Well maybe it wasn't that, but there was a marked improvement in quality when frame serving to handbrake, as opposed to encoding in Vegas. If I'll still need to do that in V15, then I can't see any reason to upgrade?

The only possible deal-changer with progressive source that I've heard of since our 2011 tests in Vegas Pro 8-12 is the addition of the adaptive deinterlace filter, which definitely bears further testing. If you have time to run some downscale-only (progressive) tests, which I haven't done in a while, I'd love to see your impressions! If you have time to compare the newer Vegas and HB deinterlacers again in 2018, that would be even better!

It's easy enough for me to remount my earlier tests in Vegas 14, but without hardware encoding.

I don't know about "necessary" or about that particular resize workflow, but I have yet to see any encoder that is BETTER than Handbrake / x264. It is quite amazing, considering it is free!

If you haven't had a lip sync complaint by now with Vegas2Handbrake don't let me discourage you from using it. You should read this and know that for sources containing PCM audio, the audio leads the video. I was  never fond of the 44 kHz limitation of debugmode frameserver in its last iteration. I have used lossless intermediates to pass to Handbrake. I still use the Vegas2Handbrake scripts when I have to bang something out quickly where audio is absent or not critical.

I never thought Handbrake was absolutely necessary 1) if all source media was progressive 2) if all source media matched the pixel dimensions of the delivery 3) if one could deliver a high enough bit rate to make other encoders competitive.

I have lots of family video files that I watch that were encoded in Handbrake, but I have as many or more that were encoded with Mainconcept or Sony AVC for Blu-ray. (I can't remember the last time I mounted an actual Blu-ray disc. It was probably from Red Box some months ago.)  

When you say that the audio leads the video, are you talking about the issue where you have to extend the loop region by 1 additional second for audio and video to sync? That's no problem for me. With the help of someone from this forum a while back, I have a script that gives me a button that does just that, adds one second to the loop region, then I just click the button to send it to handbrake.

My needs at the moment are encoding from DV AVI, or DVD to 480P for upload to Vimeo. As well as taking footage from the Canon XA10 and which is 1080 (progressive according to a lonnnng discussion I had with John Meyer) and making it 720P.

So for the cases cited above, it seems the handbrake route is still the better way to go

Here's a comparison of Lanczos, Bicubic, and Bilinear 1080->720 downscaling. No sharpening has been allowed, and chroma subsampling has been avoided, as have interlace, etc.

Only the bilinear (third down) shows a little softness in the edges in the second quadrant. Even a perfect scaler would show no definition in the first quadrant. Sorry the forum software messes with 720p images (which seems miserly).

Now to confess: I recommend Vegas for 720p because it's quick, and I use Handbrake because at 68, I still like to twiddle. Might want to compare rendering times using different encoders.

If one is upscaling for any reason, the more thorough method is the Handbrake CLI version, which allows Lanczos upscaling.

I'd say thank you but I wouldn't know what I was thanking you with those images. I'm not so much concerned with the rendering time itself. What concerns me with Vegas2Handbrake is that I have to do each video manually. So if I have a bunch of events on the timeline to render separately, I have to get the loop region, add the 1 second, send it to handbrake, render it, and repeat the process over and over again. I can't figure out any way to do batch renders this way. But now the fact that I'm aware of this audio drift issue, it has me second guessing using Handbrake, even though I don't see any lip sync issues.......

I use handbrake all the time, but not via the Vegas2Handbrake workflow. I just render out a lossless avi and pass that to handbrake. Obviously this means two renders, but it is not twice the time, as the MAgicYUV avi render is quite fast. You could do batch renders this way using something like Vegasaur. (Handbrake will render by folder IIRC).

Not sure why you resizing fragments instead of rendering the whole project in 1080 (original resolution) and then downscale the resulting video to 720 for sharing/uploading. I'm doing this all the time.

Well for example, I have a long video that was filmed in 1080, say 2 hours, and it's edited and broken down into a bunch of chapters/sections on the timeline. But I want each section to be a separate file, so they need to be encoded separately.

My comparison of downscalers above suggests the algorithm (Lanczos, bicubic) in itself is not the cause of quality differences between Vegas and Handbrake.

There are dozens of reasons to compare and contrast the two on quality considerations, but that is not one of them. So times "may" become a factor, if deinterlacing is not.

Well for example, I have a long video that was filmed in 1080, say 2 hours, and it's edited and broken down into a bunch of chapters/sections on the timeline. But I want each section to be a separate file, so they need to be encoded separately.

This does not explain me why you'd like to have all these downscaled fragments on one timeline.

I'd create one project per section and rendering them to 1080/720 (what is required).

If, at the end, I need to combine these sections (with some transitions in-between) -- I'd create another project and included already existing projects (nesting) or rendered sections (if additional quality loss is acceptable).

Probably I'm missing something important.

Well they're not fragments at all. The camera is filming pretty much non stop for the entire shoot, then the footage is all captured to one timeline and edited, markers inserted for each segment, etc. Organization would be a nightmare if I had separate projects for each section. Maybe I'm missing something important, but I don't know any other way to do it.

Probably I'm missing something important.

Probably.

Every year we record our figure skating live performance of the Nutcracker on Ice. Each act is between 30-50min with several scenes. Each act is a multicam, 3-4 cam edit and some parents buy the whole show or only a scene of their child performing. Like @MikeLV, I would like to use Vegas2Handbrake to render each section but that doesn't work. Instead, I render each section with Vegasaur because I can add a common lead in and out with MC AVC CUDA with my GTX580. However, I have to keep the bitrate high to get a similar quality as Handbrake which is unbeatable at lower bitrates.

Last year, I decided to render a master to XAVC-I, out it in a new project, apply NeatVideo and from that I rendered sections to XAVC-S because Handbrake can't read the XAVC-I in MFX wrapper properly. Fortunately, Handbrake can batch render the XAVC-S sections and that worked out well. The complete XAVC-I master gets rendered to 720 60p for BluRay and 480i for DVD

To sum it up, there is a use/need for rendering sections of a project. Others may have other reasons but I am sure they are all good ones.

I still would love the idea of having an integrated VP module "Magix Batch Converter" as an Pay-For-Add-On to the basic NLE instead of 3rd party FX Bundles that expire after each VP version..

My comparison of downscalers above suggests the algorithm (Lanczos, bicubic) in itself is not the cause of quality differences between Vegas and Handbrake.

There are dozens of reasons to compare and contrast the two on quality considerations, but that is not one of them. So times "may" become a factor, if deinterlacing is not.

You were the original creator of the handbrake method way back when, and there's no doubt that it delivers much better results than rendering from Vegas. So in your opinion, if it's not the scaling algorithm, then what is the reason that HB delivers much better quality MP4 encodes?

Just to interrupt for a moment.... :)

It gives better results than any other software I have tried (not just Vegas) and I can confirm this is the case for high res (UHD), unscaled projects too (EDIT: also when scaling very high res photographs to UHD with motion.)

Rank speculation about reason: talent of the x264 library open source contributors, some of whom are probably very experienced & knowledgeable.

If x264 is open source and freely available, what prevents Sony, or now Magix, from including it as an encoding option?

I can list a dozen things off the top of my head, the "big three" being deinterlacing with scaling when needed, and more efficient use of bits (same quality, smaller file).

Suffice it to say that "way back when" (summer 2011), I concluded that the Handbrake method isn't a great tradeoff for fime vs. quality with 720p that doesn't involve deinterlacing with scaling. Watch carefully the "Good Method" in the first tutorial; although I didn't specifically qualify downscaling progressive source for that method, my impression seven years later isn't very far off that mark, and that was validated later with scalers in Vegas and Photoshop, as well as my new test earlier in this thread.

That said, except for decombing the 1080i examples, the rest of my 720p tutorial looked just as good rendered at 10 Mbps in Vegas.