[I]"So when john said to start with progressive HD footage, I am unable to do that because my camcorder only shoots pseudo progressive, correct?"[/I]

NO, this is incorrect information.

The XA10 does shoot true progressive. Canon for some obscure reason split each frame into two fields. The correct way to describe this is 30PsF except Canon don't use that term.

The challenge here is that Vegas might not recognize that it is progressive, you need to check your media properties. Hm, it seems from actual tests that Canon really are not flagging the footage correctly so EVERY NLE gets it wrong :(

http://provideocoalition.com/atepper/story/empsf8217s_missing_workflow_em_part_2_the_canon_xa10_camera/

Bob.

@John M.

Yes! I admit you are right! I tested rendering the HDV clip without changing the project settings to 60p and there is no difference! That is great! I still think something has changed between VP11 and VP13 or my initial testing 2 years back was wrong.

There are two things that do make a difference: The de-interlace setting in the project properties and the smart resample feature on the event. For me BLEND is a no go, it looks bad compared to interpolate. Smart resample OFF is a no go either as it will lead to double frames; 30p in 60p or something like that.

@Christian

I just uninstalled Edius this morning but even I was able to find the settings for MPEG2-NTSC DVD, you can even select not to render audio. So when I meant DVD compliant, I meant MPEG-2 at around 9000kpbs max and 60i, a file that DVDA wouldn't have to re-render.
I actually don't need the audio, I can get AC3 out of VP, I am looking for a better workflow/program to convert HD to DVD.
Here is a 1920x1080 60i clip from 2012 shot with a HXR-NX5U and the DVD conversion out of Vegas with my filter package applied.
Can EDIUS or your workflow do any better?

As you can see from the clip and the result, the HXR-NX5U is very soft when set to 1920x1080 60i and that was the reason I switched to 1280x720 60p the following year. I still believe that converting 1280x720p to DVD gives better results 1920x1080 60i independent from the camera but might just be me.

Hi,

I will certainly do this test out of pure interest,, but gimme a couple of days... This test is the "sibling" to my conversion, but now going going from 60i to 60i..

Christian

Sure, no hurry.

@OldSmoke,

I had time to make one test today (dropped the file in same dropbox folder). This is a genuine mpg2 file rendered to a format DVDA should digest without "chewing"...

However (still as a noob Edius user), I noticed that the print-to-file function only has two resizing alternatives (blend fields / nearest field). I selected blend fields. However, the project properties setting is set at Lanzos3 (highest quality). I'm not sure if the project setting affects other than the direct render out (that does not ask for any options - just renders out in the project format).

I posted a question in the Edius forum to get a clarification how this works....

After all, comparing this Edius mpg to the one you made in Vegas, I prefer the Edius one, it IS a bit sharper and levels are the same as the original. Your mpeg is a little darker. It seems Edius handles video levels correctly, without any user intervention. I know, this thing has been also discussed a lot and SCS should do something about it...

I did the comparison By creating a new vegas project and dropping all three videos on three tracks above each others, and then just playing with the mute buttons...


I'll let you know if I'm able to render even a better one, when I get my questions answered :)

Christian

PS: We should really appreciate SCS forum's speed, we can post with immediate effect (fast dialogues possible), the GW forum seems to be heavily moderated, every post or reply appears only after a long time - fast response from others is impossible...

@Christian

I really appreciate your time and afford but I cant find the new file. The link still gets me to the initial ZIP file.

Hi,

Dropbox is not alway updating links properly... Here is a new

Christian

Where are we with this eternal endeavor?

Everyone seems to "settle" for their own particular HD to SD DVD workflow, but no one seems to rave about their method.

John

because, there are many ways to skin a "cat" or more likely since the target is not clearly defined - most any path is ok. Have produced good DVD from SCS generic tools and settings in the software program for years. No tweaks needed, no tweaks wanted.

"Have produced good DVD from............."
--------------------------------------------------------------
That's the point. I have produced good DVDs as well. Most of us still want excellent DVDs free of artifacts. Otherwise, there wouldn't be so many really really long threads here and on most other video forums about this very same topic.

John

Where are we with this eternal endeavor? Everyone seems to "settle" for their own particular HD to SD DVD workflow, but no one seems to rave about their method. I can't speak for "we," for speaking only for myself, I came to two conclusions after all these long threads:

1. Matching field order matters. This was a big revelation for me. Matching the MPEG-2 encoder's field order to the source shouldn't matter: the DVD spec allows either, and changing field order should do nothing worse than dropping the first field of the render. However, I did a lot of testing during one of these threads and found that this might very well be the biggest reason for many of the early complaints about "soft" DVDs rendered from HD sources, because the original default in all the templates was lower field first, which doesn't match most HD sources.

I should note that Sony eventually changed Vegas' defaults in the various MPEG-2 DVD Architect render templates from lower field to upper field first so I think they know about this "bug" but chose to simply "fix" it by changing the default. Of course now that they've changed it, if you are rendering from an SD source, which is usually lower field, you'll have to change this back to lower field first.

For slightly better quality, I use Nick Hope's AVISynth resizing script. For me, that provides the "ultimate" results, and I've been quite happy with that.

Finally, once I discovered the field reversal issue, and once I had Nick Hope's script, I gave up on the workaround of using sharpening, with the values set to zero. That always produced a slightly artificial look, and added a lot of time to the encoding.

For old hands here, it goes without saying that the render quality must be set to "best," not "good."

Here's a version of Nick's script (Nick, if I didn't get it right, please correct me):

AviSource("e:\fs.avi").killaudio().assumeTFF()

ConvertToYUY2(interlaced=true, matrix="PC.601")



Spline36Resize(720,height)



TDeint(mode=1) #Smart bob



Spline36Resize(width,480)



blur(0.0,1.0)     # Low-pass filtering to deal with line twitter, moire etc.

sharpen(0.0,0.75) # Adjust 2nd value between 0.5 and 1.0 to taste



separatefields()

SelectEvery(4,0,3)

Weave()

Levels(16, 1, 235, 0, 255, coring=false) #use this levels statement with external MainConcept MPEG-2 encoder

Hi John,

I looked back at the Interlaced HD to DVD AGAIN - some test renders thread we all participated in to see if I tried that one and it looks like I experimented with a bunch of similar ones and uploaded the results. I remember Nick saying he couldn't see much difference in the last four test renders I uploaded

The main difference is that I was using Procoder 3 for all the test renders while you were using Vegas and Nick was using Cinemacraft (I think). Also, we were all trying to get good results on both CRT and flat panel displays. I am no longer convinced that is necessary anymore given the number of CRT TVs I have seen in dumpsters in the past few years.

Maybe I'll try the one you posted again using Vegas for my current project.

I'll use the following script modified for multithread rendering.

SetMTMode(5,4)

AviSource("e:\fs.avi").killaudio().assumeTFF()

ConvertToYUY2(interlaced=true, matrix="PC.601")

SetMTMode(2)

Spline36Resize(720,height)

TDeint(mode=1) #Smart bob

Spline36Resize(width,480)

blur(0.0,1.0)     

sharpen(0.0,0.75) 

separatefields()

SelectEvery(4,0,3)

Weave()

Levels(16, 1, 235, 0, 255, coring=false) 

Thanks,

John

I took out the multithreaded statements, just to avoid causing problems for those who do use AVISynth, but who don't use its MT statements. This particular script works just fine with MT.

John,

Are you still using VFAPIConv to convert the frameserved avs file to avi?

I wanted to try it, but the utility only allows files with the extension .wbp I saw other screenshots that in fact used to allow .avs files, but the versions on the internet including the one you linked in an older thread only allow .wbp under "Add file"

The version is VFAPI Reader Codec v1.05

John

Are you still using VFAPIConv to convert the frameserved avs file to avi? Once in awhile, but not often. A decade ago I bought the external MainConcept MPEG-2 encoder, and it reads AVS files directly. So, for DVDs, I can just serve into that. For HD work, I never found any need for Blu-Ray, so most of my encoding is to MP4. As has been well documented, neither the MainConcept or the Sony AVC/MP4 encoders in Vegas are very good compared to either Handbrake or MeGUI. So, since MeGUI accpets AVS files, I do all my MP4 encoding by serving into that. The results, especially for uploading to YouTube, are really great (thanks to everyone here in this forum for developing that workflow).

For various other codecs, like Lagarith, HuffYUV, or DivX, I can do those from VirtualDub, which is an other program that can read AVS files directly.

So I generally only use the excellent VFAPIConv trick for serving an AVISynth script into another instance of Vegas when I am trying to avoid having to create an intermediate render during some sort of complicated workflow. That doesn't happen too often. I don't think I've used it for at least the last 4-5 months.