This paper presents OudTabs, a microtonal oud tablature representation, a cross-tradition symbolic corpus, and an interactive software platform for makam-family and related modal repertoires. Its underlying representation aims to diversify music research beyond Western staff notation by grounding itself in the oud’s fingerboard, preserving each tradition’s native pitch system, modal vocabulary, and rhythmic categories. Every note is encoded as string, position, and duration on the instrument, with pitch expressed as exact frequency-derived fractional positions in cents. The same cent-level values drive both the visual display and the audio, so learners and researchers hear each microtone at its target pitch while simultaneously seeing it as a concrete location on the oud. This cent-level resolution lets the OudTabs representation capture microtonal intonation more finely than the notations commonly used in oud teaching. Built on this foundation, we release the OudTabs Corpus, a cross-tradition tablature corpus of 3,672 pieces and 1,957,167 note events unified from heterogeneous public symbolic corpora, made playable, searchable, and audible through an interactive playback and browsing system. Together, these contributions provide new infrastructure for pedagogy, computational musicology, and future MIR research on microtonal oud music.

Q18. Write ONE line (in your own words) with the main take-home message from the paper.

By encoding Oud music using the instrument itself, cross-cultural analysis of Oud-playing traditions becomes possible.

Q23. Main review and comments for the authors (to be completed before the discussion phase). Please summarize strengths and weaknesses of the paper. It is essential that you justify the reason for the overall evaluation score in detail. Keep in mind that belittling or sarcastic comments are not appropriate.

This paper proposes to encode Oud music through the physical position on the instrument, to build a shared representation and allow cross-cultural anlysis across the different traditions bound to the Oud. It designs an encoding using the instrument as reference point, rather than any music-theoretical system or tradition, because the systems and traditions vary more across the regions where the Oud is used than the instrument itself. It runs several pilot analyses that showcase the potential of this encoding for analysing music from different traditions at once; finally, this is also a paper that actually delivers an application with pedagogical potential. The paper is also built on solid knowledge of the history of the Oud and its tradition.

I love the idea of going to the embodied, physical experience of performance that serves as the pivot for multiple musical traditions, and I definitely recommend this paper to be accepted. I would strongly encourage the authors to experiment with video-based pipeline for encoding performances in this representation, rather than limiting it to symbolic data: the idea and representation has great potential for ethnomusicological fieldwork as well.

As for critical remarks:

I don’t quite understand why the formula for position is defined the way it is, rather than 1 — (f_s / f), or just f_s / f, which would actually result in the position encoding being the fraction of the string from either end. The formula given in the paper instead returns the corresponding fret number — which is strange for a fretless instrument, and seems to be the exact thing the paper sets out to avoid (binding the representation to a specific tuning system). It’s technically a real number, just like a string fraction, and one can of course compute the position in terms of string fraction from it — it’s just a non-intuitive choice of unit, and should certainly be at least discussed.

This choice is also reflected in the UI of the pedagogical app. What is the meaning of seeing 1.7 on the screen? Is an assumption of a fretted-guitar-playing audience the reason for choosing a position encoding that is derived from fret positions? Does Oud pedagogy itself conceptualise position this way?

I understand that when working from symbolic data, some tuning has to be selected to project the symbolic data onto the instrument. However, while the article is clearly aware of the etic/emic distinction (L.314), I miss in the discussion some acknowledgment that the positions are derived from symbolic encodings. In L.98-99, it is stated: “…without flattening any tradition into the others’ theoretical terms”, but then the choice of x-TET tunings to project symbolic data back to the physical positions does go a bit in that direction. Some acknowledgment that the tuning systems used for projecting to frequency are also a construct (though with literature backing some of these choices, and awareness of their imperfections), and do not necessarily reconstruct the auditory experience fully, would be warranted, probably best when the hear-and-see interface is mentioned (L.332-334).

Similarly, given how crucial the instrument is to this work, I would have liked perhaps a figure or at least some discussion of the instrument itself and especially the extent to which Oud-building conventions, and therefore the validity of the hear-and-see interface computed from frequency data, might differ (or not) across its rather large area of use. I think Figure 2 could be easily sacrificed to make space for this, all that it shows is already said in the text.

Related work could perhaps mention the ongoing efforts to encode lute tablatures in MEI (E-LAUTE). It is only tangentially relevant, as it deals with encoding different notations rather than embodied practice, but it could highlight well how the relationship between text, instrument, theory, and practices shape representation choices.

As stated above, I recommend the paper for acceptance.

Q25. Meta-review and final comments for authors (to be completed after the discussion phase)

All reviews agree that this is an excellent idea executed well, that the dataset is a valuable resource, and that the work clearly deserves to be at ISMIR.

Q18. Write ONE line (in your own words) with the main take-home message from the paper.

The authors proposed OudTabs, contributing with (1) a oud-based tablature representation format in the digital space, (2) a corpus of 3,600 pieces and (3) a interactive player/interface to playback and display oud tabaltures.

Q23. Main review and comments for the authors. Please summarize strengths and weaknesses of the paper. It is essential that you justify the reason for the overall evaluation score in detail. Keep in mind that belittling or sarcastic comments are not appropriate.

Overall, the paper is very well written and organized. It's contributions are clear and well presented. I particularly enjoy the interface and the demo. I believe it is of extreme importance in this day and age to tackle musical genres/styles that are often hidden from the spotlight in a more westernized global world.

Despite the title reflecting well the outcomes of the paper, I believe that emphasising the "dataset" part of the contribution would be valuable, as it is quite extense. Moreover, I'd appreciate a bigger RELATED WORK section, perhaps including references to other tablature-based corpus recently released in the literature (e.g. DadaGP, SynthTAB).

The analysis of the corpus is comprehensive, and I think the authors have done a very good job in unifying all these traditions.

Finally, I'd encourage the authors to publish a Ethical Statement section in case the paper gets accepts, disclosing how they intend for the dataset to be used.

Q18. Write ONE line (in your own words) with the main take-home message from the paper.

A new tablature-like representation for oud music is available with an online interactive UI and a huge corpus of oud music.

Q23. Main review and comments for the authors. Please summarize strengths and weaknesses of the paper. It is essential that you justify the reason for the overall evaluation score in detail. Keep in mind that belittling or sarcastic comments are not appropriate.

The paper describes a project focusing on oud music and tablature. The Project consists of a large corpus of scores that was standardized and converted into a unique tablature format. It is accompanied by a website with an interactive interface for viewing the tablatures. I find the paper generally fascinating and enriching. And I think it could serve as a great introduction for readers, researchers and players into the world of oud/maqam oriented music. I also think that the corpus and interface could be very useful to a large number of people.

I specifically appreciate the online interface. The very fact that it is possible to hear the music without having to play it is very useful such as for practicing micro tonal perception and generally getting acquainted with this type of music and tuning systems.

Major comments

• l251, the task of mapping tones to strings is interesting. Essentially it is a sort of score-to-tab problem which is a challenging thing to do for a guitar, let alone for an Oud. I would love to hear more about implementation and the rules. How were they arrived at? Could different rules be defined, and if so, how different could they reasonably be? Generally speaking regarding Oud playing, do different players finger the same phrases differently?
• Section 5.1, This is an interesting analysis. But could this be circular logic? The fingerboard representation was generated from scores using the authors rules. It is then analyzed. So the analysis actually pertains to the output of the authors' rules rather than the musical tradition's ground truth. I do understand that this is the most that is possible at this stage, and I find the analysis interesting nonetheless. Still, perhaps acknowledging this point would be helpful.
• Section 5.2 This is very interesting and potentially very useful to novice players. What could be interesting is to verify the classification vs some (subjective) ground truths. Let some experienced players rate a small selection of representative compositions, and compare to the analysis.

Minor comments:

• Figures would work well as vector graphics, especially fig1
• Is it possible to position figures earlier in the text? Currently figures appear much later to the corresponding text.
• L225-227 perhaps define these terms, surely they would be unfamiliar to many readers.
• Personally I've found it excessive to include the entire corpus in the submission material. 3 representative files would have been sufficient, and the rest of the corpus could be accessible online via the existing interface.

Reviewer and meta-reviewer comments are addressed below, with two suggestions noted as future work.

A new figure gives a reference oud tuning and coordinate geometry, distinguishing continuous pitch position p_s (one unit = 100 cents above the open string, so a displayed 1.7 denotes 170 cents) from normalised stopping distance x_s = 1 - f_s/f = 1 - 2^(-p_s/12), which needs a scale length for physical units. The earlier system-overview figure was dropped to make room, as its content was already in the text. "Fractional semitone" is replaced by 100-cent units throughout. The paper now states that the x-TET grids are those of the integrated symbolic datasets rather than the traditions themselves, that OudTabs is not bound to them, and that the player supports alternate tunings. The Discussion adds that performance varies in fingering, intonation, timing, and ornamentation beyond what the system shows. Section 4.3 now gives the string-selection procedure (strings scanned highest to lowest; strings 2-5 accept -0.2 <= p_s <= 5, string 1 accepts p_s >= -0.2; first eligible string selected), its open-string and low-position rationale, and that it is one deterministic canonical projection, not a model of how players of different regions finger phrases. Section 5.1 now notes that the heatmaps describe OudTabs output under this mapping rule, not observed performer fingerings. Related work adds a paragraph on tablature corpora (E-LAUTE, JosquIntab, DadaGP, SynthTab). The dataset is now named the OudTabs Corpus, foregrounded in the abstract and contributions with exact counts, and archived under a DOI. An Ethics Statement was added. Future work now includes expert rating of the difficulty bands and a performance-to-tab pipeline from video or audio. Figures are vector PDFs, placed nearer their text, with alt text. Emic terms (makam, seyir, usul, dastgah, gusheh, nawba, tab', mizan) are glossed at first use. Authors de-anonymised. The demo link now points to the public player.