DVD picture quality from Digital8

[tuquet]

Hi,

I recently created some DVDs from my digital8 tapes using Ulead Studio 6 SE (which comes with the Firewire card). Though limited in features, the software has not given me any problems. However, the result DVD picture quality seemed not as good as the tape played directly from the camera using S-video. I captured in DV format and selected eveything at "best" quality.

Is the cheap/bundled software the culprit in conversion or am I imagining things? If software, does anyone have similar experience and how much difference can you tell when using more expensive software?

Thank you,
tuquet

[borghe]

You are using Microsoft's DV compressor. Some better ones are Adaptec's DVSoft or Radius SoftDV. Radius is a Quicktime compressor and Adaptec is a VFW compressor.

Microsoft's gives horrible quality right out the gate and the more times you recompress with it the more horrible it gets...

[kazzer]

You are using Microsoft's DV compressor

How do you choose the DV codec you wish to use?

I've always thought with DV you were simply transfering digital information to the PC, and therefore the quality would always be the same regardless of the software you use. Now im confused!

[rsuave5]

you are right the transfer of dv yo your computer is not a compression but a copying basically. the file should remain pristine and even through editing and resaving you should not lose quality. the microsoft dv "compression " being garbage is a foolish comment.

now encoding is another matter once you have your edited file ready to put to dvd then save it and encode with tmpgenc. for the better quality use 2 pass vbr and a bit rate from 6000 (min) to 8000 (max) if you are confused as to what these are then read here http://dvd-hq.info/Compression.html i dont use all this stuff like splitting audio etc. but i do use the 2 pass vbr and the higher bit rate as well as a few other things and my dvds come out looking awesome! i also use a digital 8 email me if you like for more info

[tuquet]

I am glad I am not the only one got confused. Let me give it a jab in making DVD from digital8 tapes: I would need to (i) capture, (ii) render, and (iii) burn. Editting and all other fancy stuff in between are for a better production but not required. I always thought the Ulead software I have would do all three but looks like I am wrong. Then which software/component involve in which step? Can I control the whole process from the same software assuming I am using a choice software for each process? I need some brief education, forgive me if this has appeared elsewhere multiple times. I did do some DD. Thanks.

[rsuave5]

well render is the wrong term, encode is the term and yes most software titles do it bad! tempgenc is the standard here and about everywhere. the process would be, capture,edit(optional) encode via tmpgenc, then author to dvd. youll find there are alot of software titles and i use 4-5 on every project it seems. tmpgenc is in the tools section to your left and i think youll find the encode with it to be magnificent. you can use the dvd template but its not so great better to set it up for a good encode. ulead will do all three for sure but the encode is definately sub - par i think the author (you call it burn) is also basic but i use ulead for some stuff.

[rkr1958]

The steps (and software) I use for DV Camera to DVD-R are:
1) Capture via firewire DV camera to PC using Studio 7.
2) Edit and save edit in .avi using Studio 7.
3) Encode to DVD Quality (MPEG-2) using TMPGEnc Plus.
4) Author using either Dazzle DVDComplete or Ulead DVD MovieFactory.
5) Burn to DVD-R using RecordNow.

[DVbuilder]

rsuave5 Wrote:

... better to set it up for a good encode.

What settings do you use for your encoding?

[tuquet]

Thank you all. I think I will need some time to experiment with different software.

[cdcox]

I use the same method as rkr1958 and I am very happy with the results. I also came across some software called DVFilmmaker that I am going to test. Adds extra step but it may be worth it?

http://www.dvfilm.com/maker/index.htm

[borghe]

rsuave - before you criticize, I suggest looking at the video file properties of a capture inside on Premiere or any other such program. It very much passes through a codec when it comes onto the computer. If I capture with MS it will say dvsd MS DV. When I capture with DVSoft it will say dvsd DVSoftTM. There is also a noticeable difference when you capture a program with Premiere's default and say, MotoDV which defaults to Radius SoftDV. There is no such thing as a "pristine" transfer from DV tape. The DV is sent over firewire protocol using DV but then still needs a codec on the other side to interpret. That is where the quality difference occurs. About the only "pristine" part of DV is that you can guarantee that the data being sent across the firewire interface is perfect everytime. But what the receiving codec does to it is entirely based on which one you use. In that respect Microsoft's is well known to be one of the worst DV codecs out there.

I accept your apology.

tuquet - I am not familiar with Ulead. In Premiere I can etll you to go to capture settings, IEEE1394/DV capture, VFW compression, and then select the codec you want to work with.

[rsuave5]

pleas search this site and read up on dv. it is a copy process and makes no difference as to the compression you use. its often argued and i wont begin another. but dv capture from a camcorder is not a compression at all. it is a file copy or transfer from one source to another, much like if you move a file from one folder on your computer to another the program you use to read this file is irrelevant and should not degrade the output file. i also use premiere 6.5 to cap and adding , using other codecs to capture with is silly and a waste. any compression happens from your camcorder

pristine- remaining in a pure state, uncorrupted (websters)

my dv captures are pristine and uncorrupted there is no such thing as perfect but pristine is acheivable and after i hit my 1000 th dv transfer and author all my custiomers agree the video is pristine. my last stating that the video was better than being there! i can provide samples and or tutorage for the ignorant.

if i must prove it and look up the links or copy from my text books then i suppose i will in the interest of education and shuttin you up


i put the tmpgenc settings in a link i posted earlier look um up but the most important is upping the bit rate and setting to 2 pass vbr

[Jester700]

I don't think Microsoft's current codec is that bad, especially if you're not doing multiple renders. There was one in the past that was horrid, but it was fixed up, for the most part. A DX update will fix it, IF you have the troublesome codec version.

But I use the one built into Vegas - MainConcept.

[tuquet]

All right, I have managed to download VS7, MF2 and TMPGEN, in addition to my VS6, to experiment with my production and I have achieved to desired level of quality but not without a bunch of questions. Here is what I did, my goal is to preserve quality without having to buy a lot more software before justifying the need:

1. Captured, editted, added effects using VS6. VS7 did the same thing with a bit more options. I have not done anything with audio since it is not my main interest right now.

2. Saved the whole thing to an avi file. I had field order A set as I read somewhere that this was "bottom first", and it did seem to work that way. I guess here is where the codec gets into the picture, where VS6 has to put everything I add/delete into one smooth product. BTW, I looked at the codec seclection and saw about 10 of them, had no ideas how and when they all got there, the default one which I used was DV type 1. I am not sure if others would be better, I do not have the list with me now.

3. Converted the AVI file using TMPGEN. At highest quality (motion search), a 2 minutes clip took 50 minutes to encode on my Intel 1.6GHz, 512MB RAM with WinXP. I tried again with the lower quality and it took about half the time, quality seemed not to noticeably differ on my Zenith progressive scan DVD player to the 36XBR450. I may stick with this.

4. Authored using either MF2 or VS6. Again, MF2 is more flexible but that is not what I would be jumping on right now. The smart-render in VS6 did work quite well once I chose and set the template right.

Over all, thanks to all your inputs/arguments I have learned quite a bit and am feel much better seeing the final result. Now, here are the questions:

1. Is 50 minutes to encode 2 minutes of video on my system right? Do I screw up/over-kill somewhere along the line? I used VBR 2-pass, average 6Kbps, min 2Kbps, max 8Kbps, highest quality motion search. The rest was similar to the recommend in the earlier webpage.

2. TMPGEN died on me once. I do not feel comfortable doing 90 minutes video and have it die on me at the 37 hour. Does anyone experience this and is there any fix? I encoded to "System (video+audio)" file. I tried "ES (video+audio)" but MF2 and VS6 would not recognize the video stream. Perhaps I use the wrong setting or was it just it? How much better is ES compared to System? If the difference is noticeable, I am willing to work that out, how?

3. Is it advisable to encode short clips and added effects later to form the video or created the whole video first then encode? With TMPGEN dying, perhaps short clips is the way?

Thank you,
tuquet