The Time-Domain Lattice in EHT Visibilities: A Glass–Crystal Audit Across Sgr A and M87**
Fuente:
Zenodo
Enregistré dans:
| Auteur principal: | |
|---|---|
| Format: | Recurso digital |
| Publié: |
Zenodo
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866901808623910912 |
|---|---|
| author | Diogenes |
| author_facet | Diogenes |
| contents | <p>## Abstract</p> <p>We present a conservative update of the R^8 real-signal note after direct uvfits re-analysis, a lane-matched no-floor rerun, and a SACGU-based closure-domain audit. The main change is interpretive rather than cosmetic. The direct uvfits feeder remains compatible with structured sparse remapping, and in the stricter lane-matched no-floor test the full source template is preferred over all tested ablations in both baseline and post-veto states. However, the newer analyses do not justify a unique 24-seeded interpretation as the primary evidential claim. The correct present seed-level verdict is therefore **inconclusive**.</p> <p>At the same time, the companion SACGU pipeline remains methodologically valuable. File-level audits built from closure phase and log-closure amplitude recover structured closure-domain signal in the current unique M87 and Sgr A* uvfits files. But after deduplication and consensus gating, the presently loaded set still contains only one independent epoch per dimension, so no source-level consensus claim is yet justified. The resulting position is deliberately conservative: the current data support an **audit-level statement of structured closure-domain behavior**, while the toy harmonic feeder should now be treated as a **robustness / feeder-sensitivity layer**, not as the primary evidential pipeline.</p> <p>## SACGU: What it is and how to use it</p> <p>SACGU (Scattering-Aware Closure Geometric Unblender) should be published as a separate, general-purpose uvfits-domain tool. It is **closure-first**, **source-agnostic**, and **plugin-friendly**. Its core observable layer is **closure phase + log-closure amplitude**, and its inference engine uses **switching-state latent modeling**, **null doctrine**, and **synthetic injections**.</p> <p>### Minimal practical workflow</p> <p>1. Start from raw uvfits files and preserve the native time, frequency, and lane/band structure.<br>2. Add or reconstruct scan structure when needed, and reorder the array before extracting minimal closure sets.<br>3. Build maximal-independent closure-phase and log-closure-amplitude bases on the **available graph** for each slice.<br>4. Fit a switching-state latent model in the closure domain rather than a naive linear closure mixture.<br>5. Run the null battery first: time-scramble, surrogate, and gate-destruction controls.<br>6. Run synthetic injections to verify that the apparent structure is recoverable.<br>7. Only then compare lanes, pipelines, epochs, or bands and apply consensus gates.<br>8. Keep any toy harmonic feeder as an appendix-level diagnostic unless it survives explicit complexity penalties and repeatability tests.</p> <p>## Current take-away</p> <p>The present uvfits set is enough to demonstrate that SACGU can recover structured closure-domain signal on real data, but it is not yet enough to establish source-level consensus or a mechanism-level claim. The next meaningful advance requires **independent repeats** and, where possible, **same-day multi-band confirmation**. </p> <p>## Abstract</p> <p>This release closes the present R^8 search for cosmological-signal signatures in M87 with a deliberately conservative and observable-driven interpretation. We revisited the real-data analysis using direct uvfits reprocessing, penalized feeder tests, lane-matched no-floor reruns, SACGU-style closure-domain audits, band-by-band sweeps across the 2018 EHT data, and forward toy detectability studies on the exact 2018 uv coverage. The main outcome is not a new positive detection claim, but a clarification of what the current data can and cannot support.</p> <p>Across the stricter reanalyses, the earlier stronger 24-seed narrative does not survive. The current real-data sweeps do not justify a unique 24-seeded interpretation, do not show a stable source-level mechanism claim, and do not provide repeatable consensus across the presently examined M87 files alone. At the same time, the data are not trivial: SACGU-style audits still recover structured closure-domain behavior in selected files and lanes, showing that the uvfits carry real geometric information at the robustness and audit level. The correct present claim is therefore limited: the current M87 data are sufficient for audit-level statements about structured closure-domain behavior, but not for a robust cosmological-signal detection claim.</p> <p>A key result of this release is that the observable itself matters. For a strictly centered even-m ring perturbation, closure phase is the wrong primary detector. A forward toy built from a thin ring with centered m=24 modulation, sampled on the exact 2018 M87 uv coverage, produces no measurable closure-phase response and no measurable log-closure-amplitude response in the present toy setup. This does not rule out theory-side structure at the source. Instead, it shows that the presently tested perturbation class is not accessible to the 2018 EHT observable chain in a straightforward way. The feeder-layer harmonic diagnostics therefore remain useful as exploratory appendix-level probes, while the SACGU framework remains the correct general uvfits-domain method note for closure phase, log-closure amplitude, switching-state inference, null doctrine, and synthetic injections.</p> <p>The resulting interpretation is best read as an observational sensitivity bound: the current dataset constrains what the 2018 EHT could have seen, rather than establishing that no source-side structure exists. Two independent suppression mechanisms — symmetry and Bessel suppression — render the 2018 EHT insensitive to centered m=24 ring perturbations. This is an observational bound, not a theoretical failure.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_19144552 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | The Time-Domain Lattice in EHT Visibilities: A Glass–Crystal Audit Across Sgr A and M87** Diogenes Crystalline Axiverse <p>## Abstract</p> <p>We present a conservative update of the R^8 real-signal note after direct uvfits re-analysis, a lane-matched no-floor rerun, and a SACGU-based closure-domain audit. The main change is interpretive rather than cosmetic. The direct uvfits feeder remains compatible with structured sparse remapping, and in the stricter lane-matched no-floor test the full source template is preferred over all tested ablations in both baseline and post-veto states. However, the newer analyses do not justify a unique 24-seeded interpretation as the primary evidential claim. The correct present seed-level verdict is therefore **inconclusive**.</p> <p>At the same time, the companion SACGU pipeline remains methodologically valuable. File-level audits built from closure phase and log-closure amplitude recover structured closure-domain signal in the current unique M87 and Sgr A* uvfits files. But after deduplication and consensus gating, the presently loaded set still contains only one independent epoch per dimension, so no source-level consensus claim is yet justified. The resulting position is deliberately conservative: the current data support an **audit-level statement of structured closure-domain behavior**, while the toy harmonic feeder should now be treated as a **robustness / feeder-sensitivity layer**, not as the primary evidential pipeline.</p> <p>## SACGU: What it is and how to use it</p> <p>SACGU (Scattering-Aware Closure Geometric Unblender) should be published as a separate, general-purpose uvfits-domain tool. It is **closure-first**, **source-agnostic**, and **plugin-friendly**. Its core observable layer is **closure phase + log-closure amplitude**, and its inference engine uses **switching-state latent modeling**, **null doctrine**, and **synthetic injections**.</p> <p>### Minimal practical workflow</p> <p>1. Start from raw uvfits files and preserve the native time, frequency, and lane/band structure.<br>2. Add or reconstruct scan structure when needed, and reorder the array before extracting minimal closure sets.<br>3. Build maximal-independent closure-phase and log-closure-amplitude bases on the **available graph** for each slice.<br>4. Fit a switching-state latent model in the closure domain rather than a naive linear closure mixture.<br>5. Run the null battery first: time-scramble, surrogate, and gate-destruction controls.<br>6. Run synthetic injections to verify that the apparent structure is recoverable.<br>7. Only then compare lanes, pipelines, epochs, or bands and apply consensus gates.<br>8. Keep any toy harmonic feeder as an appendix-level diagnostic unless it survives explicit complexity penalties and repeatability tests.</p> <p>## Current take-away</p> <p>The present uvfits set is enough to demonstrate that SACGU can recover structured closure-domain signal on real data, but it is not yet enough to establish source-level consensus or a mechanism-level claim. The next meaningful advance requires **independent repeats** and, where possible, **same-day multi-band confirmation**. </p> <p>## Abstract</p> <p>This release closes the present R^8 search for cosmological-signal signatures in M87 with a deliberately conservative and observable-driven interpretation. We revisited the real-data analysis using direct uvfits reprocessing, penalized feeder tests, lane-matched no-floor reruns, SACGU-style closure-domain audits, band-by-band sweeps across the 2018 EHT data, and forward toy detectability studies on the exact 2018 uv coverage. The main outcome is not a new positive detection claim, but a clarification of what the current data can and cannot support.</p> <p>Across the stricter reanalyses, the earlier stronger 24-seed narrative does not survive. The current real-data sweeps do not justify a unique 24-seeded interpretation, do not show a stable source-level mechanism claim, and do not provide repeatable consensus across the presently examined M87 files alone. At the same time, the data are not trivial: SACGU-style audits still recover structured closure-domain behavior in selected files and lanes, showing that the uvfits carry real geometric information at the robustness and audit level. The correct present claim is therefore limited: the current M87 data are sufficient for audit-level statements about structured closure-domain behavior, but not for a robust cosmological-signal detection claim.</p> <p>A key result of this release is that the observable itself matters. For a strictly centered even-m ring perturbation, closure phase is the wrong primary detector. A forward toy built from a thin ring with centered m=24 modulation, sampled on the exact 2018 M87 uv coverage, produces no measurable closure-phase response and no measurable log-closure-amplitude response in the present toy setup. This does not rule out theory-side structure at the source. Instead, it shows that the presently tested perturbation class is not accessible to the 2018 EHT observable chain in a straightforward way. The feeder-layer harmonic diagnostics therefore remain useful as exploratory appendix-level probes, while the SACGU framework remains the correct general uvfits-domain method note for closure phase, log-closure amplitude, switching-state inference, null doctrine, and synthetic injections.</p> <p>The resulting interpretation is best read as an observational sensitivity bound: the current dataset constrains what the 2018 EHT could have seen, rather than establishing that no source-side structure exists. Two independent suppression mechanisms — symmetry and Bessel suppression — render the 2018 EHT insensitive to centered m=24 ring perturbations. This is an observational bound, not a theoretical failure.</p> |
| title | The Time-Domain Lattice in EHT Visibilities: A Glass–Crystal Audit Across Sgr A and M87** |
| topic | Crystalline Axiverse |
| url | https://doi.org/10.5281/zenodo.19144552 |