibration into sound. Maple, heavier and harder, makes the back, ribs and neck. The numbers that matter are density (mass per volume), Young’s modulus (stiffness along the grain), speed of sound (how fast a vibration travels through the wood) and the radiation ratio (speed of sound divided by density, a rough guide to how much sound a thin plate gives off for a given vibration).
The headline is how far apart the two species sit on that last measure. Using the nominal values from a 2024 finite element study of a complete violin, the speed of sound along the grain is c = √(E/ρ): for spruce √(13,350 MPa / 440 kg/m3), about 5,510 m/s and a radiation ratio near 12.5 m4/(kg s), against √(14,920 MPa / 640 kg/m3), about 4,830 m/s and near 7.5 for maple. Spruce is built to radiate; maple is not. Notice that the maple figure lands inside the 5 to 8 window that the Wegst criteria set for back and side woods.
| Metric | Value | Source |
|---|---|---|
| Density and Young’s modulus along the grain, spruce vs maple (nominal values in a violin model) | 0.44 vs 0.64 g/cm3; 13,350 vs 14,920 MPa | Viala et al. 2024, violin finite element study |
| Speed of sound along the grain, calculated as c = √(E/ρ), spruce vs maple | About 5,510 m/s vs about 4,830 m/s | Calculated from Viala et al. 2024 nominal values |
| Radiation ratio, calculated as c/ρ, spruce vs maple | About 12.5 vs about 7.5 m4/(kg s) | Calculated from Viala et al. 2024 nominal values |
| Ultrasonic speed of sound along the grain in violin spruce, at its typical density | 6,000 m/s at about 400 kg/m3 | Bucur, Acta Acustica 2025 |
| Density of spruce earlywood (pale spring growth) vs latewood (dark summer growth) | 300 kg/m3 vs 900 kg/m3 | Bucur, Acta Acustica 2025 |
| Curly maple density, and its highest compression-wave speed | About 700 kg/m3; 4,300 m/s | Bucur, Acta Acustica 2025 |
| Wegst criteria for back and side woods: radiation coefficient; damping (tan δ) | 5 to 8 m4/(kg s); 0.008 to 0.018 | Wegst criteria as cited in a Scientific Reports 2026 tonewood study |
Context note: the ultrasonic 6,000 m/s figure and the 5,510 m/s calculation come from different methods (1 MHz ultrasound on 16 mm cubes vs arithmetic on nominal model values), so compare values only within one method. Bucur’s 2025 review in Acta Acustica also reports a very low shear speed of 300 m/s across the grain of spruce and describes the wood as extremely anisotropic from an acoustical point of view (its properties change sharply with direction).
2. Natural Variation: Why No Two Tonewood Boards Match
The real story in tonewood is spread, not averages. Spruce top plates sold to violin makers ranged from 0.32 to 0.55 g/cm3 in density, so the heaviest board was about 72% denser than the lightest (0.55 / 0.32 = 1.72), and specific modulus (stiffness divided by density) almost doubled across the same set (35 / 18 = 1.94). A maker who orders “spruce” can receive boards that will behave very differently once carved.
Makers know this and grade wood mostly by eye and hand. In a 2018 European survey of violin makers, every responding luthier called weight or density important or very important, yet only 17% rated measuring tools as important. The modeling result in the same table asks a sharper question: if plate thickness and outline explain more of the vibration than the wood does, careful graduation of the plates may matter more than the board itself.
| Metric | Value | Source |
|---|---|---|
| Density range across 49 spruce top plates from France, Italy and Switzerland | 0.32 to 0.55 g/cm3 | Carlier et al., WoodMusICK COST Action 2018 |
| Specific modulus along the grain across the same spruce plates | 18 to 35 GPa (as printed) | Carlier et al., WoodMusICK COST Action 2018 |
| Growth ring width across the same spruce plates | 0.77 to 2.52 mm | Carlier et al., WoodMusICK COST Action 2018 |
| Spruce density variation within one tree: whole tree; from pith to bark | Almost 25%; 15% to 20% | Stoel and Borman, PLOS ONE 2008 |
| Violin makers rating weight or density as ‘very important’ when choosing wood (15 respondents) | 75% (the other 25% said ‘important’) | Carlier et al., WoodMusICK COST Action 2018 |
| Violin makers rating measuring tools as ‘important’ in wood choice | 17% | Carlier et al., WoodMusICK COST Action 2018 |
| Share of influence on modeled violin vibration: geometry vs material properties | 46.3% to 56.8% vs 38.9% | Viala et al. 2024, violin finite element study |
| Share of influence on matched eigenfrequencies (natural vibration frequencies): soundboard thickness vs back thickness | 24% vs 14% | Viala et al. 2024, violin finite element study |
Context note: in a 2021 ultrasonic grading of resonance spruce (24 specimens, 720 measurements, four quality classes), the speed of sound across the grain in the radial direction was the variable that best separated the classes, and classes A and C were about 60 kg/m3 heavier than B and D (Forests 2021). Most recent available data for the maker survey: Carlier et al., 2018, with only 15 complete responses.
3. Blind Playing Tests: Old Italian vs New Violins
Two blind playing experiments set the modern baseline. In the 2010 Indianapolis test, players wore modified welders’ goggles in a hotel room and compared three new violins with two by Stradivari and one by Guarneri del Gesu. The old violins were worth about $10 million combined, roughly 100 times the new ones, yet only 8 of 21 players took an old one home. The Paris follow-up with 10 elite soloists moved the test into a rehearsal room and a 300-seat hall, and the result held.
The lesson is not that new violins are better. Both studies found big individual differences, and the 2012 Indianapolis paper itself warns that its small sample cannot be generalized with confidence. What the data does rule out is the claim that top players can reliably pick out the old instruments by playing them. The soundpost result (the small wooden dowel wedged between top and back) shows how fine the margins are.
| Metric | Value | Source |
|---|---|---|
| Players and instruments in the Indianapolis test: old violins’ combined value vs new | 21 players, 6 violins; about $10 million, roughly 100 times the new ones | Fritz et al. 2012, Indianapolis blind playing test |
| Players who chose an old violin to take home | 8 of 21 (38%; CI 18% to 62%) | Fritz et al. 2012, Indianapolis blind playing test |
| Soloists who chose a new violin to replace their own for a concert tour | 6 of 10 | Fritz et al. 2014, Paris soloist study |
| Soloists’ old-or-new guesses: right, wrong, indeterminate (69 guesses) | 31 right, 33 wrong, 5 indeterminate | Fritz et al. 2014, Paris soloist study |
| Violinists and violins in a test of hearing vs touch in quality judgments | 20 violinists, 5 violins | Wollman et al., PLOS ONE 2014 |
| Players’ optimal soundpost length vs the original 53 mm post; smallest change they could not detect | 0.132 to 0.616 mm off; 0.11 mm or less not recognized above chance | Fu et al., JASA 2021 |
Context note: in the Indianapolis test, new violins were rated significantly higher for playability and response (P < 0.02), while projection showed no significant difference (P = 0.62). Most recent available data: the blind playing tests date from 2012 and 2014.
4. Blind Listening Tests in Concert Halls
Players hear a violin from a few centimeters away. Audiences hear it after the hall has done its work, and projection (how well the sound carries to the back of a hall over an orchestra) is a quality often credited to old Italian violins. The 2017 listener study tested exactly that, with listeners marking a continuous 0 to 1 scale for each pair of violins.
Both audiences judged the new violins to project better, and the New York audience also preferred them (observed difference 0.27, 95% CI 0.23 to 0.30). Paris used a 300-seat hall in Vincennes and New York the 860-seat Great Hall at Cooper Union; hall size and reverberation shape what reaches each seat (see our concert hall reverberation and clarity statistics). An earlier, less controlled test by the Swiss lab Empa pointed the same way.
| Metric | Value | Source |
|---|---|---|
| Listeners in the Paris and New York hall tests | 55 and 82 | Fritz et al. 2017, Paris and New York listener study |
| Projection advantage of new over Stradivari violins on a 0 to 1 scale: Paris; New York | 0.19 (95% CI 0.16 to 0.21); 0.27 (95% CI 0.23 to 0.31) | Fritz et al. 2017, Paris and New York listener study |
| Listener guesses of old vs new that were correct | 122 of 273 (44.7%) | Fritz et al. 2017, Paris and New York listener study |
| Audience at the Empa blind test of September 1, 2009 | More than 180 | Empa, 2009 blind test report |
| Votes for best tone: fungus-treated violin ‘Opus 58’ vs a 1711 Stradivarius | 90 vs 39 | Empa, 2009 blind test report |
| Listeners who thought ‘Opus 58’ was the Stradivarius | 113 | Empa, 2009 blind test report |
Context note: the Empa result came from a single conference audience voting after hearing one violinist play five violins behind a curtain; treat it as a demonstration, not a controlled study. The ‘Opus 58’ wood had been treated with fungus for nine months. Most recent available data: hall listening tests from 2017.
5. Signature Modes, Radiation and the Violin’s “Voice”
A violin body vibrates in a few dominant patterns at low frequency, called signature modes. The best known is A0, an air resonance where air pumps in and out of the f-holes like the note you get by blowing across a bottle. Higher up, near 2.4 kHz, the “bridge hill” (a broad peak in the response linked to the rocking motion of the bridge) shapes brilliance.
The striking finding is how little separates good from bad on these measures. Across 17 violins rated from bad to excellent, including Stradivari and Guarneri del Gesu instruments, mode frequencies and total damping showed no obvious quality trend; only A0 radiation was a robust differentiator. The vowel-formant analysis in the table offers a different lens: Stradivari violins resemble shorter vocal tracts, closer to tenors and altos than to the basses and baritones of other old Italian violins. Radiation and directivity measurements need rooms with controlled reflections; our anechoic chamber testing statistics cover that side.
| Metric | Value | Source |
|---|---|---|
| Violins measured in a modal and radiation study, rated from bad to excellent | 17 (including the Titian and Willemotte Stradivari and the Plowden Guarneri del Gesu) | Bissinger, JASA 2008 |
| Signature modes shared by all 17 violins | Same 5 modes below 600 Hz | Bissinger, JASA 2008 |
| Only robust quality differentiator found | A0 air-mode radiation near 280 Hz, significantly higher in excellent violins | Bissinger, JASA 2008 |
| Peaks in the violin radiation profile: B1 body modes; bridge hill; bridge filter; falloff | Near 500 Hz; near 2.4 kHz; near 3.5 kHz; above about 4 kHz | Bissinger, JASA 2008 radiativity model |
| Old Italian violins and singers recorded for vowel-formant comparison | 15 violins; 8 male and 8 female singers | Tai et al. 2018, PNAS |
| Equivalent vocal tract length, Stradivari vs other old Italian violins | 16.15 cm vs 17.02 cm (P = 0.006, Welch’s t test) | Tai et al. 2018, PNAS |
| Andrea Amati violin: first two formants (vowel-like resonance peaks) | F1/F2 = 503/1,583 Hz, similar to singers up to about 4.2 kHz | Tai et al. 2018, PNAS |
Context note: in the good and bad violins study, the Stradivari violins showed both the highest and the lowest directivity (how much sound is beamed in one direction) of all instruments tested, a reminder that famous makers did not produce one uniform sound. The radiation model paper and the vowel-formant study are the sources for the last four rows. Most recent available data: 2008 and 2018.
6. Wood Treatments and Chemistry
If the old violins are not reliably better, they are still chemically different. Two samples from Andrea Guarneri’s grandson held 1,200 and 2,900 ppm aluminium, against under 50 ppm in two Amati and eight Stradivari samples, consistent with alum treatment, according to a 2021 study tracking Cremonese wood over 120 years (1619 to 1741). The same group found sodium levels that point to salt seasoning in some Stradivari wood.
Modern treatments show the trade-off clearly. Heat and solvent treatments make wood lighter, which helps radiation, but they also make it less stiff, which hurts. That is why the plane wood study saw a radiation gain of only up to 4% alongside density losses of up to 13%. Thinner plates are a parallel route: the Cremonese tops are thinner at the center than a typical modern top.
| Metric | Value | Source |
|---|---|---|
| Aluminium in two samples from Andrea Guarneri’s grandson vs Amati and Stradivari samples | 1,200 and 2,900 ppm vs under 50 ppm | Su et al., Angewandte Chemie 2021 |
| Sodium in Stradivari spruce sample SC4 and maple sample MC3 | 13,000 ppm and 11,000 ppm | Su et al., Angewandte Chemie 2021 |
| Top plate center thickness: Stradivari; Guarneri; average modern top | 2.0 to 2.8 mm; 2.2 to 2.9 mm; about 3.0 mm (up to 3.5 mm in German schools) | Su et al., Angewandte Chemie 2021 |
| Rise in Young’s modulus with 1,000 ppm aluminium added (lab simulation reported in the same study) | 14% | Su et al., Angewandte Chemie 2021 |
| Na2O in a Stradivari violin’s maple vs Bosnian, Slovenian and Slovakian commercial maple | 11, 5 and 8 times higher | Nagyvary et al., PLOS ONE 2009 |
| Spruce after thermal modification at 230 °C: density; modulus of elasticity; shear modulus | Down 16%; down 9.8%; down 9.7% | Merhar et al., BioResources 2025 |
| Plane wood after solvent washing: density; elastic modulus; radiation coefficient; damping | Up to 13% lower; up to 31% lower; up to 4% higher; up to 18% higher | Scientific Reports 2026 plane tonewood study |
Context note: Nagyvary and colleagues also report chloride ratios of 52, 11 and 7 against the same three commercial maples, and boron in both Stradivari instruments sampled. The thermal modification study tested 180, 200 and 230 °C, and the plane wood study washed samples with water and ethanol-acetone. Cremonese samples were tiny (30 to 300 mg), taken during restorations.
7. Climate, Tree Rings and Where Stradivari’s Spruce Grew
Dendrochronology dates wood by matching its pattern of wide and narrow tree rings against reference chronologies from known forests. The 2026 Dendrochronologia study presents itself as the first large-scale tree-ring evidence of where his wood came from, and it traced most of Stradivari’s soundboard spruce to the Eastern Alps, with golden-age tops most likely from Val di Fiemme in Trentino. It also found that he often cut several instruments from the same tree.
Climate enters through ring width. The Maunder Minimum hypothesis holds that the long winters and cool summers of 1645 to 1715 produced slow, even growth and narrow rings. The CT density data fits that picture of evenness better than one of heaviness: classical and modern violins did not differ in median density, but the gap between light earlywood and dense latewood was significantly smaller in the classical instruments.
| Metric | Value | Source |
|---|---|---|
| Tree-ring series; authenticated instruments; reference chronologies | 314; 284; 197 | Bernabei et al., Dendrochronologia 2026 |
| Average number of growth rings in a Stradivari violin top | 122 | Bernabei et al., Dendrochronologia 2026 |
| Mean ring width in Stradivari tops (standard deviation) | 0.95 mm (0.31 mm) | Bernabei et al., Dendrochronologia 2026 |
| Share of samples in the two largest of 11 wood groups | About 80% | Bernabei et al., Dendrochronologia 2026 |
| Soundboards made of two joined pieces vs one single piece | 230 vs 24 | Bernabei et al., Dendrochronologia 2026 |
| Maunder Minimum, the cold period proposed to have shaped the wood | 1645 to 1715 (70 years) | Columbia Climate School (Lamont-Doherty Earth Observatory), 2003 release on Burckle and Grissino-Mayer |
| Earlywood vs latewood density gap in top plates, modern vs classical violins | 274 (SE 26.6) vs 183 (SE 11.7) g/L; difference significant (p = 0.028 for tops, 0.008 for backs) | Stoel and Borman, PLOS ONE 2008 |
| Median density difference, classical vs modern violins | Not significant (p = 0.884 top, p = 0.143 back) | Stoel and Borman, PLOS ONE 2008 |
Context note: the CT densitometry study compared only five classical (three Guarneri del Gesu, two Stradivari) and eight modern violins. The Maunder Minimum link was put forward as a hypothesis and has not been tested acoustically. The 2026 study does not give a percentage for Val di Fiemme alone.
Summary: Violin Tonewood Density & Acoustic Radiation by the Numbers
| Metric | Value | Source |
|---|---|---|
| Radiation ratio, spruce vs maple (calculated) | About 12.5 vs about 7.5 m4/(kg s) | Calculated from Viala et al. 2024 nominal values |
| Ultrasonic speed of sound along the grain in spruce | 6,000 m/s | Bucur, Acta Acustica 2025 |
| Spruce top plate density range, 49 plates | 0.32 to 0.55 g/cm3 | Carlier et al., WoodMusICK COST Action 2018 |
| Influence on modeled vibration: geometry vs material | 46.3% to 56.8% vs 38.9% | Viala et al. 2024, violin finite element study |
| Spruce density variation within one tree | Almost 25% | Stoel and Borman, PLOS ONE 2008 |
| Soloists choosing a new violin for a tour | 6 of 10 | Fritz et al. 2014, Paris soloist study |
| Correct old vs new listener guesses | 122 of 273 (44.7%) | Fritz et al. 2017, Paris and New York listener study |
| Best-tone votes, fungus-treated violin vs 1711 Stradivarius | 90 vs 39 | Empa, 2009 blind test report |
| Signature modes shared by 17 violins | 5 below 600 Hz | Bissinger, JASA 2008 |
| Robust quality differentiator | A0 radiation near 280 Hz | Bissinger, JASA 2008 |
| Equivalent vocal tract length, Stradivari vs other old Italians | 16.15 vs 17.02 cm | Tai et al. 2018, PNAS |
| Aluminium, samples from Andrea Guarneri’s grandson vs Stradivari samples | 1,200 and 2,900 ppm vs under 50 ppm | Su et al., Angewandte Chemie 2021 |
| Top plate center thickness, Stradivari vs average modern | 2.0 to 2.8 mm vs about 3.0 mm | Su et al., Angewandte Chemie 2021 |
| Spruce density drop after 230 °C thermal modification | 16% | Merhar et al., BioResources 2025 |
| Tree-ring series and instruments in the Stradivari origin study | 314 and 284 | Bernabei et al., Dendrochronologia 2026 |
| Mean ring width in Stradivari tops | 0.95 mm | Bernabei et al., Dendrochronologia 2026 |
Methodology and Sources
- Blind tests by Claudia Fritz, Joseph Curtin and colleagues, all in PNAS: Player preferences among new and old violins (2012), Soloist evaluations of six Old Italian and six new violins (2014), Listener evaluations of new and Old Italian violins (2017).
- Tai et al.: Acoustic evolution of old Italian violins from Amati to Stradivari, PNAS 2018.
- Viala, Placet, Foltete and Cogan: Model based ranking of the influence of geometry and materials on the dynamical behavior of the violin, Scientific Reports 2024.
- Bucur: Ultrasound and Wood Science at the Forestry Research Centre in Nancy, Acta Acustica 2025.
- Carlier, Alkadri, Gril and Bremaud: Revisiting the notion of resonance wood choice, WoodMusICK COST Action FP1302 book 2018.
- Correlation between Anatomical Grading and Acoustic-Elastic Properties of Resonant Spruce Wood, Forests 2021.
- Wollman et al.: Investigating the role of auditory and tactile modalities in violin quality evaluation, PLOS ONE 2014 (abstract via Europe PMC).
- Fu et al.: Perception of violin soundpost tightness through playing and listening tests, JASA 2021.
- Bissinger: Structural acoustics of good and bad violins, JASA 2008 and Structural acoustics model of the violin radiativity profile, JASA 2008.
- Empa: Fungus-treated violin outdoes Stradivarius, 2009 release.
- Su et al.: Materials engineering of violin soundboards by Stradivari and Guarneri, Angewandte Chemie 2021.
- Nagyvary, Guillemette and Spiegelman: Mineral preservatives in the wood of Stradivari and Guarneri, PLOS ONE 2009.
- Merhar et al.: Thermal treatment, moisture content, and vibration direction’s effect on dynamic properties of spruce wood for musical instruments, BioResources 2025.
- Evaluating urban oriental plane wood as a tonewood: effects of solvent extraction on acoustic performance, Scientific Reports 2026 (also the source of the Wegst criteria ranges quoted in section 1).
- Bernabei et al.: Tracing the origins of Stradivari’s resonance wood, Dendrochronologia 2026.
- Columbia Climate School (Lamont-Doherty Earth Observatory): release on Burckle and Grissino-Mayer’s Maunder Minimum hypothesis, December 5, 2003.
- Stoel and Borman: A comparison of wood density between classical Cremonese and modern violins, PLOS ONE 2008.
- Data watch: the speed of sound and radiation ratio rows marked “calculated” are our arithmetic from published nominal values (c = √(E/ρ), radiation ratio = c/ρ), not measured figures; the ultrasonic 6,000 m/s value in section 1 comes from a different method and should not be compared directly with them. The 2018 maker survey covers only 15 makers, and its specific modulus range is printed in GPa as given by the authors. The Indianapolis study’s 38% take-home figure has a wide confidence interval (18% to 62%) because only 21 players took part. The Empa 2009 vote was a single, uncontrolled audience test. The thermal modification paper also reports humidity effects on sound velocity and radiation that we left out because the abstract’s wording was ambiguous, and we dropped a damping range from the 2018 maker survey chapter because its printed upper bound was ambiguous. The Maunder Minimum explanation remains a hypothesis, and the 2026 tree-ring study gives no percentage for Val di Fiemme alone. Most blind-test, modal and chemistry data predate 2023; the 2024 to 2026 studies on modeling, thermal treatment, plane wood and tree rings are the newest available.
Last updated: November 7, 2026. We update this roundup quarterly, and the next refresh is expected when Acta Acustica publishes its first 2027 issue.