{"id":"7dea6cc4-403a-4de8-a132-441dfb78f5cc","arxiv_id":"2502.02929","paper_version":4,"verdict":"REJECT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"An AR headset interface for placing and moving 3D audio objects with hand gestures was rated more usable, less mentally demanding, and more creative than a 2D desktop panner, though the comparative test disabled the headset's 6DoF motion.","lead":"AudioMiXR is an augmented reality app for the Apple Vision Pro that lets users pick up and move virtual sounds in 3D space with hand gestures. The authors report that the AR interface felt more usable, less mentally taxing, and more creative than a desktop 2D panning tool, and they offer design lessons for future AR sound design.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Study 2 restricted AR to 3DoF (§5.3.3), so the measured SUS/TLX/creativity gains cannot be attributed to 6DoF; the headline claim is unsupported.","rationale":"The paper's central claim, stated in the abstract and conclusion, is that '6DoF AR interaction yields measurable gains in user experience and creative output.' The only comparative evidence for this is Study 2. For that claim to hold, Study 2 must manipulate 6DoF. It does not: §5.3.3 explicitly restricts the AR condition to 3DoF to match the 2D panner. The reported SUS, NASA-TLX, and creativity advantages are therefore advantages of the 3DoF AR condition over a 2D mouse-based panner, confounded by display and input modality. They cannot be attributed to 6DoF. The qualitative Study 1 did use 6DoF but had no baseline and did not collect SUS or the creativity subscales. This is not a matter of interpretation; the paper's own limitation statement confirms the mismatch. A secondary concern is that all paired t-tests in §6.5 report t(11), implying 12 participants, while §5.3.2 reports N=18. This internal inconsistency must be resolved; if the sample was actually 12, the reported effect sizes are not trustworthy. But even correcting the statistics would not fix the missing 6DoF condition. The proposed test—a three-arm study with 2D, 3DoF, and 6DoF AR—would directly determine whether translation adds any measurable benefit over 3DoF for these outcomes. Until such a test is run, the central claim is unsupported.","tokens_in":31795,"tokens_out":4519,"duration_ms":41167,"concrete_test":"Run a three-arm within-subjects study with at least 18 participants: (A) 2D panner, (B) 3DoF AR, (C) 6DoF AR, using the same AudioMiXR visuals, audio settings, tasks, and measures as Study 2. Preregister pairwise comparisons of SUS, NASA-TLX, and creativity. If (C) does not significantly outperform (A) and (B), or if (C) is not significantly better than (B), the 6DoF attribution fails. Additionally, verify the Study 2 degrees of freedom against the raw data: with N=18, paired t-tests should have df=17; report corrected statistics or provide a clear explanation for t(11).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that 6DoF AR interaction yields measurable gains in usability, workload, and creativity. For that claim to hold, Study 2 must compare a 6DoF AR condition against the 2D panner. However, §5.3.3 explicitly states: 'we restricted it to 3DoF (rotation + orientation only) to match with the directional placement capabilities of the 2D interface.' Thus the comparative experiment tested 3DoF AR, not 6DoF. The abstract and conclusion nevertheless attribute the observed SUS, NASA-TLX, and creativity improvements to '6DoF interaction.' Since no 6DoF condition was included, the data cannot support that attribution. The only 6DoF data come from Study 1, which had no 2D baseline and did not collect SUS or the creativity subscales. The design also confounds DoF with interaction modality (free-hand gestures vs. mouse) and display type (HMD AR vs. desktop monitor). A separate internal inconsistency: all paired t-tests in §6.5 report t(11), implying N=12, while §5.3.2 reports N=18. If the actual sample was 12, the reported effect sizes are not trustworthy; even a corrected N would not rescue the 6DoF claim because the AR condition was not 6DoF.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents AudioMiXR, an AR head-mounted display interface for manipulating virtual spatial-audio objects with free-hand gestures, and reports two user studies. Study 1 (N=27) is an exploratory 6DoF sound-design study that produces qualitative themes and two design lessons (proprioception for AR sound design; balancing audio-visual modalities). Study 2 (reported N=18) compares AudioMiXR against a 2D desktop panner and reports significant improvements in SUS, NASA-TLX, and creativity, which the abstract and conclusion attribute to '6DoF AR interaction.' The quantitative comparison is claimed to validate the design lessons from Study 1.","tokens_in":32078,"tokens_out":2689,"duration_ms":29403,"significance":"If the central claim were supported, the contribution would be substantial: AudioMiXR would be among the first systems to provide evidence that 6DoF embodied manipulation improves usability, workload, and creative output for spatial audio design, and the two design lessons would offer actionable guidance. The qualitative analysis of Study 1, including placement heat maps and semantic clustering of audio objects, is a useful exploratory contribution. However, the central comparative claim is not supported by the data as reported: Study 2 restricted the AR condition to 3DoF, so no 6DoF comparison exists, and the statistical reporting contains an unresolved sample-size inconsistency. These issues bear directly on the paper's headline conclusion.","major_comments":[{"comment":"The central claim that '6DoF AR interaction yields measurable gains' is not tested. Study 2's procedure explicitly states that the AR interface was 'restricted it to 3DoF (rotation + orientation only) to match with the directional placement capabilities of the 2D interface.' No condition in Study 2 provided translational 6DoF movement, yet the abstract and conclusion attribute the SUS, NASA-TLX, and creativity improvements to 6DoF interaction. The measured differences between the AR and 2D conditions are therefore differences between a 3DoF HMD gesture interface and a 2D mouse interface, confounded by interaction modality and display type. The data cannot support the paper's headline attribution.","section":"§5.3.3, Abstract, §8"},{"comment":"The reported sample size is internally inconsistent. Section 5.3.2 states N=18 participants in Study 2, but every paired t-test in Section 6.5 reports t(11), which implies N=12. The paper gives no explanation for this discrepancy, so it is impossible to determine which statistics correspond to the described study. Because all of the SUS, NASA-TLX, and creativity conclusions rest on these tests, the quantitative results are not verifiable as reported.","section":"§5.3.2 vs. §6.5"},{"comment":"The creativity measure is partly circular. Section 5.3.4 states that the creativity questions were 'informed by Study 1's results,' and Section 7.3 uses Study 2's creativity results to 'validate and expand on the design lessons from our exploratory study.' If the outcome instrument is constructed from the same qualitative findings it is later used to confirm, the confirmatory value is weakened. In addition, the creativity subscales named in Section 7.3 (originality, elaboration, expressiveness) do not appear in the measures description in Section 5.3.4, making it unclear what was actually administered and analyzed.","section":"§5.3.4, §6.5.3, §7.3"}],"minor_comments":[{"comment":"The NASA-TLX description lists five dimensions and omits frustration, but Section 6.5.1 reports frustration results; please correct the list. Also, the SUS is described with a 7-point Likert scale, whereas the standard SUS uses a 5-point scale; the scoring-to-100 procedure should be specified.","section":"§5.3.4"},{"comment":"The normality test is referred to as the 'Shapiro-Wilkinson test'; the correct name is the Shapiro-Wilk test.","section":"§6.1"},{"comment":"The phrase 'took some trail and error' should be 'trial and error.'","section":"§7.2"},{"comment":"The caption states that users are shown with a Meta Quest Pro, while the study text says the deployment target was the Apple Vision Pro; please clarify whether the figure shows a different hardware version or a separate demonstration.","section":"Figure 8 caption"}],"recommendation":"reject","confidential_remarks":"The mismatch between the 3DoF Study 2 condition and the 6DoF headline claim is not a presentation issue but a fundamental gap between the evidence and the conclusion. The sample-size inconsistency compounds this. If the authors re-run the comparison with a true 6DoF condition and clarify the statistics, the exploratory Study 1 material could form the basis of a useful future submission, but the present manuscript does not support its central quantitative claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nTwo things you should know about this one. First, the headline result in the abstract—that 6DoF AR interaction yields measurable usability, workload, and creativity gains over a 2D panner—is not actually tested in the paper. Study 2, the only controlled comparison, explicitly restricted the AR condition to 3DoF (§5.3.3) to match the 2D panner's directional placement. So the data cannot attribute any observed gains to 6DoF. Second, all the paired t-tests report t(11), which implies N=12, while §5.3.2 says N=18. That inconsistency is never addressed.\n\nThat said, the paper isn't empty. The system itself is a real contribution: a working AR sound-design interface on the Apple Vision Pro with free-hand pinch/push/pull manipulation, and it appears to be the first 6DoF AR tool for this task. Study 1 is a reasonable exploratory study—27 participants, thematic analysis, heat maps of audio object placements—and the two design lessons (proprioception for AR sound design; balancing audio-visual modalities) are plausible and grounded in the qualitative data. The heat-map analysis showing scene-dependent spatial distributions is a nice touch. The related work is thorough.\n\nThe soft spots are not minor. The central claim is overreach, and the N inconsistency undermines trust in the statistics. The creativity measure is self-designed and built from Study 1's own themes, so the confirmation is partly circular. There's no data or code released, which makes it hard to check the analysis.\n\nWho is this for? Someone building AR/VR audio tools will get useful design hints and a working reference implementation. Anyone looking for evidence that 6DoF beats 2D will be disappointed. As it stands, the paper overclaims; the good parts are the system description and the exploratory lessons.\n\nMy take: send it to peer review—the system and Study 1 deserve referee time—but it needs major revision before acceptance. The authors should reframe the claims to what Study 2 actually tested (AR at 3DoF vs 2D), reconcile the N, or run a true 6DoF comparison. If they don't fix that, it's not publishable in its current form.","headline":"Study 2 tested 3DoF AR, not 6DoF, so the abstract's 6DoF claim is unsupported; the system and Study 1's design lessons still warrant a referee.","tokens_in":32616,"tokens_out":1824,"would_cite":false,"duration_ms":18922,"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":"This paper claims that a 6DoF augmented-reality interface for spatial audio manipulation yields measurable gains in usability, workload, and creativity over a 2D desktop panner.","keywords":["augmented reality","spatial audio","sound design","6DoF","head-mounted display","usability","NASA-TLX","creativity"],"falsifier":"Run a second controlled comparison with three conditions — a 2D panner, an AR headset restricted to 3DoF rotation, and an AR headset with unrestricted 6DoF translation and rotation — holding all audio stimuli, headset, and task identical. If the 3DoF and 6DoF conditions produce the same SUS, NASA-TLX, and creativity scores, then the gains are not caused by 6DoF translation, and the paper's central attribution would be falsified.","tokens_in":31593,"feed_emoji":"🎧","tokens_out":4343,"duration_ms":38094,"temperature":0.7,"pith_summary":"This paper argues that spatial audio sound design, traditionally done on 2D desktop panners that force users to mentally map three-dimensional sound onto flat screens, can be made more usable and more creative by letting users physically walk around and grab audio objects in augmented reality. To establish this, the authors built AudioMiXR, an AR interface running on a head-mounted display in which binaurally rendered sound objects are pinched, pushed, and pulled through 6DoF space, and they ran two within-subjects studies. The first, exploratory study with 27 expert and non-expert sound designers produced two design lessons: proprioception supports AR sound design, and audio-visual modalities need careful balancing in AR GUIs. The second, comparative study with 18 novice users found that AudioMiXR significantly improved System Usability Scale scores, reduced NASA-TLX frustration and mental workload, and improved all creativity subscales relative to a 2D panner baseline. A sympathetic reading takes the central claim to be that 6DoF embodied interaction yields measurable gains in user experience and creative output for spatial audio authoring.","feed_headline":"AR sound mixing beats 2D panner in usability, creativity","feed_subtitle":"In two studies, head-mounted 6DoF object manipulation cut mental load and frustration while lifting creative output.","key_machinery":"The carrying mechanism is AudioMiXR itself: a Unity-built AR system on a head-mounted display that renders each sound as a spherical audio object with 3D spatialization, distance-based logarithmic attenuation, and real-time FFT-driven visual pulsation. Users interact through free-hand gestures (pinch to grab, push to move away, pull to bring closer) while the system tracks head position and orientation in 6DoF, so walking around the room changes the binaurally rendered perspective. The design lessons build on the user's body as an input-output interface and on Gestalt-like semantic clustering of audio objects in physical space.","core_discovery":"The central discovery claimed is that a 6DoF augmented-reality interface for manipulating spatial audio objects outperforms a conventional 2D desktop panner on perceived usability, workload, and creative output. Concretely, in a matched within-subjects comparison, the AR condition scored significantly higher on the System Usability Scale (with large effect sizes on overall usability, confidence, learnability, ease of use, and reduced cumbersomeness), significantly lower on NASA-TLX frustration and mental demand, and significantly higher on every creativity subscale measured, including rapid idea generation and creative insight. The authors further report that in the exploratory study, expert and non-expert users experienced similarly low workload, and that users' placement of audio objects followed semantically meaningful spatial patterns, such as placing birds high, the river on the floor, and frogs near the river. These results are used to argue that embodied, proprioceptive interaction, rather than point-and-click mapping, makes immersive mix creation more intuitive.","pith_inferences":["The paper's quantitative comparison omits a third condition: full 6DoF with translation. If embodied motion itself, rather than rotational head tracking plus gesture, drives the gains, future 6DoF-enabled versions might show even larger benefits, but the current data cannot distinguish these effects.","The spatial clustering results (birds above, river on the floor, frogs near the river) suggest that 6DoF AR mixing may be especially valuable for diegetic sound-design tasks rooted in real-world spatial semantics, and a testable design guideline would be to match the mixing space to the represented scene.","One implicit consequence is that the same interface could let users walk an auditorium and verify spatial cues from multiple seats for loudspeaker deliveries, a use case the paper only sketches."],"forward_implications":["A head-mounted AR mixing interface can be integrated with existing DAWs, letting producers switch between desktop editing and in-situ spatial manipulation in the execution environment.","Users can audition a spatial mix by physically walking through it from multiple listener positions, reducing reliance on sweet-spot listening in post-production.","The two design lessons give concrete starting points for designing XR audio-authoring tools: exploit proprioception and carefully balance visual and audio representations of sound.","Even novice users with no DAW background can produce spatially arranged mixes with lower mental workload than with a familiar desktop panner."],"supporting_citations":[{"why":"Supplies the NASA-TLX workload instrument used in both studies to measure mental demand, frustration, and effort.","marker":"[31]"},{"why":"Supplies the System Usability Scale used in Study 2 to compare perceived usability between the AR interface and the 2D panner.","marker":"[13]"},{"why":"Provides the definition of immersion used in the task instructions given to participants in both studies.","marker":"[2]"},{"why":"Supplies the sound design principles (layering, spatial awareness, Gestalt grouping) that the design lessons build on.","marker":"[25]"},{"why":"Provides AR design heuristics that guided the interface design and later interpretation of participant behavior.","marker":"[23]"},{"why":"Documents the 2D DAW workflow that AudioMiXR is contrasted against, motivating the claim that desktop tools constrain spatial awareness.","marker":"[47]"},{"why":"Surveys limitations of existing 3D audio authoring tools, supporting the need for new interaction paradigms like 6DoF.","marker":"[49]"},{"why":"Provides evidence that walking in 6DoF develops more accurate cognitive maps, used to motivate the interface's embodied interaction model.","marker":"[60]"},{"why":"Shows that walking improves cognitive mapping in large-scale environments, cited to justify the 6DoF design approach.","marker":"[62]"}],"fun_headline_variants":["6DoF AR audio mixing: better usability, lower workload","AR 6DoF audio design cuts mental load, boosts creativity","6DoF AR audio mixing: less frustration, more creativity","Spatial audio in AR: 6DoF beats 2D panner on usability"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The measured usability and creativity gains are attributed to 6DoF interaction, yet Study 2 deliberately restricted the AR condition to 3DoF rotation to match the 2D panner's directional capabilities, so the argument assumes that the observed advantages of the restricted AR condition still hold for full 6DoF translation.","fun_headline_variants_meta":{"raw":{"variants":["6DoF AR audio mixing: better usability, lower workload","AR 6DoF audio design cuts mental load, boosts creativity","6DoF AR audio mixing: less frustration, more creativity","Spatial audio in AR: 6DoF beats 2D panner on usability"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001099,"raw_usage":{"total_tokens":4662,"prompt_tokens":1097,"completion_tokens":3565,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":713,"completion_tokens_details":{"reasoning_tokens":3486}},"tokens_in":713,"tokens_out":3565,"duration_ms":22962,"temperature":1.0,"reasoning_tokens":3486,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T10:34:32.750162+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a second controlled comparison with three conditions — a 2D panner, an AR headset restricted to 3DoF rotation, and an AR headset with unrestricted 6DoF translation and rotation — holding all audio stimuli, headset, and task identical. If the 3DoF and 6DoF conditions produce the same SUS, NASA-TLX, and creativity scores, then the gains are not caused by 6DoF translation, and the paper's central attribution would be falsified.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the NASA-TLX workload instrument used in both studies to measure mental demand, frustration, and effort."},{"cited_title":"1996.SUS: A quick and dirty usability scale","cited_arxiv_id":null,"evidence_quote":"Supplies the System Usability Scale used in Study 2 to compare perceived usability between the AR interface and the 2D panner."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the definition of immersion used in the task instructions given to participants in both studies."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the sound design principles (layering, spatial awareness, Gestalt grouping) that the design lessons build on."},{"cited_title":"Endsley, Kelly A","cited_arxiv_id":null,"evidence_quote":"Provides AR design heuristics that guided the interface design and later interpretation of participant behavior."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Surveys limitations of existing 3D audio authoring tools, supporting the need for new interaction paradigms like 6DoF."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides evidence that walking in 6DoF develops more accurate cognitive maps, used to motivate the interface's embodied interaction model."}],"review_version":1}