Implement the complete physical chirp Fourier consumer

[jmbarrancoml]

I complete source08's physical Q_freq F_t chirp consumer after PR3709. I generate the numerical kernel from encoded residues and dyadic phase powers, execute both endpoint products, and run the complete twist/ring-convolution/exact-narrowing/untwist sequence in between.

For the declared p-grid contract, input norm at most .6 and p-A>=13, I return error below (4sqrt(2)M+22)*2^-p and norm below .601. I preserve the natural tensor order and borrowed ports, consume protected child returns and restore the fixed tape owners. I charge the phase and consumer work at O(CT(ell+1)p^(8/5)+rp^3+(A+p+log(C+1))^3), in addition to three complete inherited Fourier calls. Under the stated outer hypotheses, the added power .603 fits the corrected full Fourier power .99827619697.

I build on Chafik Boukhalfa's PR479 additive residue schedule and Ross Richardson's PR948 symbolic checks. I add their encoded physical lowering, certified numerical phase evaluation and the whole consumer. I retain corrected PR3709 at 62264289adb01f3ba4930cd0154c75887f96203c, its nested attribution and source07/source08 notices.

I started from canonical main 20ecf2891d80b5acf18d980b2164fb44ea7faa90, priority revision 6, and recorded the expansion gate and fresh overlap in Discussion3664. This PR changes only research/physical-chirp-middle/. Arham Tahir's separate PR3709 audit remains pending; I claim no acceptance result from it.

Validation at 6bb7dfdae182a69f7c8836a9ec1be0b2b747e124:

• Seven registered suites passed locally, including the full p16 consumer against staged and dense exact Fourier models, 48 certified phase values, complete odometer cubes, a 40-axis prefix and 13 fault controls.
• I executed the full packet with the original reduced q2 setup and with an exact specialization of three certified metadata values. I obtained the same 288-bit payload. The registered specialization preserves the physical data, child, normalization and restoration execution; EVIDENCE.md explains its scope and exact certificates.
• Metadata, source-history and committed-byte checks passed. I replayed all seven commands outside Git on Darwin arm64, Python 3.12.13, and verified that all 30 regular files, including the receipt, remained unchanged. The complete middle-map check took 31.242 seconds.
• Official export SHA256: e769b750eb650c4b2f7822a6e4780aa1b360c207ff9a724caa6dd2555ab3f54f; receipt SHA256: dca075b93c723087c167f01c1074031eeeed2d09d9eeb877481b735a0cecb495. Hosted CI39161347031 passed all seven suites and the unchanged-input check. I matched its merge tree to the published head tree 0533e11436d5036bfa04a616bf370947637156c6.

Read PROOF.md and EVIDENCE.md for the argument and finite coverage. Reproduce with python3.12 scripts/research_package.py --package research/physical-chirp-middle.

I retain the inherited Fourier supplier/refinement premises. Production GL64/O73 execution and the Gaussian expansion/compression, CRT and final integer-recovery join remain open. I leave the selected frontier unchanged. Jose Manuel (jmbarrancoml), with substantial OpenAI Codex assistance, including source-analysis agents; I claim no independent human review or formal-kernel verification.