DV to DVD what bitrate do you use.

[Craig Tucker]

I have just bought a Pioneer A04 and intend to convert my miniDV tapes to DVD. I would like some feedback on what bitrates people use (CBR or VBR), and hence the length of video they can get on a DVDR 4.7Gb. I want to convert my tapes with no noticible loss in quality. 90 minutes per DVD would be nice but not essential.

Thanks for your participation.

[ItsMe]

I normally use VBR 6000 average - 8000 Max. You could use a lower average without much visible loss of quality. The best thing to do is to experiment on a short clip, encoding it at different bitrates until you find the bitrate that suites you.

[bugster]

If these are home video's made using a handheld camera then they are notriously difficult for mpeg-encoders to handle well due to camera shake. However, as ItsMe said, an average of 6000 and max of 8000 using Tmpgenc 2-pass VBR will get you 90mins on a DVD. Going above 6000 will probably not see much improvement.

[ItsMe]

Camera shake is indeed a big problem for mpeg, using a tripod whenever possible makes a world of difference though.

[DJRumpy]

I wonder how many dual layer DVD's the Blair Witch Project comes bundled on?

[snowmoon]

Loss of visiually perceptible quality is a subjective matter and depends alot on how you are viewing the final materal. Light pre-processing to take the DV-edge off of the video can do wonders for it's compressability. When viewed on a standard TV you would probably be unable to tell the difference.

AVISynth and VDub can be used for pre-filtering.

[teljeur]

I am using the Sony pc100 and the pinnacle dv500 DVD. To edit the movie i use premiere. From the timeline in premiere i frameserve to cce 2.5. Setting are AVG 2500 Min 0 Max 5500 Multipass 3 times.

So then we have a mpv file and a wav file. The wav file is converted to a ac3 file with sonic foundry soft encode.

With this settings i can have about 3 hrs of movie with Dolby Digital sound (upto 5.1) and the picture quality is perfect.

[albatros]

What about the resolution of input and output? Also other filters? Does it have any influence? What do you set?

[teljeur]

The input resolution and output resolution are the same PAL 720x576.

Only the image quality setting can make a differance in the file size, normale i use a setting of 5. Other setting i use are.

under the video tab.
Add sequence end code
upper field first
Zigzag scanning order
DVD compliant

Other settings are not used.

Hope this helpes.

[Xesdeeni]

When putting together the DVD of my daughter's first year, I found that 6.25Mbps, which fits 1 1/2 hours / DVD, was excellent quality when encoded with CCE 2.64 from interlaced DV. I used VBR (6.25Mbps avg), and left the default noise reduction on. I have read that zig-zag scan order is for progressive frames, and that the linear quantization is for MPEG-1, so I left those off.

I found that the same bitrate using the 2-pass VBR of TMPGEnc was not as good. It left some "pulsing" in static images that I think were occurring at the I-pictures, and there was some visible mosquito noise.

Unfortunately, everything I tried in the area of noise reduction was a bust. If I reduced the noise enough to have an effect, it caused degradation of the picture. Spatial noise reduction decreased the detail, while temporal noise reduction ended up with ghosting and streaking. I re-encoded the video at least two dozen times with various filters, and the above approach was really the best I could get.

This year I'm trying to use the progressive scan feature of the camcorder. Hopefully this will give me better quality when I make the second year DVD

Xesdeeni

[ronnylov]

I normally use CBR 8000 kbit/s and the resolution 704x576 with 16 pixels black borders on all sides (generated from FitCD). Then I can fit 1 hour without problems on a DVD-R.

If I need more play time I use VBR but with minimum 6000 kbit/s average and 8000 kbit/s maximum. For best results I use Canopus Procoder inmastering quality. I also use avisynth preprosessing with the Convolution3D filter. If the video is shaky I use DynaPel Steadyhand to stabilize the picture. I do the editing after DynaPel Steadyhand and I edit in MediaStudio Pro.

It is also possible to use the 352x576 resolution with average 3000 kbit/s if you can accept "almost VHS"-like quality on the output.

I do not deinterlace! I encode interlaced with bottom field first.

[Wizeman]

Contrary to popular belief, the Apple MPEG-2 codec is outstanding - especially when DV is source. I've done 720x576(PAL) at CBR2500 and had crisp, ultra-sharp MPEG-2, and it took about twice the movielength (a ten minute clip encoded in almost exactly 20 minutes). I know most of you don't use Mac, but... that's your loss

[snowmoon]

I normally use CBR 8000 kbit/s and the resolution 704x576 with 16 pixels black borders on all sides (generated from FitCD). Then I can fit 1 hour without problems on a DVD-R.

If I need more play time I use VBR but with minimum 6000 kbit/s average and 8000 kbit/s maximum. For best results I use Canopus Procoder inmastering quality. I also use avisynth preprosessing with the Convolution3D filter. If the video is shaky I use DynaPel Steadyhand to stabilize the picture. I do the editing after DynaPel Steadyhand and I edit in MediaStudio Pro.

CBR 8000 is a waste of bits for all but the most demanding sources. Just use a CQ or CCE once pass VBR Q=10 you will get more space and virtually no loss of quality. I also find that avisynth TemporalSoften(1,3,6) works well for steady scenes without bluring motion. I don't blacken my borders since my HDTV has less than 5% overscan.

Cheers