Thanks for the reply:

Project Prop matches video

Resample set at "SMART RESAMPLE"

Did you read this very useful thread?
HD to SD downconversion

Mike

I have been reading lots of posts on this topic and have even contacted a few professionals...and it seems they all have the same response - "been there" its all about trial and error.

I read the link you suggested (thanks for that), but I'm not sure what part of the thread pertains. There is discussion about progressive/UFF/LFF (pros about filming progressive). My video is progressive and I've insured that both Project Prop and Render Prop are set to Progressive Scan.

I know that my HD video will never looks as good when using SD-DVD, but I have seen work of others, and I know I can get better results than what I am getting.

Is there something specific on the thread that I should be referencing? I'm pretty new at this - so some of the discussion was completely foreign - my video is almost 2 hours long.

chris

2 hours is a lot to put on a single DVD. Whenever one of my projects gets over 90 minutes, I split it onto 2 disks. Supermediastore.com has cases that hold 2 disks.

A 2 hour program crammed onto a single 4.7 GB DVD will have a very low bitrate.

Finally, how are you rendering this out? You need to do 2 renders, 1 to the MainConcept MPEG2 "DVD Architect NTSC Widescreen video stream" template (without audio) and a second render to the Dolby Digital AC-3 Pro "Stereo DVD" template using the same file name for both (except for the extensions). If both rendered files are in the same subdirectory with the exact same names, DVDA will find the audio when you drag in the video. Following this workflow prevents DVDA from rerendering your files.

Use a bitrate calculator (I like http://dvd-hq.info/bitrate_calculator.php) to determine what bitrate you should set as you render the video. Try and keep it above 8Mbit/second.

I have found over the years that rendering with a third party program, TMPGEnc Mastering Works, via the Frameserver gives the best DVD results from HD source.

edit: Here is another thread you should look at. In particular, I recommend reading JohnMeyer's post about half way down the thread.

http://www.sonycreativesoftware.com/forums/ShowMessage.asp?ForumID=4&MessageID=918436

Did you try setting the switch to "No-Resample"?

The following is copied from a reply I gave to someone else a few months ago.Before you spend too much time chasing a better workflow, you should see how well you can do making DVDs from HD using just Vegas, and nothing else. Do these three things when rendering from HD interlaced to MPEG-2 DVD interlaced, and see if you like the results.

1. Make sure you set a Deinterlace Method in the project properties. Set it to Blend if you have nothing but talking heads; set it to Interpolate if you have lots of fast moving motion. Don't spend a lot of time worrying about this, but just make sure that it is NOT set to None.

2. Set the project and rendering quality to Best. This uses a different resizing algorithm, one that delivers better quality when downsizing HD to SD. This is not a huge deal that is going to make a night/day difference, but it is worth doing.

3. Render to MPEG-2 using upper field first when your source material is HD (which is almost always upper field first). In theory this should not matter at all, but someone in this forum discovered that Vegas degraded quality when the field order was switched during rendering. I did some tests and confirmed that this was true. I don't know if this ever got fixed in later versions of Vegas, but I suspect it did not because very little has been done to the MPEG-2 (DVD) rendering in the last few versions.

This last one makes a surprisingly big difference, and may explain why some people many years ago were getting such disappointing results when downconverting and then rendering their HD video to SD MPEG-2 for DVD.

So use lower field first when rendering MPEG-2 from SD material, and use upper field first when rendering from HD source material.

Some people add sharpening, with the settings set to zero (which still actually does some sharpening). At one time I thought this was a good idea, but I no longer recommend it.

Finally, you should always start with the DVD Architect Widescreen template (NTSC or PAL, depending on where you live in the world). Then, make sure that the various quality settings are set to maximum quality. Do no alter the min or max bitrate, but set the average bitrate to the highest number that will let your material fit (use a bitrate calculator). If the total length of your material is under 75 minutes, you can use 8,000,000 bps VBR (don't go higher than this). I am not a fan of CBR. If the average bitrate needed is below 6,500,000 bps, then you definitely need to use 2-pas VBR.

If you are not satisfied with the results you get from this, then you can take the plunge into the advanced ways of resizing video prior to encoding to MPEG-2. The usual way to do this is to install the Debugmode frameserver (free) and then use that to frameserve into an AVISynth script. You open this in VirtualDub and render your video from there. Nick Hope and I (and John Cline and others) posted several alternative scripts that do a better job with the resizing than does Vegas, preserving more detail. If you decide you want to do this, and if you don't want to read 200+ posts, just let me know and I'll post the simplest script that will give you better results than Vegas.

Finally, let me try to clear up a few other points of confusion you have expressed.

First, you can definitely put 30p video into 60i (usually called 29.97 interlaced) video container. 30p is just 60i video with no time difference between the even and odd fields.\One last thing. The reason for looking for solutions outside of Vegas has little to do with what resizing algorithm Vegas uses. I know there have been lots of discussions and arguments about this, both in this forum and also over at the doom9.org forums. However, most people are not going to notice much difference between all the difference resizing algorithms.

Instead, the real problem everyone was trying to wrangle in those tortuously long posts linked to at the beginning of this thread, is the issue of how to temporarily align. in time, the odd and the even fields in an interlaced video. The worst thing to do (which Vegas does not do unless you make the mistake of setting Deinterlace Method to None) is to resize the whole frame, using both the odd and even fields. The reason for this -- and this leads to an understanding of the underlying problem everyone is trying to solve -- is that those fields are from different points in time. If the motion is fast enough, then when you do the resizing, you end up combining pixels from what amounts to two almost unrelated images, and the result is an absolute, unwatchable mess.

The simple solution is to resize the odd fields (they all come from the same point in time) and then separately resize the even fields, and then combine the two resized fields back together. The problem with this approach is that you are throwing away half the spatial information, and you end up losing a lot of fine detail, like the detail you might find in the fabric on a shirt.

The "solution" is to create a really, really smart algorithm that can take the even fields, which happened 1/30 of a second apart, and "estimate" where each pixel in those fields would be at the point halfway between those two moments in time. If done perfectly, each resulting even field would now be at the same moment in time as the matching odd field, and the two could be taken together and resized to a lower resolution, and that result separated back into individual fields. You could also keep the result as a 60p video, but the DVD specification doesn't permit 60p video.

Vegas most definitely does not contain any of this technology, despite my private emails to the development team many, many years ago urging them to invest in "motion estimation" technology which is at the heart of the solution not only to this problem, but also many encoding issue, elegant slow motion, frame rate changes (such as PAL <--> NTSC), and more.

The other thing to note is that the resizing algorithm is not that big a deal and is not the reason why Vegas can't produce the "ultimate" quality from HD interlaced material. There are dozens - perhaps hundreds -- of posts in this forum and also over at doom9.org where people argue about the "best" resizing algorithm. There are differences, to be sure, but they are minor compared to the issue of how to handle resizing even and odd fields that are both spatially and temporaly different (i.e., at different points in time).

Given your source video properties: 1920x1080x32 29.970p

I used a similar source as follows:
Complete name : C:\Users\John\Desktop\Delete Me\Source XDCAM EX 1080-30P.MP4
Format : MPEG-4
Commercial name : XDCAM EX 35
Format profile : Base Media / Version 2
Codec ID : mp42
File size : 943 MiB
Duration : 4mn 45s
Overall bit rate mode : Variable
Overall bit rate : 27.7 Mbps
Encoded date : UTC 2015-04-16 20:42:53
Tagged date : UTC 2015-04-16 20:42:53

Video
ID : 1
Format : MPEG Video
Commercial name : XDCAM EX 35
Format version : Version 2
Format profile : Main@High
Format settings, BVOP : Yes
Format settings, Matrix : Custom
Format settings, GOP : Variable
Codec ID : 61
Duration : 4mn 45s
Bit rate mode : Variable
Bit rate : 26.2 Mbps
Maximum bit rate : 35.0 Mbps
Width : 1 920 pixels
Display aspect ratio : 16:9
Frame rate mode : Constant
Frame rate : 29.970 fps
Standard : Component
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Compression mode : Lossy
Bits/(Pixel*Frame) : 0.421
Time code of first frame : 00:00:00:00
Time code source : Group of pictures header
Stream size : 891 MiB (94%)
Language : English
Encoded date : UTC 2015-04-16 20:42:53
Tagged date : UTC 2015-04-16 20:42:53
Color primaries : BT.709
Transfer characteristics : BT.709
Matrix coefficients : BT.709

Audio
ID : 2
Format : PCM
Format settings, Endianness : Big
Format settings, Sign : Signed
Codec ID : twos
Duration : 4mn 45s
Bit rate mode : Constant
Bit rate : 1 536 Kbps
Channel(s) : 2 channels
Sampling rate : 48.0 KHz
Bit depth : 16 bits
Stream size : 52.3 MiB (6%)
Language : English
Encoded date : UTC 2015-04-16 20:42:53
Tagged date : UTC 2015-04-16 20:42:53

I used the following Project Properties:


... the following Render Settings:


and got very good results making a DVD in DVD Architect.

Well thanks for all the info. Here is where I'm at now: I've tried it all and the results are the basically the same. Note: I have tried to disable resample and that didn't seem to make a difference.

I did try this:

Rendered in Vegas to HDV (1080) and then took the file into TMPGenc and rendered to DVD (720x480). As it was encoding, the quality actually looked pretty good. However, once I took that file into DVDA and burned the DVD - BYE BYE Quality - HELLO jagged lines.

UGH - Very frustrating!! Any other suggestions? I'm sure I am just overlooking something and it doesn't help that I don't fully understand the meaning of the terms. Is it possible that my footage or file is bad? I did bring in a mp4 (format AVC) file which and a jpeg file - which I matched to the project properties. Could those two files be the root of the problem?

Thanks again!!!

[I]Rendered in Vegas to HDV (1080) and then took the file into TMPGenc and rendered to DVD (720x480).[/I]

That doesn't make any sense at all. You are converting 1920x1080(your source from the X70) to HDV 1440x1080 and then to DVD 720x480.

Keep in mind that reducing from HD down to DVD you will loose a 1/4 of the information, more or less. You are also converting down from 4:2:2 to 4:2:0 and from progressive to interlaced.
You can try and render straight from Vegas to a 29.97p (progressive). Yes it isn't an official DVD format but it will work on 99% of DVD players.

If you still want to do 60i DVD then try and apply this filter package in your video output bus. You will need Sony's free preset manager to install it.

As mentioned earlier, it is important that you render the audio separately from your video as AC3. As long as the file name is the same as the video file, just different extension and both are in the same location, DVDA will combine them.

Hi John - Thanks. I tried taking the CREDIT ROLL portion of my project (about 1m 34s) and did exactly what you did - both project properties and render properties.

The results - no change. On my TV - the letters move as they scroll up the screen - almost making me dizzy. It almost gives the appearance that there are transparent lines on top of my footage (in this case a still shot) and when the letters move through the line - they distort...if that makes any sense.

Now I am not sure - maybe I am overcritical and that's the best it can look????

Any chance it has something to do with my computer? As I watch the video render in Vegas - I can already tell that the end result is going to be bad.

Have a good one..and thanks in advance for your expertise.

Perhaps you could post a few seconds of you rendered 720x480 video elementary stream so we could see what you're seeing? Use dropbox.com or other file hosting site and include the link here.

DVD Architect tells you if it is re-rendering the video. First of all, if it is simply preparing the project, and not re-rendering, the whole process only takes a few minutes. If it is taking thirty minutes or more, then it is re-rendering.

An even more certain way to tell if the video is being re-rendered is to open the Properties dialog and see if there is a green check mark next to the name of the video or, instead, a yellow exclamation mark. If you see a green check mark, then DVD Architect is doing absolutely nothing (zero, nada, zilch) to the video, and anything you are seeing is the result of something further down the chain (e.g., how your DVD player is attached to your TV set).

https://www.dropbox.com/s/k1rmlvns8oqoz0k/TestClip.mpg?dl=0


Hope this works!!

I don't think DVD Architect is re-rendering. I'm getting no error messages or yellow exclamation marks. It does take about 30 minutes or more to "prepare" but the actual burn is pretty quick. The only "warning" I'm getting is that the Audio is being compressed.

Note that the quality when I preview on DVDA (prior to burn) doesn't look good either.

Blu-Ray connection to television via receiver, HDMI connection

[I]The only "warning" I'm getting is that the Audio is being compressed.[/I]

when you render the audio separate as AC3 you will not get that message either.

Judging from your sample I would say that's as good as it gets with this kind of fonts and scroll speed.
You can try and apply the filter package from my link; just disable the sharpening FX in it. The filter package will "soften" the fonts a bit and reduce the line chitter you get.
Another thing that sometimes helps is setting the Text media resolution to DVD. When you work in a HD project, the text media will be HD too and now needs to be resized. You can set it to 720x480 in the media properties and try rendering again; just try a short region first and see if that help.

Have you tried rendering to a 29.97p file?

I am rendering audio separate as AC3 (same name and file location as video) - but I did get the message???. FYI, when I add the video file, I do not add the separate audio file because as I understand it, DVDA does that automatically. Correct?

I tried changing the project properties for the text media - didn't help. But honestly, I'm ok with what I've got - its the actual video that is "unacceptable". I provided a link to the Credit Roll portion thinking the problem would be real obvious...not realizing that text media has other issues all on its own.

Included another link - of video footage for opinions

I haven't tried rendering to 29.97p file - simply because I didn't understand what you meant...can you explain further?

https://www.dropbox.com/s/0ysm8dmy98wtg4u/TestClip1%28disresample%29.mpg?dl=0

Thanks!!

[I]I haven't tried rendering to 29.97p file - simply because I didn't understand what you meant...can you explain further?[/I]

Your source file is 29.97p (progressive) and the official DVD standard is 60i (interlaced) which is also the setting in the DVDA render template.
In the DVDA render template, set the Field Order to progressive and render it out and use it in DVDA.

However, your second sample looks just as good as DVD will look when it is interlaced. If you know that you have to deliver DVD, next time try and shoot 60i instead of 30p.

"[I]However, your second sample looks just as good as DVD will look when it is interlaced. If you know that you have to deliver DVD, next time try and shoot 60i instead of 30p.[/I]"

I've thought all along that 30p was an unfortunate choice if DVD was the target. Looking at the second clip, it is a wider shot than the subjects with a number of small items all around. You are likely bumping up against the DVD resolution limit.

Since you have a 1080p source, you could use key-framed Pan/Crop to move in on the subjects while still maintaining greater than the 720x480 DVD resolution.