Diary Of An x264 Developer

05/08/2010 (1:47 pm)

Taking submissions for encoder comparison

Filed under: Uncategorized ::

With VP8 supposedly going to come out in about 2 weeks, it’s time to get a rough idea as to the visual state of the art in terms of encoders.  Accordingly, I’m doing a small visual codec comparison in which we will take a few dozen encoders, encode a single test clip, and perform score-based visual tests on real humans using a blind test.  There will be no PSNR or SSIM results posted.

See the doom9 thread for more information and feel free to submit streams for your own encoders.  I’m particularly interested in some newer proprietary encoders for which I wouldn’t be able to get the software for due to NDAs or similar (such as VP8, Sony Blu-code, etc) — but for which I would be able to get a dump of the decoded output.

17 Responses to “Taking submissions for encoder comparison”

  1. Louise Says:

    You can use Flix Cloud to get your VP8 video.

    http://www.flixcloud.com/

    That is On2′s VP8 cloud encoder.

  2. Dark Shikari Says:

    @Louise

    I don’t see VP8 listed…

  3. Louise Says:

    This this PDF On2 compares VP7 with x264

    http://www.on2.com/file.php?229

  4. Louise Says:

    @Dark Shikari

    Very sorry about that. I was pretty sure I saw it on the format list…

  5. Dark Shikari Says:

    @Louise

    Please stop posting On2 marketing material on my blog. We all know it’s bullshit, including you, and it’s certainly not grounds for an encoder comparison. On2 have spent the last 10 years claiming absurd gains over industry leaders and never actually delivered on any of their promises.

    FYI, in my tests so far, VP7 got beaten pretty badly by every single good H.264 encoder that I’ve tested. It was still pretty respectable though (if you can get past the disastrous level of blurring) and beat basically everything pre-H.264.

  6. Tomoni Says:

    MSU Video Quality Measurement Tools

    Provides objective (“Metrics”) and subjective (“Blind Test”) video quality evaluation. Might be useful (including AviSynth-Support).

  7. Shiretoko Says:

    Why not wait until VP8 is released officially? You can’t just accepted the “decoded output” without knowing what the guy actually used, what settings, and whether he’s competent to do it in the first place.

    I think that’s unfair.
    (btw, I don’t care about On2 or advertising their codecs)

  8. Dark Shikari Says:

    @Shiretoko

    True, there’s no guarantee I can trust said person, but how can you guarantee that I’ll know the “best settings” to use with it either?

    And what if VP8 doesn’t come out as promised by the gossip tech media?

    I could just delay it a week or so and hope that it does come out, in which case I can include it in the main visual comparison (it’s hard to go and add anything to the comparison afterwards, because everyone’s ratings are going to be made in the context of all the other images).

  9. Carl Eugen Says:

    Any reasons you didn’t add FFmpeg MPEG-2?

  10. Dark Shikari Says:

    @Carl

    MPEG-2 is basically the same as MPEG-1 without interlacing support. There are minor differences but it’s not worth bothering.

  11. n Says:

    XVD Video (priority=quality, 2 pass, key interval=250, qp=15)
    http://www.mediafire.com/?mtmzzdzjdgt

    I used trial version encoder of deceased BHA (Japanese company) released in 2003.
    it’s too dark and blurring.

  12. Jeremy Says:

    I don’t know if it’s possible, but what I’d really be interested in seeing are some hardware implementations of H264; specifically those from TI and Analog Devices, for example.

    I can encode you video with a TI DM365, if you’re interested. But what really irks me about the embedded encoder space is these people get away with marketing shit products under the guise of “it’s H264!” when everybody well versed in the industry is acutely aware of the differences in encoder quality.

  13. Marsian Says:

    Using Unreal Live Server v7.0, you can compare one-pass, real-time live encoding of VC1 and x264 codecs.

    As I wrote in another post here, we find VC1 slightly better at low bitrates and x264 better (more detailed picture) at medium-high bitrates

  14. cb Says:

    Are you going to be posting all the results and decompressed videos somewhere? How is the blind testing being done?

    It would be nice if the actual app and command line used for each case was posted so they can be independently verified and checked for correctness.

  15. R.T. Systems Says:

    Any news on this? It’s been a month or two since this was announced — any ETA on the results?

  16. Dark Shikari Says:

    @R.T.

    Very busy at my current company because of upcoming E3. After E3, I’ll get back to this. Harass me if I don’t.

  17. R.T. Systems Says:

    E3 is long over now, and a bit over a month has passed. I understand that you’re probably quite busy, but it’d be nice to know if this is dead.

Leave a Reply