FAFranck Aumont

Franck Aumont

*This is an algorithmically generated author record based on publicly available information.

Franck Aumont has a total of 14 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    OLED’s power attenuation on streaming context

    by Emmanuel Sampaio, Franck Aumont, Olivier Le Meur, Thomas Maugey

    Rising power demand from video streaming heightens the need to reduce energy use across the end-to-end streaming chain [2], where end-user devices, especially the display, dominate consumption [3]. Display energy can be adapted through pixel intensity [4], motivating pixel-value manipulation methods such as Deep Pixel Value Reduction (DeepPVR) [4]. DeepPVR estimates a per-pixel attenuation map that is applied linearly to the luminance signal, reducing display rendering energy while maintaining visual quality. To limit the computational cost of the client devices, the attenuation map is generated on the server side and transmitted alongside the original content, requiring efficient map compression to optimize the total bitrate. This paper evaluates transmitting lower-resolution, low-bitrate attenuation maps, encoding them using high quantization parameter (QP), and streaming them alongside the original video encoded at its native resolution. At the client, the attenuation map is decoded, uniformly dequantized and upsampled to the decoded video resolution before applying the attenuation, as presented in the Figure 1. Experiments were conducted on a subset of 10 ten-second 4K (3840×2160) sequences from Waterloo4K [5]. Attenuation maps were computed from the luminance (Y) component of the YUV videos. For this test we consider no compression of the videos, while the attenuation maps were uniformly quantized from 32-bit to 8-bit, downsampled using bicubic downsampling and encoded with HEVC (H.265) at QP = 51. We tested three reduced map resolutions (256 × 144, 1280 × 720, 1920 × 1080 and 3840 × 2160). Figure 1 summarizes the end-to-end pipeline. After decoding, the attenuation map is upsampled, dequantized, and applied over the original video content. We evaluate the impact of attenuation map compression along three axes: (i) output quality, using objective metrics (VMAF and Y-PSNR); (ii) power rate reduction, which measures the ratio between the difference of power consumption against the original power. This value was accessed both using a luminance power model [4], and real measures conducted on an TV LG OLED42C35LA using an inline power meter; and (iii) bitrate overhead, to quantify the cost of signaling the attenuation information. For bitratate overead calculus, we compare the bitrate of the content encoded in its original resolution (3840x2160) using HEVC with target bitrate equal to the bitrate used by the HLS bitrate ladder [1] (16.8 Mbps). Results in Table 1 indicate that stronger attenuation-map compression increases objective quality (VMAF, Y-PSNR) while slightly reducing theoretical power-rate reduction, suggesting a modest trade-off between visual quality and energy-saving potential. A subjective study is still required to validate these trends. As expected, the bitrate overhead decreases with compression, reaching only 0.08% when transmitting a 144p map alongside 4K content. We also observe a large gap between measured PRR and theoretical TPR, likely related to the luminance model used in DeepPVR [4]. Finally, variations in measured power reduction across resolutions indicate that the attenuation-map resolution should be selected based on the target power-reduction level.

  • Open Access

    End-to-end VVC encoding for energy-aware HDR rendering

    by Olivier Le Meur, Franck Aumont, Thomas Guionnet, Frédéric Plissonneau, David Touze, Pierre-Loup Cabarat, Daniel Ménard, Juan-Carlos Vargas, Oussama Hammami, Bill Redmann


  • Open Access

    Energy-aware images: Quality of Experience vs Energy Reduction

    by Olivier Le Meur, Claire-Hélène Demarty, Franck Aumont, Laurent Blondé, Erik Reinhard

    The production, transmission and display of video content requires significant amounts of energy. Whether broadcasting or streaming a video, its display on modern televisions is responsible for a significant proportion of the energy consumption. This paper proposes a framework for analysing and processing video frames that allows a modern screen to use less energy when displaying these frames. The content is analyzed prior to encoding and transmission, generating metadata that is attached to the content. After reception by a display, the metadata is then used along with display parameters and user settings to adapt the content prior to display. Two use cases are discerned: one whereby the energy is reduced as much as possible under the constraint that visual quality is maintained, and another where the highest visual quality is sought under the constraint of a fixed reduction of energy use. These use cases would well serve broadcasting and streaming services, respectively.