DVD Studio - HD to SD hell!

[Randall Leong]

Originally Posted by Randall Leong

So, I have reverted to simply applying the Resize filter with "Interlaced" checked within the filter. The field dominance of the newly downconverted file, in this case, remains the same as the original HD interlaced footage (TFF/UFF). Modern NTSC DVD players will automatically convert any TFF footage to BFF before outputting through the S-Video or composite video outs.

As I previously stated, this workflow only works with native 1080i60 video. Also, as I previously stated, 1080p24 video encoded inside a 1080i60 stream should be IVTC'd first in order to produce a 1080p24 video file.

On the other hand, if I were working with a 720p60 stream, I have to determine what the original source footage was prior to the original conversion. One program was largely in 24p but the pulldown was done as a 1-1-1-2 pulldown instead of a 2-3 pulldown. And some of the source video footage within that 720p60 stream was originally interlaced.

[Randall Leong]

Originally Posted by Robert M Wright

I have no idea why you are separating the fields, resizing, and then recombining them. If you want to resize interlaced footage, and get interlaced footage as a result, simply check the box next to "Interlaced" in the dialog for resizing and, unless I'm utterly mistaken, the resizing algorithm you choose will be applied to the fields individually (no need for all the extra fooling around).

DVD video does not allow 60p footage. If you want 60i, then no deinterlacing should be done at all (at any point in the workflow). If you do want to deinterlace the footage (and make a DVD using 30p footage), with 1080 line video, you have enough lines to be able to simply drop a field (no further action required to deinterlace the footage) and yet have enough lines in each frame left to still be downsizing (result has a full 480 lines of resolution).



Of course dropping a field and then "deinterlacing" will result in a mess. Dropping a field deinterlaces the footage, in and of itself. Every line in each individual field is acquired at the same time. The individual fields are progressive images by nature. If you drop one of the fields from 60i footage, what you are left with is defacto 30p footage. Applying a deinterlacing filter, at that point, is completely nonsensical.



No. Not at all. I am in no way confusing 60i for 30p source footage. What I said was, if you downsize a 1080i60 frame (shot at 1/60 shutter) to 480 lines, without telling the resizing algorithm that the source footage is interlaced (2 individual fields per frame), doing so will effectively deinterlace by blending the two fields together, and the end result will be very similar to looking the same as if the source was shot as 30p at a shutter speed of 1/30th of a second and resized to 480 lines (entirely different sources yet a very similar end result when resized that way, to SD resolution).

Okay, I am trying to clarify this resizing procedure right now:

All of VirtualDub's native filters - including resize - work properly only if the source input is in RGB (4:4:4) color space. Unfortunately, the videos that most of us have to work with use only YV12 (4:2:0) color space. Therefore, the incoming video has to be converted to RGB before the VirtualDub filter(s) can correctly process the video. The color space conversion can be done in AVISynth using the "ConvertToRGB" script. This in itself is no problem because the resulting video can be re-imported into an NLE of your choice (most NLEs are designed for RGB color space imports from lossless AVIs) and then reconverted to the video format of your choice.

The big disadvantage to using VirtualDub for your processing is that it takes very long (because the program has to reprocess the videos and convert the color space during recompression), and some of the filters included in VirtualDub simply aren't very good. Which is why some video diehards separate the fields and resizing and reinterlacing (weaving) within AVISynth.

[Peter Manojlovic]

Alrighty......There should be a sticky about HD to SD converts....
Here's an old picture i drew up (i realize it's not a Picasso), to better understand myself as to proper video processing...

This is of course, interlaced HD resized to interlaced SD...

The fundamental requirements are in my diagram. NO TWO WAYS ABOUT IT!! Deinterlacers, and resizers are all different, and therefore become subjective...
Otherwise, simply resizing interlaced material will destroy information....

Speed issues are usually bottlenecked at the deinterlacer. But understand that there's a HUGE amount of information being processed...
Any other way of resizing will destroy information.

Programs like AviSynth allow you to do this in native YUY colourspace, while the filters in Vdub require RGB conversions...Unless you're using multiple colourspace conversions (YUV>RGB), it shouldn't be a hit in quality, and it's simply a matter of editor preference...

Randall wrote:

Unfortunately, the videos that most of us have to work with use only YV12 (4:2:0)

.
This is only if it's ripped from a DVD...Which in case needs to be treated with care, since colourspace is shared differently.

Otherwise, most camera's are recording YUV colourspace.
Most computers prefer number crunching in RGB, and final output goes to YV12..

[Randall Leong]

Originally Posted by Peter Manojlovic

Alrighty......There should be a sticky about HD to SD converts....
Here's an old picture i drew up (i realize it's not a Picasso), to better understand myself as to proper video processing...

This is of course, interlaced HD resized to interlaced SD...

The fundamental requirements are in my diagram. NO TWO WAYS ABOUT IT!! Deinterlacers, and resizers are all different, and therefore become subjective...
Otherwise, simply resizing interlaced material will destroy information....

Speed issues are usually bottlenecked at the deinterlacer. But understand that there's a HUGE amount of information being processed...
Any other way of resizing will destroy information.

Programs like AviSynth allow you to do this in native YUY colourspace, while the filters in Vdub require RGB conversions...Unless you're using multiple colourspace conversions (YUV>RGB), it shouldn't be a hit in quality, and it's simply a matter of editor preference...

Randall wrote:
.
This is only if it's ripped from a DVD...Which in case needs to be treated with care, since colourspace is shared differently.

Otherwise, most camera's are recording YUV colourspace.
Most computers prefer number crunching in RGB, and final output goes to YV12..

Also, for those who have TV tuner cards, MPEG-2 files of TV shows are also in YV12. And some non-DV consumer camcorders record in the YV12 color space. (I know for a fact that my CX100 camcorder records in YV12.) Of the new 2010-model Canon camcorders, only the HF-R series records in YV12 (this reflects in the maximum bitrate of 17 Mbps); the others record in YUV.

Anyway, most AviSynth filters can work in either YV12 or YUV, so there's no need to convert YV12 to YUV if using such filters (although you may want to convert to RGB after applying any AviSynth filters if your NLE prefers RGB input).

And I would have to agree that simply resizing interlaced content will result in artifacts or blur. And even if you separate fields, simply resizing the fields and then weaving is not enough. Each field should be converted (interpolated) to full-resolution frames using a deinterlacer like Yadif, then resized and have selected lines in each downsized interpolated frame dropped (this is where the "SelectEvery" command in AviSynth comes in) before weaving (re-interlacing).

Robert M Wright wrote:

I have no idea why you are separating the fields, resizing, and then recombining them. If you want to resize interlaced footage, and get interlaced footage as a result, simply check the box next to "Interlaced" in the dialog for resizing and, unless I'm utterly mistaken, the resizing algorithm you choose will be applied to the fields individually (no need for all the extra fooling around).

Actually, I found that the "interlaced" feature in VirtualDub's Resize filter still does not do this correctly. The resizing alogarithm in VirtualDub actually applies to the entire frame (same as if "Interlaced" isn't checked), but now it blends the two fields as it resizes. (Without "Interlaced" checked, one field would have been discarded and the other resized and duplicated.) This results in noticeably blurry videos in the downsize to SD. I used the filter on a few videos, and they all came out noticeably blurry - about as blurry as if I did my interlaced HD to interlaced SD downconversions using just Pinnacle Studio alone. Therefore, until someone comes up with a free or low-cost simple-to-use resizer for interlaced content which works properly, I still need a ton of fooling around using software which may require a Ph.D in rocket science just to get the downconversions of interlaced material right.

[Steve Rotter]

i'm kind of at the end of my rope. i thought i had a good thing going but my footage looks like good vhs tape footage and not something that was shot in 1920 HD scaled to 720 SD. i took a couple month break, thinking i resolved it. made a wedding dvd last week....and yes it looks good but i have done so much better! i switched to mac completely a year ago. went from pc and premiere cs4. my footage always looked great! pc didn't have all these little gotchas you had to select or these many combinations to remember to get good video. i'm at the point where i believe i can only achieve so much with the mac. that can't be.

i shoot in 1080 30p on XHA1. i make the entire timeline HDV1080 30, and export to QT movie via final cut FILE / EXPORT / QT MOVIE. i bring that into idvd. i stopped trying dvd studio pro and compressor doesn't seem to be much better.

my main thing, as silly as it sounds....i don't know when to deinterlace and when not. there is of course NO DEINTERLACE FEATURE in file / export / QT movie....only when using QT conversion but i don't want that since i want my chapter markers to come over.

so then i thought i would take the effect / filter of deinterlace and paste it to all my clips in the timeline. maybe in need to deinterlace, i thought. nope, same thing. the disc looks good but i'm not happy and my pc turned out outstanding work. with the pc it was clean and sharp and great! with final cut, it's like good vhs quality, a little soft, not as much sharpness.

i don't know if i should shoot in 30p or 60i. i have always shot in 30p and will continue to do so. so i shoot in 1080 30p, then capture in HDV30p, then export to QT movie with same settings....or should i change to pro res 422 duing export?

one thing that always confused me is the upper and lower fields. how do you know when to use what? do you have to deinterlace 30p or only 60i? and if so, there are no settings for deinterlacing during file / export....stupid. i'm lost and confused.

my dvd menus ... the drop zones...look pixelated and soft. it's driving me nuts. not sure what to do. film and edit in same setting, but then export to pro res 422 or keep in the same setting for export as well?
thanks

[Perrone Ford]

If you shoot progressive (the "p" in 30p) then there is nothing to deinterlace. Progressive is the opposite of interlaced.

And there are no fields in progressive footage either. So if you are shooting progressive, there are no fields to worry about.

Stunning footage can be produced on the Mac, but they sure as heck don't make it obvious. Mac software spends a lot of time trying to make everything "easy", but not nearly enough time trying to make things their "best". There should be a button for best that REALLY does do best, including the frame controls in Compressor. That seems to be the magic potion that makes this downres actually look decent on a Mac. But I don't have a Mac, so someone else will have to talk you through the workflow.

And for the record, the only NLE that I've found that has a decent downres was Avid. All the rest looked AWFUL.

[Bryan Sellars]

Just to add to what Perrone said, if you do decide to shoot in 60i then AVCHD is upper field first, which is the opposite to standard DVD so if you make a DVD at 720x480 from interlace AVCHD then I've found it better to change the default lower field to "upper field first".
If however you are just down scaling to 720P then 30P is probably preferably.

I to am not familiar with Mac and have never used Avid so can't compare, but from my limited experience with AVCHD 1920x1080i I've found the best program to convert to 720P is Premiere CS4 using the MainConcept H264 Video codec,

[Steve Rotter]

Thanks guys. So if I shoot in 30p I should do lower field first? I love the mac but cs4 was better as f'r as you got what you got! No hidden switches! Most people say to stay away from H.264 when exportng frm timeline. The quality is excellent though! The only gotcha to that is chapter markers you created on your timeline don't come over! And DVD pro doesn't see h.264 but iDVD does!

[Perrone Ford]

Originally Posted by Steve Rotter

Thanks guys. So if I shoot in 30p I should do lower field first?

How did you get this from what we wrote? If you shoot 30p there ARE NO FIELDS. They don't exist in your source. If the format you are converting *TO* requires fields (meaning it's going to interlace the footage) then follow whatever it demands. The only time you should be getting a lower field first option is if you are recording or converting to miniDV. Other than that, everything should be either progressive or upper field.

Does this make sense?

Originally Posted by Steve Rotter

I love the mac but cs4 was better as f'r as you got what you got! No hidden switches! Most people say to stay away from H.264 when exportng frm timeline. The quality is excellent though! The only gotcha to that is chapter markers you created on your timeline don't come over! And DVD pro doesn't see h.264 but iDVD does!

It's not an issue of the quality. If you are going to DVD, then the ONLY format DVD understands is Mpeg2. If you export using the proper templates from FCP or Compressor, or whatever, you make a compatible Mpeg2 file that goes straight into your DVD authoring program and no further conversion is necessary. If you export with mpeg4 or some other format, you footage will be compressed once when you export it from the NLE, and then be compressed AGAIN when you try to author the DVD since it *will* change it to Mpeg2. There is a quality loss at both those compression steps so you want to try to avoid that if possible.