Encoding quality questions

[hapz]

I was wondering about the impact of various encoding steps.

I.e., is there a difference in quality between a Hi8 tape captured directly to a DV format via a converter box

versus

capturing the Hi8 tape first lossless with YUV2/HuffYUV and then converting the resulting avi files to an avi DV format using software such as MSP with the MainConcept or another DV codec?

By the way, what is the difference between DV Type 1 and Type 2?

Thanks.

[Dr Mordrid]

If I want to capture analog to DV for non special effects use I use my RT-2000 for the task, which I guess would be similar to using a converter box.

The utility of these analog =>DV files is pretty much the same as normal DV, but with a bit lower sharpness & contrast as would be expected.

HuffYUV is better if effects will be added as it's allows lossless compositing of the effects during the edit.

Generally speaking;

Type 1 DV uses DirectShow/OpenDML and can be used in Win2K or XP on an NTFS drive to create files >4 gigs.

Type 2 DV uses Video for Windows and is thereby limited to 2 or 4 gigs, depending on the codec & DV device drivers on older non-OHCI devices.

Variants of Type 2 exist that have different specs, but generally speaking this is what it comes down to.

Dr. Mordrid

[hapz]

Thanks Doc

[hapz]

Another quality question.

I render a losslessly compressed HuffYUV video after editing with MSP6.5 to a compressed format (lets say mpeg-2) to save HD space.

At a later time I load these mpeg-2 files into MSP to add some more mpeg-2 clips, possibly with additional transitions.

Will that degrade quality, or is MSP6.5 smart enough to recognize that the final format is equal to the source format and thus does not introduce further qualiy loss?

Would it be different if I'd mix mpeg-2 clips with HuffYUV clips in the editing process with the final format being mpeg-2?

[wkulecz]

My experience trying to use Video Studio or MSP to do simple mpeg2 edits suggests smart render is not too smart and seems to want to render the whole project. But all timeline and editing operations are so cumbersum when mpeg is the project type and mpeg files are on the timeline that I hardly consider it useful. YMMV.

Any transition requires re-rendering the overlap at the very least. The temporal compression of mpeg (unless I-Frame only) makes it difficult to edit so I'm not faulting Ulead for poor performance, just pointing out that editing mpeg is IMHO a method of last resort. This is why I think the "new" direct mpeg recording camcorders (disk or tape) are a step in the wrong direction.

I think its best to think of mpeg as only a final output format like VHS tape.

If you cut in the middle of an IPB sequence it may be necessary to decode and re-encode the entire project to make the final IPB sequence come out correctly. Perhaps Doc can elaborate as he knows more about the low level details of the mpeg format than I do.

--wally.

[Dr Mordrid]

Even I-frame MPEG can be difficult to edit because there isn't any standard GOP layout.

Sure, each compressor has it's defaults, as does each standard MPEG profile, but if anything differs even a smidgen then smartrender can't keep up.

IMHO it's always best to edit in an I-frame non-temporal format and only encode to MPEG once the editing is finished.

For "normal" video (no f/x) DV or MJPeg are fine.

For compositited effects in a DV project absent a 4:4:4:4 compositing engine like the RT-2000's it's better to capture the footage as uncompressed or high data rate MJPeg and add the effects in that format. Once this is done then the effects shot can be included in the DV project.

Dr. Mordrid