Video is also tied to the quality of the encoder. For the last couple of years of my professional career I encoded videos for theaters and National Park visitor centers. These screens were 30 feet or larger. The MPEG encoders provided with Vegas would not hold up at that size of projection. Artifacts that you would not notice on a 52" screen became distracting at 30'. I used am MPEG hardware encoder that at that time sold for $30,000. The images were almost flawless. So in addition to bitrate and codec used, you also have to factor in the algorithm of the encoder.
Video is also tied to the quality of the encoder. For the last couple of years of my professional career I encoded videos for theaters and National Park visitor centers. These screens were 30 feet or larger. The MPEG encoders provided with Vegas would not hold up at that size of projection. Artifacts that you would not notice on a 52" screen became distracting at 30'. I used am MPEG hardware encoder that at that time sold for $30,000. The images were almost flawless. So in addition to bitrate and codec used, you also have to factor in the algorithm of the encoder.
Good to know. I'm definitely no expert on encoders when it comes to video....actually far from it and thus the reason why I have an interest in this topic. I would have thought MPEG is MPEG when it comes to encoders. My comparison would be like comparing a Maincepts MP3 encoder to a LAME MP3 encoder where as long as the bitrates are at the same setting in each, it would be difficult to tell the difference between them. What I currently thought is that as long as you selected an encoder which supported x264 encoding and chose a high enough bitrate, then things tended to render out pretty well. I have no idea what the difference would be between using a Sony AVC, or Magix encoder or something else but I'm willing to learn.
I also currently have no idea if I edited a video together in Vegas, how I would go about using another/better encoder not already included with Vegas to create the final composite video file. This is a frustration point for me, since these are things I never had to be concerned about when doing audio work inside of Vegas. It's also another frustration point that it sounds like most every present encoder which is built into Vegas needs to be edited just to get decent video quality renders out of them.
In the case of major motion pictures released on DVD or bluray, those systems used often offer the possibility of tweaking individual scene bitrates. They don't always use an automatic "one size fits all" encoder.
In the case of major motion pictures released on DVD or bluray, those systems used often offer the possibility of tweaking individual scene bitrates. They don't always use an automatic "one size fits all" encoder.
Interesting. Any idea how they go about selecting the different bitrates per scene? I'm guessing if a scene has fast action movement, then the bitrate gets bumped up, but if it's mostly stationary slow movement, then lower bitrates? Light vs dark scene bitrate adjustments?
Would that be a good use case for selecting a variable bitrate where the encoder attempts to do that for you?
but there eventually comes a point where going with a higher bitrate doesn't seem to gain you anymore additional quality and only results in larger file size.
Ah, now I get the question and yes, that is something I've been interested in learning about. Here's a chart I made comparing x264 bitrates versus quality over increasing RF settings. A measurement called SSIM is used to compare accuracy to the original, with 1.000 being a perfect copy. It's been found that "optimal" quality for modern encoders ( including x265) is at SSIM 0.995, the maximum reasonable quality without throwing LOTS of bits at it.
As you can determine from the chart, "optimal" bitrate for high-end motion x264 (1080 p30) occurs at 20-28 Mbps, or RF 18-20. Above that, the eyes can perceive no differences (as their resolving power has long since been surpassed), so quality is essentially as good at 25 Mbps as it is at 144 Mbps, despite wasting nearly 6x times more space!
I ran another test, this time uploading various bitrates from the same source to Youtube, 4k p50. The results are predictably meh, showing we can't force quality with excessive bitrate, despite colorful theories.
That is what the automatic VBR attempts to do. Adjust itself per scene. But sometimes it misses. For instance, if you have a fade to black scene and it sits on the black for a while, it will usually use a low bitrate. Upon viewing though, you will notice artifacting and blocks during the black. It is not always a smooth black screen.
Most consumer encoders do not offer manual scene settings. Just an overall setting. I tend to use Fixed Bitrate more when the size of the file isn't an issue because it avoids low bitrate blocky scenes.
but there eventually comes a point where going with a higher bitrate doesn't seem to gain you anymore additional quality and only results in larger file size.
Ah, now I get the question and yes, that is something I've been interested in learning about. Here's a chart I made comparing x264 bitrates versus quality over increasing RF settings. A measurement called SSIM is used to compare accuracy to the original, with 1.000 being a perfect copy. It's been found that "optimal" quality for modern encoders ( including x265) is at SSIM 0.995, the maximum reasonable quality without throwing LOTS of bits at it.
As you can determine from the chart, "optimal" bitrate for x264 occurs at 20-28 Mbps, or RF 18-20. Above that, the eyes can perceive no differences (as their resolving power has long since been surpassed), so quality is essentially as good at 25 Mbps as it is at 144 Mbps, despite wasting nearly 6x times more space!
I ran another test, this time uploading various bitrates from the same source to Youtube, 4k p50. The results are predictably meh, showing we can't force quality with excessive bitrate, despite colorful theories.
Thank you! That is AWESOME! This is exactly the type of stuff I was looking. It moves the grey fuzzy areas into the more scientific and quantifiable areas that I can use as general rule of thumb guidance.
That is what the automatic VBR attempts to do. Adjust itself per scene. But sometimes it misses. For instance, if you have a fade to black scene and it sits on the black for a while, it will usually use a low bitrate. Upon viewing though, you will notice artifacting and blocks during the black. It is not always a smooth black screen.
Most consumer encoders do not offer manual scene settings. Just an overall setting. I tend to use Fixed Bitrate more when the size of the file isn't an issue because it avoids low bitrate blocky scenes.
I think it may have been mentioned earlier and I have yet to explore VBR encoding and I've avoided using it on the audio side of things as well. Does VBR rendering have a minimum bitrate setting adjustment so you could potentially reduce the likelihood of the black scene artifacts from occurring?
That is what the automatic VBR attempts to do.
@Former user
Now that this thread is totally off-topic, this old Handbrake tutorial has been helpful to many in understanding the differences between VBR and CQ/CRF encoding, and why the concepts and algorithms to do each are totally unrelated, because they serve opposite purposes.
https://handbrake.fr/docs/en/latest/technical/video-cq-vs-abr.html
Basically, CQ/ CRF is not VBR because no target bitrate can be pinned down, as illustrated in th OP's earlier test, only a loose "range" defined by the quantizer. Once again, the target is Quality, not Bitrate.
Mainconcept uses VBR, or linear math. CQ/ CRF Encoders, of which Sony AVC is an early adopter, cannot be described in those terms, but only in nonlinear regression algorithms, which I can't explain unassisted. Hope this furthers your understanding.
Thank you! That is AWESOME! This is exactly the type of stuff I was looking. It moves the grey fuzzy areas into the more scientific and quantifiable areas that I can use as general rule of thumb guidance.
That's why I've related to your questions, because I've had the same ones, and I hate hearing answers that should begin with, "What if ..."
Now, for an answer to another great question. Mainconcept MPEG-2 has always had Minimum Bit Rate control, which is literally a godsend, as raising MBR to 2,000 Kbps allows us to enjoy the efficiency and space-saving features of VBR without the blocky shadows. Unfortunately we've had to train hundreds of users not to leave it set at 900 Kbps, the default. Although legal for BluRay, MPEG-2 runs nominally 5 times larger than AVC.
Mainconcept AVC/MP4 does not have Minimum bitrate control. One can force a higher minimum bitrate by setting the Average and Maximum bitrates closer together, losing some efficiency, of course.
In x264, regrettably, there is no bitrate floor. One can set --vbv-maxrate lower to level the playing field, which is a really good idea for streaming, especially! Unfortunately, no such rate control settings are exposed in Sony AVC.
I've got more, just tell me when to stop.
When using VBR activating 2 pass encoding can help because, although a lot slower, the encoder can evaluate the video contents before the final encode to adjust the bit rate better to accommodate the changing video data. I suspect the pro encoders may even be able do multiple passes, which is fine if you don't need real time encoding but want the best result with the efficient use of storage space.
I am rendering for Blu-ray from Vegas (17.0 build 387) using MainConcept MPEG-2 with the 'Save as separate elementary streams' ticked which states "This will create separate files for the video and audio streams with extensions .m2v and .mpa respectively instead of the .mpg file" (This 'greys' the 'Include audio stream' button under the 'Audio' tab in the export settings) However, no '.mpa' is being created when I do this. I saw in another thread that it is advisable to render audio as .ac3 which I did using the Vegas 'Dolby Digital AC-3 Studio' template.
I want to create the Blu-ray using DVDA (build 100). I can import the m2v to DVDA but there does not seem to be any way to add my ac3 to the 'audio-less' m2v file once I import it so I'm stuck.
"I can import the m2v to DVDA but there does not seem to be any way to add my ac3 to the 'audio-less' m2v file once I import it so I'm stuck."
Cheers that solved it. I double clicked on the movie in DVDA and navigated to 'track media' under Media Properties and added the ac3 track.