{"id":"aa452486-7040-4134-a751-4c90eb502a06","arxiv_id":"2507.18723","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"SCORE-SET provides Guitar Pro tablature files derived from piano MIDI datasets with randomly injected guitar expression techniques.","lead":"SCORE-SET is a dataset of Guitar Pro tab files built by converting piano MIDI from MAESTRO and GiantMIDI into guitar tabs and adding random expressive techniques like bends and palm muting. It aims to give music-generation researchers guitar-specific training data, but the expressions are synthetic, not from real performances.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Random expression injection with metal-derived ratios makes the 'real-world guitar performance' claim unsupported; context-free placement undermines performance-aware learning.","rationale":"The reader's weakest_assumption identifies random expression injection and the unvalidated playability heuristic; this stress-test agrees but sharpens the concern to the context-free nature of that injection. Random placement is not merely an evaluation gap; it is an internal property of the pipeline described in Section 3, and it directly contradicts the abstract's claim that the dataset reflects real-world guitar performance. A performance-aware model trained on this data would learn that expression is independent of phrase context, which is musically false and would make generated performances sound arbitrary. The genre mismatch between the metal-derived ratios and the classical piano source corpus compounds the issue. The proposed test is concrete and feasible because the dataset is released and includes .gp5 files with expression annotations. If the test shows context-independent placement, the central claim fails as stated; if it shows context-dependent placement, the author's pipeline description would need correction but the dataset could be repairable. Therefore the existing CONDITIONAL verdict remains appropriate: the paper should not be accepted as-is, but the concern is addressable through additional validation and documentation rather than being a fundamental flaw in the dataset concept.","tokens_in":3919,"tokens_out":3652,"duration_ms":41433,"concrete_test":"Sample 200 notes from the released SCORE-SET .gp5 files and 200 notes from human-transcribed Guitar Pro files of comparable repertoire. For each note, record the presence or absence of each expression plus local features: note duration, metrical position, preceding interval, and chord versus single-note context. Fit a logistic regression predicting expression presence from these features. If the SCORE-SET model achieves near-chance AUC (≈0.5) while the human-transcription model achieves AUC > 0.7, then random context-free injection is confirmed and the realism claim is unsupported; if SCORE-SET's conditional placement matches human patterns, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that SCORE-SET reflects the nuances of real-world guitar playing. The only mechanism for adding expression is Section 3: 'Accents were added to the original dataset at random, following the expression ratios derived from Figure 5.' Random injection makes each technique's occurrence independent of musical context—pitch contour, duration, phrasing, chord structure, and metrical position—yet these are precisely what makes expression typical of guitar performance and what a performance-aware model must learn. Expression ratios alone, even if accurately measured from the undocumented manual metal-song transcription, encode only marginal frequencies, not placement. Moreover, those ratios come from metal transcriptions and are applied to classical piano MIDI from MAESTRO and GiantMIDI, a genre mismatch that further weakens representativeness. The paper provides no evaluation of musical coherence, no comparison against human-written tablature, and no validation of the Section 2.3 playability heuristic, which can alter octaves ('pitch was adjusted by adding or subtracting octaves') and thereby change the musical content. Consequently, the abstract's claim is an assertion of intent, not a demonstrated property of the released dataset.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents SCORE-SET, a dataset of Guitar Pro tablature files (.gp5) created by converting piano MIDI from MAESTRO and GiantMIDI into rhythm guitar parts, then augmenting them with guitar-specific articulations such as bends, slides, vibrato, palm muting, dead notes, hammer-ons, and harmonics. The augmentation uses expression frequencies derived from a manual transcription of modern metal songs, with accents added randomly at those ratios. The authors position the dataset for guitar music generation, sequence modeling, and performance-aware learning, and they provide a statistical overview of the articulation distribution.","tokens_in":4287,"tokens_out":5618,"duration_ms":53622,"significance":"If its realism were established, SCORE-SET would be a useful public resource for symbolic guitar research, with strengths including its release in a standard Guitar Pro format, use of well-known MIDI sources, and reliance on an open-source exporter. However, the paper currently establishes neither that the expression content is representative of real-world guitar playing nor that the musical transformations preserve musical coherence. The central claim depends on undocumented statistics and a random placement model, and no validation against human-written tablature or downstream learning experiments is offered. With careful revision—documenting the transcription, justifying genre transfer, quantifying the octave-shift heuristic, and adding validation—the dataset could become a meaningful contribution.","major_comments":[{"comment":"The paper's central claim in the abstract that the processed dataset 'better reflect[s] the nuances of real-world guitar playing' is not supported by the construction described in Section 3: 'Accents were added to the original dataset at random, following the expression ratios derived from Figure 5.' Random injection makes each technique's occurrence independent of pitch contour, duration, metrical position, chord structure, and phrasing—exactly the contextual cues that determine idiomatic guitar expression and that a performance-aware model must learn. The expression ratios encode only marginal frequencies, not placement. No musical-coherence evaluation or comparison against human-written guitar tablature is provided to close this gap.","section":"Section 3 (Statistics); Fig. 5"},{"comment":"The expression statistics that drive the augmentation come from a manual transcription of 'modern metal songs' that is not described as a method: no song list, number of songs, annotation procedure, or inter-annotator/error assessment is given. The text only states that transcriptions were 'created by ear.' These undocumented ratios are then applied to classical piano MIDI from MAESTRO and GiantMIDI, a genre and instrument mismatch. The paper does not justify why metal-specific expression frequencies transfer to non-metal piano-derived material, so the representativeness claim rests on unverifiable data.","section":"Section 3, Figure 5"},{"comment":"The playability heuristic can alter the musical content: 'If multiple notes fell on the same string, their pitch was adjusted by adding or subtracting octaves until a playable fret position was found on different strings.' The paper does not quantify how often octave shifts occur, nor does it validate the 'plausible playability' heuristic against any standard (e.g., human transcription or a playability checker). Since octave changes alter the pitch content of the original source, this transformation is load-bearing for the dataset's fidelity and must be measured and discussed.","section":"Section 2.3 (String and fret)"}],"minor_comments":[{"comment":"The first paragraph contains an article error ('curated dataset' should be 'a curated dataset') and a typo ('adapted innto' should be 'adapted into').","section":"Section 1"},{"comment":"The statement that 'MIDI formats represent time signatures using ticks' is technically inaccurate; time signatures are carried by meta-events, while ticks provide only a resolution for note timing.","section":"Section 2.2 (Quantization)"},{"comment":"The caption reads 'Tremlo types and their pitch variations'; 'Tremlo' should be 'Tremolo'.","section":"Figure 2 caption"},{"comment":"The figure's axis is labeled 'Count ratio' while the text refers to 'Percentages of techniques'; the log-scale tick labels (10^1 through 10^4) and the precise definition of 'count ratio' should be clarified.","section":"Figure 5"},{"comment":"The treatment of tremolo is an exception ('applied to an entire beat, which may contain multiple notes') but the figure appears to count TREM_BAR occurrences like other note-level techniques; please state explicitly how tremolo counts are computed and whether the 25.39% accent figure includes tremolo.","section":"Section 3 (Statistics)"},{"comment":"The dataset description would benefit from a dataset card giving the number of .gp5 files, total note count, total duration, license, versioning, and whether the conversion script is also released.","section":"Section 2 (SCORE-SET Dataset)"},{"comment":"The claim that 'there remains a lack of large-scale, high-quality resources tailored specifically to the guitar' omits existing Guitar Pro datasets such as DadaGP; please cite and differentiate SCORE-SET from them.","section":"Section 1 (Introduction)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is very short for a dataset paper and may need additional content to meet the journal's standards for dataset contributions. A key novelty-disclosure concern is the absence of any comparison with existing Guitar Pro datasets (e.g., DadaGP) that already contain expressive techniques. The central representativeness claim cannot be accepted without either substantially revising the claim or providing validation of the random-injection model."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Vishakh, quick take on 2507.18723. The usable thing here is the artifact: a set of Guitar Pro files converted from MAESTRO and GiantMIDI, with guitar-specific expressions (bends, slides, palm mute, harmonics) written in. That is genuinely useful to people doing guitar-tab generation, and the author deserves credit for putting it on GitHub and using PyGuitarPro so the files are inspectable. The expression statistics in Figure 5, based on a manual transcription of metal songs, also show real effort.\n\nThe problem is the framing. The abstract says the dataset 'reflects the nuances of real-world guitar playing,' but Section 3 states plainly that accents were 'added at random' using ratios from a separate, undocumented transcription of metal songs. Random injection means the expressions carry no musical context—no relation to pitch contour, rhythm, phrasing, or harmonic function. That undermines the 'performance-aware learning' selling point, because a model trained on this will learn frequencies, not placement. The genre mismatch is also real: metal-derived ratios are being applied to classical piano pieces. The octave-adjustment heuristic in Section 2.3 is another soft spot, since it can alter the musical content of the original MIDI.\n\nThe paper is honest about the random injection—it is right there in the text—so this is an overclaim rather than a hidden flaw. But the overclaim matters. Without any evaluation of playability or musical coherence, and without details about the manual transcription (which songs, how many, how verified), the dataset's value is speculative.\n\nBottom line: the author has created a resource that could be useful for benchmarking guitar-tab generation, but the claims need to be pulled back and the method documented more fully. I would send this to peer review, because dataset papers can be valuable even when raw, and the flaws are fixable—add a playability check, release the conversion code and the random seed, describe the transcription, and reframe the contribution as 'a synthetic tab resource' rather than 'real-world performance.' As written, I would not cite it yet, but I would keep an eye on a revised version.","headline":"A small but real dataset artifact whose abstract overclaims: random expression injection from metal-song ratios cannot, by itself, support the 'real-world guitar performance' framing.","tokens_in":636,"tokens_out":823,"would_cite":false,"duration_ms":25442,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"SCORE-SET converts piano MIDI recordings into Guitar Pro tablature with bends, slides, vibrato, and palm muting, giving sequence models guitar-specific training data.","keywords":["Dataset","Guitar tablature","Guitar Pro","Sequence learning","Music generation","Expressive techniques","MIDI-to-tablature","Transformer"],"falsifier":"A fret-span calculation on the exported .gp5 files would test the playability claim: if a substantial share of chord voicings require stretches wider than an average hand, the dataset cannot be said to reflect real-world guitar performance.","tokens_in":3702,"feed_emoji":"🎸","tokens_out":9700,"duration_ms":95185,"temperature":0.7,"pith_summary":"This paper introduces SCORE-SET, a dataset of Guitar Pro tablature files built by converting existing piano-oriented MIDI recordings into rhythm guitar tracks. Each note is mapped to a string and fret, rhythms are quantized to a sixteenth-note grid in 4/4, and then guitar techniques such as bends, slides, vibrato, palm muting, hammer-ons, harmonics, and dead notes are added at frequencies taken from a set of manually transcribed metal songs. The stated goal is to give sequence models a large body of guitar-specific symbolic music with realistic expression, which piano-heavy datasets do not provide. If the conversion works as claimed, researchers can train guitar generation models directly on tablature with playable, expressive output rather than plain note sequences.","feed_headline":"New dataset turns piano MIDI into expressive guitar tablature","feed_subtitle":"Guitar Pro files with bends, slides, vibrato, and palm mutes give sequence models guitar-specific training data.","key_machinery":"The load-bearing object is the Guitar Pro tablature file (.gp5): each note is stored as a string-fret combination with articulation flags, so a generated file is a complete performance instruction rather than a bare pitch-duration sequence. The conversion machinery has three components: a MIDI-to-tablature heuristic that fixes the root note and assigns chord tones to consecutive strings while keeping frets low; a quantization step that snaps off-beat notes to the nearest beat and aligns all measures to 4/4; and a random accent-injection step that applies techniques at ratios computed from a manual transcription of metal songs. The .gp5 format and the injection ratios are what carry the claim that the data contains realistic, playable guitar expression.","core_discovery":"The paper's central claim is that a large, guitar-specific symbolic music dataset can be built by converting piano-oriented MIDI into rhythm guitar tracks and then enriching the tracks with guitar techniques. The conversion maps each MIDI pitch to a string-fret pair using a heuristic that favors lower frets and continuous melodic motion, quantizes durations to a sixteenth-note grid in 4/4, and attaches articulations such as bends, slides, vibrato, palm mutes, hammer-ons, natural harmonics, and dead notes at frequencies measured from manually transcribed metal songs. The result is a set of .gp5 tablature files that preserve note-level expression attributes, which the paper offers as training material for guitar phrase generation and sequence learning.","pith_inferences":["A direct extension would be to train a sequence model on SCORE-SET and check whether the articulations it generates match the injected ratios from Figure 5; this would test whether the dataset actually teaches expression rather than just note content.","Because the expression ratios were measured only on metal transcriptions, applying the same injection pipeline to other genres would likely require genre-specific ratio data, suggesting a conditional extension of the dataset design.","The paper does not include human evaluation of the tablature, so a perceptual study in which guitarists compare random SCORE-SET files to human transcriptions would be the prudent next step before relying on the dataset for performance-aware generation."],"forward_implications":["Generative models trained on the dataset have access to articulation labels, so they can learn to place bends, slides, vibrato, and palm mutes in musically plausible positions rather than treating every note as a plain pitch.","Because outputs are .gp5 files, generated phrases can be loaded into tablature software and listened to directly, making qualitative evaluation much easier than with raw MIDI.","The reported expression statistics give modelers a numerical prior for how often each technique should appear in generated guitar music.","The dataset repurposes existing piano-oriented symbolic corpora for the guitar domain, which is useful because guitar-specific symbolic data with technique annotations is scarce."],"supporting_citations":[{"why":"Supplies the source MIDI performances from which rhythm guitar tracks are adapted.","marker":"Hawthorne et al. (2019)"},{"why":"Supplies the classical piano MIDI collection used as the other source of adapted tracks.","marker":"Kong et al. (2022)"},{"why":"Provides the library used to export the processed tracks to .gp5 format.","marker":"Abakumov (2014)"},{"why":"Makes the SCORE-SET dataset publicly available for download.","marker":"Begari (2025)"}],"fun_headline_variants":["New dataset: piano MIDI turned into expressive guitar tablature","Guitar tab dataset with bends, slides, and vibrato for ML","SCORE-SET: GuitarPro files from MIDI for sequence learning","Expressive guitar tabs from piano MIDI for phrase generation","Dataset turns piano MIDI into guitar tabs with techniques"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that randomly injecting guitar techniques at ratios measured from a manually transcribed set of metal songs, combined with a heuristic fret placement, yields tablature that reflects how guitarists actually play.","fun_headline_variants_meta":{"raw":{"variants":["New dataset: piano MIDI turned into expressive guitar tablature","Guitar tab dataset with bends, slides, and vibrato for ML","SCORE-SET: GuitarPro files from MIDI for sequence learning","Expressive guitar tabs from piano MIDI for phrase generation","Dataset turns piano MIDI into guitar tabs with techniques"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00063,"raw_usage":{"total_tokens":2821,"prompt_tokens":769,"completion_tokens":2052,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":385,"completion_tokens_details":{"reasoning_tokens":1963}},"tokens_in":385,"tokens_out":2052,"duration_ms":13139,"temperature":1.0,"reasoning_tokens":1963,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:07:48.868381+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A fret-span calculation on the exported .gp5 files would test the playability claim: if a substantial share of chord voicings require stretches wider than an average hand, the dataset cannot be said to reflect real-world guitar performance.","supporting_citations":[{"cited_title":"Pyguitarpro","cited_arxiv_id":null,"evidence_quote":"Provides the library used to export the processed tracks to .gp5 format."},{"cited_title":"Score-set","cited_arxiv_id":null,"evidence_quote":"Makes the SCORE-SET dataset publicly available for download."}],"review_version":2}