Microsoft DV vs MainConcept DV preview codecs - Premiere 6.5

[lordsmurf]

This is kind of vague to me. Going off things I see in briefs...

I've seen it mentioned before where some have replaced the Microsoft DV preview codec with a MainConcept DV preview codec in Premiere.

If this is true, and there are advantages to doing this, how do I go about changed the DV preview codec?

And what are the advantages to doing this?

[racer-x]

I've tried this with the Cannapus DV codec, and quite honnestly I couldn't tell a differance.

To use it, you'll have to install the codec, then when you open premiere and it asks what project settings, click custom and choose the codec in the video dropdown.

[q1aqza]

My understanding is that the advantage in quality is only gained when you output your project i.e. export the timeline to back to DV, then there is supposed to be a quality gain over the Microsoft DV codec.

I personally export straight to MPEG2 so I have no reason to do this although I tried the MC DV codec once out of curiosity and it screwed up Premiere 6.5 - curiosity killed the cat - what a true catchphrase !!

[proxyx99]

I have MC and tried Canopus as well. Frankly, can't tell the difference on a good quality material unless one would like to examine it under a microscope. For safety reasons I decided to stick to MS DV ver. (not MC) since I've noticed some issues with other tools that I use. To eliminate guesswork and minimize number of variables I switched back to default codec. Can't say anything bad about MC though. It is a really good codec for editing, I've never had any trouble. Encoded stuff looks really good and rich in color. My favourite though is the Pinnacle implementation of DV codec (it may very well be the MS codec although they've had their own Pinnacle DV codec in the past) in Edition 5 and Studio 8 (yes...). Pixelation seems to be really fine (finer then other codec versions), but again, it is my own subjective impression

[andie41]

For the past week or so, I have been trying to determine if a DV codec makes any difference and, if so, which is "better." My conclusion is the codec is used not just to render to another DV file, but the mpeg encoder uses the codec to read the DV. If you take a DV AVI into Vdub, select information, you can see which codec is being used.

The Microsoft codec seems to expand, or not limit, the luminance values to the NTSC spec. Of course, this can be accomplished with a filter, when encoding. The canopus codec also shows this effect. Canopus provides filters to expand or limit the color space. The Sony codec does appear to limit the color space. My understanding is NTSC can only handle values of 220 steps, unable to produce values from 0 - 16 or 240 - 255. So any values in these ranges are wasted. The brightness of the display device (TV) can be adjusted if the range is too high or low, to move the entire range up or down, but still cannot display more than the 220 levels.

I am still trying to sort all of this out. When playing the mpeg on TV, I found the Canopus render to look "nicer"; the TMPGenc render, with the Sony codec, looked fine until compared with the Canopus. When I looked closely, I realized some of the difference (perhaps all of it) was because of higher contrast in the Canopus render, even though some of the whites and blacks were lost. Overall it appeared "snappier", while the TMPGenc/Sony mpeg was a little flat.

[donpedro]

Can't you use different codec to read than capture ?

BTW: My uderstandig (cofirmed in one long post) is that capturing in DV-AVI is just plain data copy = no additional conversion and loss.

[andie41]

donpedro

As you stated, there is really not a capture with DV from camera through firewire, so the codec used would be the one in the camera - in my case Sony. When an app on the computer reads the DV file, the app uses whatever DV codec is active on the computer to decode.

From what I have seen, there are definitely differences in the way the software codecs handle luma values. The Sony codec appears to clamp the luma values (even more restricted than the NTSC specs); Canopus seems by default to allow values in the 0-255 range, but then includes a filter in its encoder to clamp these values to NTSC specs.

If you select the "output YUV data as basic ..." in Tmpgenc, it then expands the luma out to 0 - 255, even using the Sony codec.

The only codec I do not like is the Mainconcept; it appears to definitely provide sharpening and, as a result, in many cases increase noise, etc. You can always increase the sharpness on the TV , I don't want it encoded into my video.

These are my obsevations - still learning.