Thank you for these suggestions which I have followed with some improvement. However the result is not what I had hoped.

As you say, a lot of the video data is lost at the lower resolution but when I view a commercial DVD the quality of detail is considerably higher.

I am puzzled. I don't understand why it is not possible to produce a DVD which is comparable to a commercial DVD in terms of detail and image quality when I have originals which are crisp and clear.

Are there resources available to commercial makers which we do not have?

Jim Brook

Suffice it to say this question is asked once a week. And the answer is always the same as it has been in the past. YES, commercial DVD preparation goes through a host of steps and tools you do not have. Think about the fact that DVD sources out of Hollywood also have a whole lot better sourcing from cameras and film than you could ever have.

NOW - Blu-ray prep and burn is another matter. This is the highest quality workflow accessible by Joe sixpack on the consumer realm.

You can take an HDV source (ie., not full HD) shot with a good 3-chip prosumer camera in the $4,500 range and workflow it to Blu-ray such that with color-grading and careful management of edits you can roll your own consumer hi-grade Blu-ray disc.

If you do a search here you'll find many many similar posts. There are complicated work flows involving other 3rd party programmes (most free) that produce extremely good results. It depends how much time you want to spend on it. One of the things Vegas isn't good at is deinterlacing and resizing at the same time. That sounds like what your "flickering" issue is? For this I use TMPGEnc authoring works 5. It does a MUCH better job. The down side to that is no smart render.

That's not much of a down side, considering that if you're resizing, deinterlacing, or both, smart rendering isn't possible in any software anyway.

Ehhhhhh!!?? Render an m2t (HDV) to mpeg2 for DVD using Vegas. This is resizing. Render that file again using the exact same settings, it smart renders. Add in deinterlacing...it still smart renders.

"but when I view a commercial DVD the quality of detail is considerably higher.

One camera, lens, lights, and MPEG encoder to do a "commercial DVD" costs more than my home, SUV, and savings combined.
Just how much money did you plan to invest in a standard definition product??

Want a better prosumer encoder than Mainconcept?
Cinemacraft (starting at $1,000) is your friend.

You have a HD camcorder. Shoot HD, burn BluRay.

Because my Sony cameras create a 'Super Bright' above 100 IRE (Above 235) I use Sony Levels to reduce the upper end so as not to lose details on the brighter areas of the video.

Input start: 0.000
Input end: 1.000
Output end: 0.920
Gamma: 1.000

I could use the Sony Broadcast filter to reduce the over 100 IRE but it just basically kills whatever is above 100 IRE and a lot of the brighter areas are lost and have no details.

I created a filter package that has both the Sony sharpness (light setting) and levels filters that make a lot of difference in the quality of my videos.

Important Note: This is what I do with the videos from my Sony cameras. The videos from your cameras may be different so what you need may also be different.

I set the Waveform monitor to where only the 'Studio RGB (16 to 235)' is checked so the blacks will be at zero which is much easier to see than 7.5 because there is no 7.5 marker on it and the brightest whites as they should be are at 100. Above 100 or below 0 and it is wrong. Solid blacks above zero is wrong. Use the NTSC test pattern from the 'Media Generators' and see that the brightest whites are at 100 and the blacks except for the PLUGE (Picture Line-up Generation Equipment)are at zero. The PLUGE has one small part that is below zero as it should be. Use the PLUGE from the 'Media Generators' to more easily see what I mean.

Use the Waveform monitor without any filters to determine what your original video consists of and then use filter(s) as needed to (within reason) to make it what it should be. I say within reason so you do not go overboard trying to make it correct but the video should still not exceed 100 IRE or go below 0 IRE. Use Google search for PLUGE and you will learn a lot. I did a search of pluge and got stuff but when I use PLUGE I got what I wanted.

If you do not start with correct video you will not finish with correct video and the quality will never be good no matter how you render it!

As others have said render the audio separately.

On videos up to 1.5 hours I use 'Best' one pass. On longer videos I use 'Best' two pass because of the need for a lower bit rate than default.

I hope this helps you.

Danny Fye
www.dannyfye.com/ccm

> "I am puzzled. I don't understand why it is not possible to produce a DVD which is comparable to a commercial DVD in terms of detail and image quality when I have originals which are crisp and clear."

That's because as others have posted they use an original source that's much crisper & clearer. In my opinion anyway, given the same source VP would do better than most of the retail DVD movies on store shelves, because nowadays that re-size & re-encode is probably more a quick task or job since they want Blu-Ray to look as much better than the DVD version as possible... Blu-Ray discs usually sell for more $, & with much better DRM than DVDs, they'd rather that was what people bought anyway.

Arthur wrote about optional, added steps/filtering... As videoITguy posted, as good as your camera footage is, it's not as good as the source for retail video Blu-Ray & DVD, & some of those steps/filtering might make up *some* of that difference. From the sound of it your biggest loss of quality comes from your source being interlaced, which doesn't re-size nearly as well as the progressive frames Hollywood uses. You're missing every other line of data, & there's no perfect way to make that up, to re-create it. The stuff Arthur mentioned can help, sometimes a lot, but it's up to you if it's worth it or not.

What you don't mention, Jim, is where you're viewing your DVD, & that can sometimes be important, e.g. if you're looking at your DVD on your PC/laptop. Assuming you're creating a DVD to be watched on std TVs, judge the results on a std. TV using a normal DVD player. Can you do without menus? If so, then perhaps you might try some tests with DivX/Xvid, which will play in a lot of DVD players & won't have the same restrictions as the DVD spec. If it's to be viewed on PCs/laptops &/or HDTVs, why not use Blu-Ray with your 1440 or 720, on either BD or DVD discs?

As I mentioned in post above, choosing a Blu-ray workflow has a lot of advantages to the consumer today. When I am doing quality control of a one-off type of produced disc in the studio (my source is high-end 3 chip) I screen it on a 42+inch widescreen LCD and measure it against a $10.00 commercial Blu-ray release playing in the same set-top player.

A lesser form of that test for DVD one-off's is to screen your DVD in a BLURAY (yes, Blu-ray player with good uprez) set-top player connected to 42+inch widescreen LCD and compare the result to Hi-definition over the airwaves direct receive broadcast. _Edit to clarify the comparison source, I am referring to a broadcast of a 1980's era movie or a 1990's outdoor filmed adventure series. DO NOT try to compare your work to a live telecast in hi-def of a football game in 2012-13. _ Then, your test should come within 5% range if you have great source for your own footage and use a workflow that is optimum.

Sometimes we HAVE TO make a dvd from a HD source..
Vegas is terrible here, i use debugmode frameserver for creating virtual avi file. I open this in virtualdub, and to the scaling here, with very good result (filters: field bob, resize, little sharpen).
Again use frameserver from virtualdub with avisynth, and that file i can import to Procoder, so i can create a high quality mpeg2 file.
Works great!

Almost all of my output is SD on DVD, although I have moved up to widescreen. My customers are mostly retirees, and BluRay is not in their vocabulary. I often have to visit their homes to set up the DVD player for widescreen.

My source is 1920 x 1080 60i from a 3 chip Panasonic, or 1440 x 1080 60i from an older Canon. I typically mix in some jpg stills, and do a lot of chromakey, which I light myself in a room that is too small (only 30 long x 12 wide - I sure wish it were bigger in both dimensions, but this is what I have been given.) You can see the environment in this Credits/Outtakes reel from one of this year's historical pieces - costumed amateur actors talking about the lives of some of the community's notable citizens. This YouTube was an MC MP4 upload VBR max. bitrate 14M, min. 10M.



To the point, I get a much much better SD render using TMPGEnc, frameserved from Vegas Pro 12, than from the Main Concept encoder. No combing, the herringbone on the sexton's brown vest doesn't shimmer, and the plaid doesn't shimmer.

I have been experimenting with the MC encoder. I have found no difference in output quality in 1-pass and 2-pass renders. Even using Vegas to compare the two output files, there just is not a difference, and the 2-pass encoding takes about twice as long. The 2-pass TMPGEnc encode takes 3 times as long as the 1-pass MC encode. But it is really, really worth it.

I have also been comparing render times between my soon-to-retire i7-950 and the shiny new i7-3770k. Renders with debug frame serve to TMPGEnc are about twice as fast, even though I have not tried to overclock it yet.

All of the above discussion is with EVGA Geforce GTX 660 ti turned on.

"One camera, lens, lights, and MPEG encoder to do a "commercial DVD" costs more than my home, SUV, and savings combined.

Much truth there.
Over the last few months I've used the following cameras, none of which are really top shelf.

PMW-350 $20K+
PMW-EX1 $8K
HVR-Z5P $5K
C5 $1.5K
Chinavision Actioncam $120

Even the $120 camera's original footage on a HDTV, shot under ideal conditions doesn't look totally tragic. Converted to a SD DVD and displayed on the same HDTV and the difference between the PMW-EX1 and PMW-350 is quite obvious.

The EX1 always has some noise, the 350 even at 6dB gain, none.

Another factor. 720p means you've thown away say half the spatial resolution compared to FullHD but a HDTV has to scale it back up, the badness factor is 4.
Convert to 720x576 SD PAL. You've now got a quarter of the original spatial resolution but the HDTV has to scale that up by 4, the badness factor is 4x4 = 16.
That's further compounded by the 720p being encoded with a H.264 codec compared to a MPEG-2 codec which is less efficient.

Bob.

So what is the workflow for using TMPGEnc when I'm using AVCHD 1440 X 1080 and putting that on a DVD?

Does TMPGEnc show up as a plugin on the Render menu, or is it standalone?

Can you use DVDA or does it re-render it and lose quality? Or is it better to use TMPGEnc Disc Author?

Render to native format of the source files from vegas. Then use tmpg to render to mpeg. DVDA will accept the file & it will look way better than what vegas can produce. Use vegas to render out the audio separately.

I'd like to thank everybody for the comments made here, which have contributed greatly to my understanding of the issues and the potential solutions. I will return to this thread in the future.

One thing seems clear, that I should acquire and use TMPGEnc as an intermediate step between native output from Vegas Pro and DVD Architect. I had no idea that this program existed.

Jim Brook

There's something that I still don't understand:

Why doesn't a professional application (as Vegas PRO) give us the best possible (professional) HD to SD downconversion quality? Why are we forced to use 3rd party software and complicated time consuming procedures to get pristine or even acceptable down-conversion quality? SCS proudly advertises that VP works with almost any video format - input or output. While this is true it seems that some of the downconversions are not up to par.

I have been wondering about this since my first HD-SD conversion a couple of years ago. There has been no improvements in this respect even if the version number has changed a couple of times... I understand that SCS uses some 3rd party codecs but could not those be improved on? I have recently compared some results rendered both using the same source material in both VP11 and Edius 6.5 and the difference is remarkable. Edius wins hands down... This should be a fair comparison since both applications cost about the same. Vegas wins hands down what comes to usability, but best possible render quality is also important.

Please SCS, put this on your to-do list.

Christian

"Render to native format of the source files from Vegas."

I have a Lumix GH2, which records AVCHD, that's just the container though, right? How would someone find the true native format?

Just to confirm MrBrook; I'm talking about TMPGEnc Mastering Works 5. There's a free trial. http://tmpgenc.pegasys-inc.com/en/product/tvmw5.html

You can either 1. Render an HDV file at highest settings in Vegas, then use TMW5 to render that down to MPEG2 for DVD. 2. Use the the Debug Frameserver method (do a search here) to render the Vegas TL to TMW5.
http://www.debugmode.com/frameserver/
As I mentioned in my last post, the only thing I miss with this workflow is the ability to 'smart render' the file if I've made a mistake. :-(

There has been a suggestion that the TMPGENC encoder is superior to MainConcept. All else being equal, they are about the same. I've run enough hours of tests a while back to feel comfortable with that statement, although one could easily state a preference for one over another.

The difference is in the filtering. The resizing algorithm in TMPGENC may be a little better, but the real difference (the one you will notice) is in the deinterlacing. Vegas uses Blend or Interpolate, which have been around since the stone tablets. TMPGENC uses some slightly better deinterlace filters and combinations.

That being said, anyone interested in more than a PHD solution (that's Push Here, Dummy) should thoroughly read Nick Hope's definitive work on this issue, as well as the volumes of background discussion linked in the first paragraph of his thread:

Interlaced HD to DVD AGAIN - some test renders

I have not run extensive tests, but my own eye sees a better rendered file from TMPGENC. It primarily shows up on dissolves and transitions. I have always had less artifacting with TMPGENC. But since I am too cheap to spend money on another encoder, I have been using Mainconcept with Vegas and have been happy with the results. As you said, "they are about the same".

Dave T2

I might add that I always use a minimum bitrate of 2,000,000 in Mainconcept, which would affect artifacting in fades and transitions as you brought up. The Vegas default of 192,000 is woefully inadequate. I have no idea what the default or typical minimum bitrate in tmpgenc is.

That brings up a starting recommendation in Mainconcept for the OP --
Two pass VBR
2,000,000 - 6,000,000 - 9,500,000

I always used a fixed rate, unless it is over an hour which is seldom.

Usually 7.5 or 8.

I know we have discussed this before. :)

Dave T2