Papers › Neural Video Compression with Feature Modulation

Neural Video Compression with Feature Modulation

27 Feb 2024CVPR 2024 1arXiv:2402.17414archive 2025-07-28

Jiahao Li, Bin Li, Yan Lu

The emerging conditional coding-based neural video codec (NVC) shows superiority over commonly-used residual coding-based codec and the latest NVC already claims to outperform the best traditional codec. However, there still exist critical problems blocking the practicality of NVC. In this paper, we propose a powerful conditional coding-based NVC that solves two critical problems via feature modulation. The first is how to support a wide quality range in a single model. Previous NVC with this capability only supports about 3.8 dB PSNR range on average. To tackle this limitation, we modulate the latent feature of the current frame via the learnable quantization scaler. During the training, we specially design the uniform quantization parameter sampling mechanism to improve the harmonization of encoding and quantization. This results in a better learning of the quantization scaler and helps our NVC support about 11.4 dB PSNR range. The second is how to make NVC still work under a long prediction chain. We expose that the previous SOTA NVC has an obvious quality degradation problem when using a large intra-period setting. To this end, we propose modulating the temporal feature with a periodically refreshing mechanism to boost the quality. %Besides solving the above two problems, we also design a single model that can support both RGB and YUV colorspaces. Notably, under single intra-frame setting, our codec can achieve 29.7% bitrate saving over previous SOTA NVC with 16% MACs reduction. Our codec serves as a notable landmark in the journey of NVC evolution. The codes are at https://github.com/microsoft/DCVC.

PaperPDFConference PDFCodeCode Syntology ranMethod as code

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 43 code samples harvested from 1 repository linked to this paper: 36 ran (of which 17 constructed an object rather than computing a result; 26 with no instrument failure: 4 honoured, 0 violated, 22 with no contract checked; 10 where Syntology's instrument failed) · 7 unverified. Where Syntology's instrument failed: 10 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 43 samples from 1 repository, 36 ran (26 with no instrument failure; 10 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

microsoft/dcvc officialmentioned in papermentioned on GitHubpytorchSyntology: 43 samples, 36 ran (10 instrument-failed) · 7 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

43 samples harvested: 36 ran (of which 17 constructed an object rather than computing a result; 26 with no instrument failure: 4 honoured, 0 violated, 22 with no contract checked; 10 where Syntology's instrument failed) · 7 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

4ran · honoured contract
10ran · our draft was wrong
22ran
7unverified

Licence: 0 of the 43 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from microsoft/dcvc. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. For 8 samples that ran, the graph records a signature this page does not show: for a class the fixture built without calling, the harness called the object with inputs it made up itself, so the recorded signature is the harness's probe, not the fixture's arguments. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

microsoft/dcvc official repository 43 samples, 36 ran (10 instrument-failed) · 7 unverified
AEHelper DCVC-family/DCVC-FM/src/models/video_model.py ran MIT (permissive) · f82bf9fe9efb4ab4 · report fixture: signature quantized_cdf:Tensor/2/int, cdf_length:Tensor/1/int, offset:Tensor/1/int · returned NoneType · level recorded 2026-09-24
BitEstimator DCVC-family/DCVC-FM/src/models/video_model.py ran MIT (permissive) · 379a90fedf952408 · report fixture: returned BitEstimator · level recorded 2026-09-22
Bitparm in 5 papers DCVC-family/DCVC-FM/src/models/video_model.py ran MIT (permissive) · c4a3ec7db0af2724 · report fixture: returned Bitparm · level recorded 2026-09-22
ContextualDecoder DCVC-family/DCVC-FM/src/models/video_model.py ran MIT (permissive) · fca5b37ad0d391c2 · report fixture: returned ContextualDecoder · level recorded 2026-09-22
ContextualEncoder DCVC-family/DCVC-FM/src/models/video_model.py ran fingerprinted MIT (permissive) · 310fe7d061e997cf · report fixture: signature x:Tensor/4/float, context1:Tensor/4/float, context2:Tensor/4/float, context3:Tensor/4/float, quant_step:Tensor/0/float · returned Tensor · level recorded 2026-09-24
ConvFFN DCVC-family/DCVC-FM/src/models/video_model.py ran fingerprinted MIT (permissive) · 419be1df805c2172 · report fixture: returned ConvFFN · level recorded 2026-09-22
ConvFFN3 DCVC-family/DCVC-FM/src/models/video_model.py ran fingerprinted MIT (permissive) · 03692913271d3292 · report fixture: returned ConvFFN3 · level recorded 2026-09-22
DepthConv DCVC-family/DCVC-FM/src/models/video_model.py ran fingerprinted MIT (permissive) · a6721c08436a1879 · report fixture: signature x:Tensor/4/float · returned Tensor · level recorded 2026-09-24
DepthConvBlock DCVC-family/DCVC-FM/src/models/video_model.py ran fingerprinted MIT (permissive) · b14677bd2a05b95f · report fixture: returned DepthConvBlock · level recorded 2026-09-22
DepthConvBlock4 DCVC-family/DCVC-FM/src/models/video_model.py ran fingerprinted MIT (permissive) · 5fcc0a5945f66527 · report fixture: returned DepthConvBlock4 · level recorded 2026-09-22
FeatureExtractor in 24 papers DCVC-family/DCVC-FM/src/models/video_model.py ran MIT (permissive) · 4d085f78c5a59bdc · report fixture: returned FeatureExtractor · level recorded 2026-09-22
GaussianEncoder DCVC-family/DCVC-FM/src/models/video_model.py ran MIT (permissive) · 659877e766994421 · report fixture: returned GaussianEncoder · level recorded 2026-09-22
MEBasic DCVC-family/DCVC-FM/src/models/video_model.py ran MIT (permissive) · fe224c85d7424b84 · report fixture: returned MEBasic · level recorded 2026-09-22
MultiScaleContextFusion DCVC-family/DCVC-FM/src/models/video_model.py ran MIT (permissive) · 320925690ef29c06 · report fixture: returned MultiScaleContextFusion · level recorded 2026-09-22
MvDec DCVC-family/DCVC-FM/src/models/video_model.py ran fingerprinted MIT (permissive) · ddc18f98237531b4 · report fixture: signature x:Tensor/4/float, quant_step:Tensor/0/float · returned tuple · level recorded 2026-09-24
MvEnc DCVC-family/DCVC-FM/src/models/video_model.py ran fingerprinted MIT (permissive) · 0272de7b35cc7549 · report fixture: signature x:Tensor/4/float, context:NoneType/0/?, quant_step:Tensor/0/float · returned Tensor · level recorded 2026-09-24
OffsetDiversity DCVC-family/DCVC-FM/src/models/video_model.py ran MIT (permissive) · d2f0192d643ab535 · report fixture: returned OffsetDiversity · level recorded 2026-09-22
ReconGeneration DCVC-family/DCVC-FM/src/models/video_model.py ran MIT (permissive) · 3d55656b67d7cbf4 · report fixture: returned ReconGeneration · level recorded 2026-09-22
ResBlock in 127 papers
census: 6 implementations from 5 papers, 6 distinct outputs across 2 buckets, one shared input per bucket
DCVC-family/DCVC-FM/src/models/video_model.py ran MIT (permissive) · 75b276630f3fe107 · report fixture: returned ResBlock · level recorded 2026-09-22
ResidualBlockUpsample DCVC-family/DCVC-FM/src/models/video_model.py ran fingerprinted MIT (permissive) · c0ac2daecdabac0b · report fixture: returned ResidualBlockUpsample · level recorded 2026-09-22
ResidualBlockWithStride DCVC-family/DCVC-FM/src/models/video_model.py ran fingerprinted MIT (permissive) · 0b4d24dfa9389a82 · report fixture: returned ResidualBlockWithStride · level recorded 2026-09-22
UNet2 DCVC-family/DCVC-FM/src/models/video_model.py ran MIT (permissive) · a12a5eee4649c19e · report fixture: returned UNet2 · level recorded 2026-09-22
bilineardownsacling DCVC-family/DCVC-FM/src/models/video_model.py ran · our draft was wrong fingerprinted MIT (permissive) · 4ee1913b1f8ce86e · report fixture: signature inputfeature:Tensor/4/float · returned Tensor · level recorded 2026-09-22
bilinearupsacling DCVC-family/DCVC-FM/src/models/video_model.py ran · our draft was wrong fingerprinted MIT (permissive) · 2c8c63d56261cb76 · report fixture: signature inputfeature:Tensor/4/float · returned Tensor · level recorded 2026-09-22
block_mc_func DCVC-family/DCVC-FM/src/models/video_model.py ran · our draft was wrong fingerprinted MIT (permissive) · da82b58aa9f6478c · report fixture: signature im:Tensor/4/float, flow:Tensor/4/float · returned Tensor · level recorded 2026-09-22
conv1x1 in 691 papers DCVC-family/DCVC-FM/src/models/video_model.py ran · our draft was wrong MIT (permissive) · ff496db6b866af12 · report fixture: signature in_ch:int/0/int, out_ch:int/0/int, stride:int/0/int · returned Conv2d · level recorded 2026-09-22
conv3x3 in 1,376 papers DCVC-family/DCVC-FM/src/models/video_model.py ran · our draft was wrong MIT (permissive) · 194462a957fb0991 · report fixture: signature in_ch:int/0/int, out_ch:int/0/int, stride:int/0/int · returned Conv2d · level recorded 2026-09-22
flow_warp in 48 papers DCVC-family/DCVC-FM/src/models/video_model.py ran · our draft was wrong fingerprinted MIT (permissive) · ca37ddb87705e620 · report fixture: signature im:Tensor/4/float, flow:Tensor/4/float · returned Tensor · level recorded 2026-09-22
get_downsampled_shape DCVC-family/DCVC-FM/src/models/video_model.py ran · honoured contract MIT (permissive) · 0c429d93ea07cde7 · report fixture: signature height:int/0/int, width:int/0/int, p:int/0/int · returned tuple · level recorded 2026-09-22
get_padding_size DCVC-family/DCVC-FM/src/models/video_model.py ran · honoured contract MIT (permissive) · 058564027e22df3c · report fixture: signature height:int/0/int, width:int/0/int, p:int/0/int · returned tuple · level recorded 2026-09-22
subpel_conv1x1 DCVC-family/DCVC-FM/src/models/video_model.py ran · our draft was wrong MIT (permissive) · 05d9e671c5f84bdb · report fixture: signature in_ch:int/0/int, out_ch:int/0/int, r:int/0/int · returned Sequential · level recorded 2026-09-22
subpel_conv3x3 in 9 papers DCVC-family/DCVC-FM/src/models/video_model.py ran · our draft was wrong MIT (permissive) · c441c9f39e918e8a · report fixture: signature in_ch:int/0/int, out_ch:int/0/int, r:int/0/int · returned Sequential · level recorded 2026-09-22
torch_warp DCVC-family/DCVC-FM/src/models/video_model.py ran · our draft was wrong fingerprinted MIT (permissive) · 4ed6bbdfa756e1be · report fixture: signature feature:Tensor/4/float, flow:Tensor/4/float · returned Tensor · level recorded 2026-09-22
write_ip DCVC-family/DCVC-FM/src/models/video_model.py ran · honoured contract MIT (permissive) · c8819fe99eeef17c · report fixture: signature f:BytesIO/0/?, is_i_frame:bool/0/bool, sps_id:int/0/int, bit_stream:bytes/0/? · returned int · level recorded 2026-09-22
write_uchars DCVC-family/DCVC-FM/src/models/video_model.py ran · our draft was wrong MIT (permissive) · cde57c4548df7988 · report fixture: signature fd:BytesIO/0/?, values:list/1/int, fmt:str/0/str · returned int · level recorded 2026-09-22
write_uint_adaptive DCVC-family/DCVC-FM/src/models/video_model.py ran · honoured contract MIT (permissive) · 1ec4dcdc3950ddfe · report fixture: signature f:BytesIO/0/?, a:int/0/int · returned int · level recorded 2026-09-22
CompressionModel DCVC-family/DCVC-FM/src/models/video_model.py unverified MIT (permissive) · b04500057b53dc0d · report
DMC DCVC-family/DCVC-FM/src/models/video_model.py unverified MIT (permissive) · d0541ab1ef1f01d7 · report
EntropyCoder DCVC-family/DCVC-FM/src/models/video_model.py unverified MIT (permissive) · 8b1f88062c054aad · report
ME_Spynet DCVC-family/DCVC-FM/src/models/video_model.py unverified MIT (permissive) · 8a52237f88c790c0 · report
NalType DCVC-family/DCVC-FM/src/models/video_model.py unverified MIT (permissive) · b50de84040ba0329 · report
add_grid_cache DCVC-family/DCVC-FM/src/models/video_model.py unverified MIT (permissive) · a3cc1c3e7b0b88aa · report
write_bytes DCVC-family/DCVC-FM/src/models/video_model.py unverified MIT (permissive) · 395e86482196ce43 · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly for Neural Video Compression with Feature Modulation (graph method: Neural Video Compression with Feature Modulation). 43 bricks; individually on a fixture: 36 ran (of which 17 constructed an object rather than computing a result; 26 with no instrument failure: 4 honoured, 0 violated, 22 with no contract checked; 10 where Syntology's instrument failed) · 7 unverified. The assembly has not been run end-to-end.

  1. DMC entry · level 0 · unverified · d0541ab1ef1f01d7
  2. CompressionModel model · level 0 · unverified · b04500057b53dc0d
  3. ContextualDecoder model · level 1 · ran · fca5b37ad0d391c2
  4. ContextualEncoder model · level 1 · ran · 310fe7d061e997cf
  5. DepthConvBlock helper · level 1 · ran · b14677bd2a05b95f
  6. DepthConvBlock4 helper · level 1 · ran · 5fcc0a5945f66527
  7. FeatureExtractor model · level 1 · ran · 4d085f78c5a59bdc
  8. ME_Spynet helper · level 0 · unverified · 8a52237f88c790c0
  9. MultiScaleContextFusion model · level 1 · ran · 320925690ef29c06
  10. MvDec model · level 1 · ran · ddc18f98237531b4
  11. MvEnc model · level 1 · ran · 0272de7b35cc7549
  12. OffsetDiversity model · level 1 · ran · d2f0192d643ab535
  13. ReconGeneration model · level 1 · ran · 3d55656b67d7cbf4
  14. ResidualBlockUpsample helper · level 1 · ran · c0ac2daecdabac0b
  15. bilineardownsacling helper · level 1 · ran · our draft was wrong · 4ee1913b1f8ce86e
  16. block_mc_func helper · level 1 · ran · our draft was wrong · da82b58aa9f6478c
  17. get_downsampled_shape helper · level 1 · ran · honoured contract · 0c429d93ea07cde7
  18. write_ip helper · level 1 · ran · honoured contract · c8819fe99eeef17c
  19. BitEstimator model · level 1 · ran · 379a90fedf952408
  20. EntropyCoder helper · level 0 · unverified · 8b1f88062c054aad
  21. GaussianEncoder model · level 1 · ran · 659877e766994421
  22. get_padding_size helper · level 1 · ran · honoured contract · 058564027e22df3c
  23. subpel_conv3x3 helper · level 1 · ran · our draft was wrong · c441c9f39e918e8a
  24. ConvFFN helper · level 1 · ran · 419be1df805c2172
  25. DepthConv helper · level 1 · ran · a6721c08436a1879
  26. ConvFFN3 helper · level 1 · ran · 03692913271d3292
  27. ResBlock helper · level 1 · ran · 75b276630f3fe107
  28. MEBasic helper · level 1 · ran · fe224c85d7424b84
  29. bilinearupsacling helper · level 1 · ran · our draft was wrong · 2c8c63d56261cb76
  30. subpel_conv1x1 helper · level 1 · ran · our draft was wrong · 05d9e671c5f84bdb
  31. ResidualBlockWithStride helper · level 1 · ran · 0b4d24dfa9389a82
  32. UNet2 helper · level 1 · ran · a12a5eee4649c19e
  33. conv3x3 helper · level 1 · ran · our draft was wrong · 194462a957fb0991
  34. flow_warp helper · level 1 · ran · our draft was wrong · ca37ddb87705e620
  35. NalType helper · level 0 · unverified · b50de84040ba0329
  36. write_bytes helper · level 0 · unverified · 395e86482196ce43
  37. write_uchars helper · level 1 · ran · our draft was wrong · cde57c4548df7988
  38. write_uint_adaptive helper · level 1 · ran · honoured contract · 1ec4dcdc3950ddfe
  39. AEHelper helper · level 1 · ran · f82bf9fe9efb4ab4
  40. Bitparm helper · level 1 · ran · c4a3ec7db0af2724
  41. conv1x1 helper · level 1 · ran · our draft was wrong · ff496db6b866af12
  42. torch_warp helper · level 1 · ran · our draft was wrong · 4ed6bbdfa756e1be
  43. add_grid_cache helper · level 0 · unverified · a3cc1c3e7b0b88aa

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2402.17414") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

BlockingQuantizationVideo Compression

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections