# Real instrument recordings / 13 September 2026 Research inputs and limitations. These MSU-MR comparisons use full original IQ. The gallery displays **SatDump + Telemetry Yield**: standalone native reception also misses some reference frames. ## Meteor M2 / Atlantic pass / 2 August 2021 [Recording by MM3IIG, OARC Wiki](https://wiki.oarc.uk/vhfsatellite-general): original SDR Console PCM16 stereo IQ, 500,000 complex samples/s, 498,192,418 bytes, 249.095 s. Embedded capture time: 10:44:41 UTC; the filename uses local summer time. SHA-256: `527682376b52f54cbb7fc120842a5c1e9d700863e0bd7eae610dc72f057bc211`. SatDump DC blocking is enabled as recommended by the recording author. Reference: 1,627 unique CADUs; standalone native: 1,616, with 101 additional and 112 missed. The explicit union adds 405 complete image packets and nine telemetry packets, with no complete reference packets lost. Channels 1/2/3 fill 12/40/27 empty 112 × 8 tiles within the common image extent. They also add 56/64/64 nonempty coverage rows; the gallery's selected common-extent crops show gap filling, not that extension. [Full verification](meteor-m2-atlantic-20210802-verification.json). Attribution: receiver MM3IIG via OARC Wiki (wiki text CC BY-SA 4.0). No separate recording licence was identified; owner-private gallery access is unchanged. ## Meteor M2-3 / North America / 30 June 2023 [Receiver's post and original IQ link](https://www.reddit.com/r/amateursatellites/comments/14nujca/063023_meteorm2_3/): full PCM8 unsigned stereo IQ WAV, 240,000 complex samples/s, 466,616,364 bytes, 972.117333 s. The capture date is reported in the post title; no exact capture UTC is established. WAV SHA-256: `870a25677a29f32d9c575d4c08566d451f0d51490c8e5068798c0368fadffaa1`. The ZIP was losslessly extracted, without resampling or degrading either arm. Reference: 4,169 unique CADUs; standalone native: 4,062, with 67 additional and 174 missed. The explicit union adds 244 complete image packets and six telemetry packets, losing no complete reference packets. Channels 1/2/3 fill 42/44/60 empty 112 × 8 tiles within the common extent. Their nonempty coverage extensions are 32/32/24 rows, not the gallery's selected common-extent views. Narrow recovered strips remain surrounded by unrecovered gaps: a modest gain, not a complete reconstruction. [Full verification](meteor-m2-3-north-america-20230630-verification.json). The first native attempt received external SIGTERM and produced no accepted output; the successful full retry used the same frozen receiver with four workers. Its 851.34 s and the reference's 16.13 s are workflow timings on a shared VM, not an isolated performance benchmark. The receiver shared the IQ explicitly; the author account is now deleted and no separate recording licence was identified. Owner-private gallery access is unchanged. ## Meteor M2-3 / MSU-MR / LRPT / 9 August 2023 [Original IQ WAV](https://www.satdump.org/files/Meteor-M/LRPT/aang23_m23_baseband_137100000Hz_10-38-27_09-08-2023.wav): 423,223,596 bytes, PCM16 stereo IQ at 150,000 complex samples/s. Complete-file SHA-256: `0cf77d57668e93caaae3f25ac90ff6dba28859791869c98703f2b73a6b65e285`. SatDump 1.2.2 uses the `meteor_m2-x_lrpt` pipeline. An initial automatic-postprocessing failure was replaced with a successful full run with only automatic postprocessing disabled; RF/FEC settings were unchanged. Analog calibration telemetry was unavailable: these are uncalibrated images. Reference: 5,727 unique CADUs; native: 5,767 (73 additional, 33 missed). The union adds 293 complete image packets and seven telemetry packets, without losing complete reference packets. It fills 140 empty 112 × 8 tiles across three channels in their common extent. [Full verification](meteor-m2-3-msumr-verification.json). ## Meteor M2 / original SDRplay IQ [Owner's recording](https://voiceoverman.wixsite.com/noaam2/receive-meteor-m-n2-with-sdrplay): PCM8 stereo WAV, 222,222 complex samples/s, 400,226,684 bytes, 900.510841 s. SHA-256: `047afccdbeaea7c3c9b0997586ebcb84b5b81da5c3e7b0535e4d730c4379a872`. No capture date is inferred from a different file linked on the page. Reference: 6,168 unique CADUs; native: 6,143, with 409 additional and 434 missed. The union adds 1,563 complete image packets and 40 telemetry packets. The three channels extend by 376 rows; channels 1/2/5 contain 328/312/320 nonempty new rows respectively, some incomplete. Almost all of this gain is beyond the previous image extent, not inside its gaps. [Verification](meteor-m2-daylight-verification.json). ## Meteor M2-2 / 3 November 2019 [Daniel Vogelbacher's recording](https://blog.chaospixel.com/linux/2019/12/receive-meteor-satellite-images-with-rtlsdr-gqrx-linux.html): original Gqrx CF32, 144,000 complex samples/s, 485,228,224 bytes, 421.205056 s. SHA-256: `0a285fa745f98e9752686337d62d06c87c629cc3fa9a37304f346d26969797cd`. The source specifies private use; owner-private gallery access is unchanged. The correct profile is OQPSK/NRZ-M, not QPSK. The incorrect initial QPSK run is excluded from comparison. Reference: 3,197 CADUs; native: 3,380 (474 additional, 291 missed, including idle frames). The union adds 1,823 complete image packets and 40 telemetry packets. Channels 1/2/5 add 352/360/368 nonempty rows beyond the original extent. [Verification](meteor-m2-2-daylight-verification.json). ## Shared validation and display rules All comparisons use the unchanged SatDump 1.2.2 MSU-MR renderer with `fill_missing=false`. No retouching, interpolation or resolution increase. All four RS(255,223) lanes verify; this is not an independent CRC. In extended-coverage crops, transparency means an area outside a decoder's output extent, not a received black pixel. Reports retain standalone gains and losses although the sliders show an explicit union. Spectral channels and selected crops are not independent recordings. For smaller downloads, distributed Meteor PNGs use lossless PNG prediction and compression. Their dimensions, bit depth, colour/alpha channels, metadata and every decoded sample remain identical; original research PNGs are unchanged. Verification records retain both original and distributed file hashes. [Compression inventory](image-compression.json) and [independent pixel verification](image-compression-check.json). ## Further IQ sources, not demonstrated gains - [NOAA-19 / AVHRR / HRPT](https://www.satdump.org/files/NOAA/HRPT/NOAA-19_10.09.2021_08-21-16_1698000000Hz.wav): RF64 PCM16 stereo, 3 MS/s, 5,419,565,136 bytes. - [MetOp-B / AVHRR / AHRPT](https://www.satdump.org/files/MetOp/AHRPT/meteosatellites_MetOpB.wav): RF64 PCM16 stereo, 6 MS/s, 10,748,952,656 bytes. - [Aqua / MODIS / direct broadcast](https://www.satdump.org/files/Aqua/DB/Aang23_13-51-57_15000000SPS_1360000000Hz.i8): 12,987,924,480 bytes; filename indicates 15 MS/s and raw i8. Complex-IQ interpretation still requires decoding confirmation. Availability and the initial HTTP Range response were checked for these three sources; complete files were not downloaded. Archive .cadu and .raw16 files are not substitutes for original IQ. No verified SAR, Terra or Aura comparison is claimed. [SatDump reference pipeline documentation](https://docs.satdump.org/md_docs_2pages_2Pipelines.html).