Why does Bitmovin produce this distribution of bitrates? Seems strange?

Looking through the manifest.m3u8 file, this is the distribution of bitrates and resolutions (I double checked them).

Target Bitrate Resolution
15.32 Mbit 2304 x 1440
14.93 Mbit 2304 x 1440
9.18 Mbit 1728 x 1080
8.93 Mbit 1728 x 1080
5.5 Mbit 1440 x 900
5.33 Mbit 1440 x 900
3.32 Mbit 922 x 576
3.19 Mbit 922 x 576
2.19 Mbit 692 x 432
2.09 Mbit 692 x 432
1.2 Mbit 460 x 288
1.12 Mbit 460 x 288
821 kbit 346 x 216
740 kbit 346 x 216
537 kbit 288 x 180
461 kbit 288 x 180
315 kbit 288 x 180
246 kbit 288 x 180

It seems strange to me that there are lots of pairs of bitrates that are very close to each other, and then there will be a big jump to the next one. Not a steady progression. They are all at 60fps too. I think I’d like to change the lower bitrate versions to use less fps. I care more about image quality than having 60 vs 30 frames per second. I’m not sure exactly if I should just use the same resolutions in that case though. The whole reason I wanted to use Bitmovin is that there are lots of moving parts, and I don’t know what an optimal distribution of bitrate, resolution, and fps tuples would be. The output has me questioning if the distribution produced is very optimal, but I’m quite open to the possibility that there are things I’m missing, since I’m not knowledgable in this domain. Anyhow, I’m wondering if I should just re-encode the video using a static ladder and use the same bitrate+resolution settings as produced in the per-title setting, but if I should just delete bitrate versions that are very close to each other and also if I should make the fps lower at the lower bitrate versions (not sure at what threshold). Perhaps I should add a full resolution+full bitrate+full fps version as well. There are 18 videos in the m3u8 manifest and only 9 in the mpd manifest, however, so I don’t know if I should just use all 18 and delete the extras out of the manifest of the mpd file produced if I do this or what. Honestly this is all just a lot of work. That’s why I made another forum post suggesting a hybrid approach between “per title” and “static ladder” strategies to produce the video representations. I would like to be able to modify the settings that the “per title” strategy produces. Maybe we should also have more settings to tweak what the “per title” strategy will produce (like fps vs image quality tradeoffs).

Your input file likely has a high resolution / bitrate and the per-title algorithm likely interpolated a large amount of renditions from very low to the default maximum (15Mbps). It is recommended to have a rather large amount of intermediate renditions, in order for players to choose the most suitable one and have less visible quality switches.

This being said, our per-title algorithm is almost entirely configurable, and you can chose your own constraints including the step between 2 renditions, as well as the max bitrate (among other settings)

I hope this helps. Please let me know in case you have any further questions.

This topic was automatically closed 60 minutes after the last reply. New replies are no longer allowed.