Will be rolling it out in a month or two.
Plays natively on Quest 3, Quest 3s and the newest Nvidia.
Because encoding takes two three times longer than usual (up to two weeks for 8K 60fps) these will be released later on additionally so we don't have to wait.

Have to check on AVP and Pico.

AV1

Indoors:
https://insights.sexlikereal.com/videos/AV1/46865_1080p.mp4
https://insights.sexlikereal.com/videos/AV1/46865_1440p.mp4
https://insights.sexlikereal.com/videos/AV1/46865_1920p.mp4
https://insights.sexlikereal.com/videos/AV1/46865_2160p.mp4
https://insights.sexlikereal.com/videos/AV1/46865_2880p.mp4
https://insights.sexlikereal.com/videos/AV1/46865_3360p.mp4
https://insights.sexlikereal.com/videos/AV1/46865_3840p.mp4
https://insights.sexlikereal.com/videos/AV1/46865_4000p.mp4

Outdoors:
https://insights.sexlikereal.com/videos/AV1/46230_1080p.mp4
https://insights.sexlikereal.com/videos/AV1/46230_1440p.mp4
https://insights.sexlikereal.com/videos/AV1/46230_1920p.mp4
https://insights.sexlikereal.com/videos/AV1/46230_2160p.mp4
https://insights.sexlikereal.com/videos/AV1/46230_2880p.mp4
https://insights.sexlikereal.com/videos/AV1/46230_3360p.mp4
https://insights.sexlikereal.com/videos/AV1/46230_3840p.mp4
https://insights.sexlikereal.com/videos/AV1/46230_4000p.mp4

h265

Indoors:
https://s3.sexlikereal.com/videos_app/h265/46865_1080p.mp4
https://s3.sexlikereal.com/videos_app/h265/46865_1440p.mp4
https://s3.sexlikereal.com/videos_app/h265/46865_1920p.mp4
https://s3.sexlikereal.com/videos_app/h265/46865_2160p.mp4
https://s3.sexlikereal.com/videos_app/h265/46865_2880p.mp4
https://s3.sexlikereal.com/videos_app/h265/46865_3360p.mp4
https://s3.sexlikereal.com/videos_app/h265/46865_3840p.mp4
https://s3.sexlikereal.com/videos_app/h265/46865_4000p.mp4

Outdoors:
https://s3.sexlikereal.com/videos_app/h265/46230_1080p.mp4
https://s3.sexlikereal.com/videos_app/h265/46230_1440p.mp4
https://s3.sexlikereal.com/videos_app/h265/46230_1920p.mp4
https://s3.sexlikereal.com/videos_app/h265/46230_2160p.mp4
https://s3.sexlikereal.com/videos_app/h265/46230_2880p.mp4
https://s3.sexlikereal.com/videos_app/h265/46230_3360p.mp4
https://s3.sexlikereal.com/videos_app/h265/46230_3840p.mp4
https://s3.sexlikereal.com/videos_app/h265/46230_4000p.mp4

    So this makes encoding faster?
    How much ram do you guys have for encoding?

    Nice and you didn't even mention the file size difference.

    46865_4000p h265 - 485MB
    46865_4000p AV1 - 225MB

    AV1 is about half the size here. Pretty big savings.

    Are you planning to encode older videos in AV1 too?

      tinderbox4357 Nice and you didn't even mention the file size difference.

      What benefits are you expecting from this? For most customers, it won't make a difference for streaming because their internet connections never were a bottleneck or metered. AV1 would have been an opportunity to offer better image quality to customers at the same file sizes as before. But as it seems to be implemented now (i.e. bitrate - how low can you go without notably deteriorating image quality), only SLR and maybe some people with shitty internet will benefit.

        SchnuppiLilac AV1 is meant to be able to compress around 30% better than H265, but this AV1 8K sample file is 50% of the H265 size. I haven't had a closer look yet but this is almost certainly a quality downgrade.

          SchnuppiLilac

          I think the move to AV1 is pretty great for all of us, not just those with slow internet.

          First off, it's great to see companies adopting royalty-free codec. HEVC licensing has been problematic - for instance, Firefox still struggles with HEVC decoding due to these issues. (https://connect.mozilla.org/t5/ideas/hevc-support/idi-p/18643)

          While I'm totally fine with the current quality of SLR videos, AV1 opens up some cool possibilities. Even if the quality stays the same, the smaller file sizes mean SLR can store and serve up more content without breaking the bank. That could lead to potentially higher quality options down the line without massive cost increases.

          Plus, AV1 generally handles certain types of content better than HEVC, like fast motion or complex scenes. So we might see some improvements even at similar bitrates.

          Sure, not everyone needs smaller files for streaming, but it's still a win-win. SLR saves on bandwidth and storage, and that savings could translate to better content or features for us in the future.

          Hairsational

          You're right to be cautious about quality, but I think there's more to the story here.

          The 50% difference you noticed might be a bit misleading due to the specific sample. Different scenes compress differently, and static backgrounds tend to compress really well with AV1.

          For a more balanced comparison, check out this outdoor scene where the background isn't static:

          AV1 46230_4000p - 464MB
          HEVC 46230_4000p - 437MB

          As you can see, the file sizes are much closer here, with AV1 being about 6% larger.

          AV1 tends to handle complex scenes and motion better than HEVC, so the quality difference might actually be negligible or even slightly better in some cases, despite the smaller file size.

          It's also worth noting that encoders are constantly improving, so we might see even better results as AV1 encoding is optimized further.

            tinderbox4357 Thanks for the detailed response. Regarding cost reduction: I guess this is necessary in these days to keep the whole thing going. But as far as I am aware, Q2 does not support AV1 and I would expect that the Q2 customer base will stay relevant for a while, so in the mid-term while both formats need to coexist I wouldn't expect that we will get an additional SLRO release per week.

            tinderbox4357 As you can see, the file sizes are much closer here, with AV1 being about 6% larger.

            When I checked it out, I think the bitrate of the AV1 was a bit higher than in the HEVC, like 29 vs. 26.

            doublevr
            I think AV1 is great but these encodes are not good. They look less fluid than the x265 version.

            I confirmed I was getting 75FPS in the headset and 60FPS for the video playback but motion was inferior to the x265 version.

            Only tested the outdoor scene:
            The av1 one doesnt work well on the Apple Vision pro, stutters extremely. I just googled, the AVP doesnt have an AV1 hardware decoder, damn Apple!!!

            It definitely looks sharper though, best seen when looking at that tree, the quality is amazing.

              harigeharry It definitely looks sharper though, best seen when looking at that tree, the quality is amazing.

              So what you're saying is SLR's current quality level is too low 🙂

              LOL. The SLR originals already have an amazing detail of course. I just did an A/B test with a screenshot, pausing at the same second, to make sure I wasn't seeing stuff that isn't there. But yeah the AV1 does seem to have some more details:

              That being said, while paused at the same second they're most likely not paused at the same frame. To do the test correctly it's probably better to just view the images in VLC on the PC for example and do such test again while pausing at the exact same frame.

                harigeharry Don't you notice the motion in the AV1 video to look less fluid?

                You seem to be showing 2 different versions of a not detailed image. Different but not really better. If you want to make a comparison you need to choose things that are actually discernible.

                  onyxone Like said, the Apple Vision pro doesnt have an AV1 hardware decoder so the video stutters extremely, so I can't say anything about how fluid it is. For $3500 Apple couldnt include that decoder, LOL. It's hard to love that company.

                  I'm not sure what you mean with 'things that are actually discernible', this obviously is an extremely zoomed in part of the video where the HVEC part compressed to less details. But again, to do a fair comparison one should compare the same frames, so not sure how valid the above is.

                    harigeharry I missed that u did not have hardware decoding, that explains why you would not notice the motion difference.

                    Differing compression can create different lossy artifacts. So unless you have something to discern such as words that are legible you can't tell which truly has more detail. You could take the same smeary image and apply sharpness or grain to it and create an illusion of detail that is not actually real. So you need something you can actually recognize to tell when real detail is present or not.

                    21 days later

                    Also some GPU encodings to benchmark v CPU

                    Indoors NVEnc:
                    https://insights.sexlikereal.com/videos/AV1/46865_nvenc_1080p.mp4	42,098,430
                    https://insights.sexlikereal.com/videos/AV1/46865_nvenc_1440p.mp4	62,043,273
                    https://insights.sexlikereal.com/videos/AV1/46865_nvenc_1920p.mp4	93,181,881
                    https://insights.sexlikereal.com/videos/AV1/46865_nvenc_2160p.mp4	111,875,567
                    https://insights.sexlikereal.com/videos/AV1/46865_nvenc_2880p.mp4	176,731,166
                    https://insights.sexlikereal.com/videos/AV1/46865_nvenc_3360p.mp4	234,348,952
                    https://insights.sexlikereal.com/videos/AV1/46865_nvenc_3840p.mp4	304,474,495
                    https://insights.sexlikereal.com/videos/AV1/46865_nvenc_4000p.mp4	343,393,007
                    
                    Outdoors NVEnc:
                    https://insights.sexlikereal.com/videos/AV1/46230_nvenc_1080p.mp4	83,424,575
                    https://insights.sexlikereal.com/videos/AV1/46230_nvenc_1440p.mp4	130,035,446
                    https://insights.sexlikereal.com/videos/AV1/46230_nvenc_1920p.mp4	200,700,562
                    https://insights.sexlikereal.com/videos/AV1/46230_nvenc_2160p.mp4	242,188,941
                    https://insights.sexlikereal.com/videos/AV1/46230_nvenc_2880p.mp4	377,329,433
                    https://insights.sexlikereal.com/videos/AV1/46230_nvenc_3360p.mp4	496,656,136
                    https://insights.sexlikereal.com/videos/AV1/46230_nvenc_3840p.mp4	622,643,983
                    https://insights.sexlikereal.com/videos/AV1/46230_nvenc_4000p.mp4	677,374,077

                    harigeharry could you have picked a more out of focus "sample" dear lord do an AB test with something in focus at least

                    rerun119 Files encoded on a GPU by Nvidia's hardware encoder for comparison with the files in the original post that were encoded on a CPU using unspecified software.

                    Problem with that is that any comparisons people do here are likely going to be both subjective and inaccurate.

                    15 days later