Not converting Interlaced footage to Progressive if viewing on BD player?

[jagabo]

I knew V210 was 10-bit, but didn't realize it is MJPEG.

Where are you getting that? V210 is uncompressed video.

Seems in order for the Motu to capture it needs Adobe

A lot of capture programs can't deal with 10 bit input.

[clashradio]

I knew V210 was 10-bit, but didn't realize it is MJPEG.

Where are you getting that? V210 is uncompressed video.

Seems in order for the Motu to capture it needs Adobe

A lot of capture programs can't deal with 10 bit input.

When I uploaded the screen shot from the Adobe capture screen, you replied saying that it's best to capture lossless, not MJPEG. I assumed you were referring to the V210 as that what was reported. The MediaInfo screen shot also reports V210.

I guess Motu is designed to work with Adobe. I didn't know that when I bought it. But I happened to have Adobe on my computer so it's been capturing fine.

Vegas seems to do fine with the raw 10-bit AVI. But I do end up having to remux the audio because when I drop the AVI file in the Timeline, the audio track doesn't appear. I had this problem when I was using the Blackmagic Intensity Shuttle but chalked it up to the fact that it captures audio at 24b/48k. The Motu is 16/44.1 so it really doesn't make sense why Vegas doesn't recognize the audio.

[jagabo]

MJPEG is a lossy compression codec. V210 and UYVY are uncompressed video formats.

Try installing UT Video Codec. It supports 10 bit 4:2:2 YUV. It may show up in your software.

[clashradio]

MJPEG is a lossy compression codec. V210 and UYVY are uncompressed video formats.

Try installing UT Video Codec. It supports 10 bit 4:2:2 YUV. It may show up in your software.

You're saying install UT Video Codec and use it with Adobe and then maybe Vegas might recognize the audio track?

[jagabo]

MJPEG is a lossy compression codec. V210 and UYVY are uncompressed video formats.

Try installing UT Video Codec. It supports 10 bit 4:2:2 YUV. It may show up in your software.

You're saying install UT Video Codec and use it with Adobe and then maybe Vegas might recognize the audio track?

No, I'm saying UT Video Codec will reduce your file size to 1/2 or 1/3 that of uncompressed video without any loss of quality. MJPEG, being a lossy codec, is losing quality. You want your captures to be the best quality you can get. Then you filter and encode to make your files for viewing and distribution.

I don't know what's causing your audio problem. I don't use Vegas.

[clashradio]

Ok, I misunderstood. I thought you meant using UT Video might help with the Vegas audio issue. So far using the V210 codec hasn't been an issue with larger file sizes as I delete the raw files before I start a new project. I've got two, 6tb drives for storage.

Oh, I think I know why you mentioned not using MJPEG...one of the tests I uploaded was captured using the Blackmagic Shuttle, set to capture MJPEG. All of the recent capturing I've been doing has been with the Motu using V210. Sorry, I forgot I did one recent capture test using the Shuttle & MJPEG. I also wanted to compare captured file sizes using 10-bit for the Shuttle and the Motu. They matched pretty close except the Motu's captured AVI was a little smaller. Probably due to the Shuttle's audio of 24/48 & Motu's 16/44.1.

[clashradio]

What do you do if you have a ntsc laserdisc of a UK movie? example Beatles Let It Be? It was filmed on 16mm I'm guessing at 25fps? The LD is from a 35mm blow-up print. I assume you would IVTC, but at 23.97?

[jagabo]

PAL to NTSC conversions usually require deinterlacing followed SRestore() to remove blended frames, delivering 25 fps:

WhateverSource()
QTGMC() # or Yadif(mode=1)
SRestore()

But this is a minefield of variations.

[manono]

What do you do if you have a ntsc laserdisc of a UK movie?

Post a 10 second sample with movement, one from the source.

[clashradio]

Here is a 11 second clip captured at 486i, 10-bit 4:2:2 V210

[jagabo]

That Beatles clip is normal telecined film. Use

LWLibavVideoSource("Let It Be.avi", format="YUY2")
AssumeBFF()
TFM()
TDecimate()

[clashradio]

Thank you jagabo. I'll have to really try to learn AVIsynth.

[clashradio]

A few more questions:

1) Shouldn't I see a 3:2 repeat pattern in the Let It Be file since it's normal telecined film?

2) I assume for MiniDV footage if I want the best quality I'd need to apply QTGMC? Do you think it's best to transfer via Firewire and keep it digital; or should I capture via s-video at 10-bit uncompressed?

3) if you don't know if the original source on the laserdisc is film or video, how do you know when to IVTC or QTGMC?

[jagabo]

1) Shouldn't I see a 3:2 repeat pattern in the Let It Be file since it's normal telecined film?

Yes. Beware that many players/editors will automatically deinterlace DV for display. So you may not see it in your editor.

2) I assume for MiniDV footage if I want the best quality I'd need to apply QTGMC?

What matters is what the underlying video is. Film footage telecined for TV, then captured via DV device, remains telecined video and can be inverse telecined. Interlaced sources captured via DV device remains interlaced. I have an old DV camcorder that can shoot interlaced video (where each field comes from a different point in time, 59.94 fields per second) or progressive video (where the camcorder shoots 29.97 progressive frames per second and encodes as interlaced DV). So that fact that it's DV is irrelevant.

Do you think it's best to transfer via Firewire and keep it digital; or should I capture via s-video at 10-bit uncompressed?

If you're talking about something that was recorded to DV tape you are best off capturing via firewire. The data on the tape is already digital video compressed with a DV codec. Capturing via firewire is little more than a file copy.

3) if you don't know if the original source on the laserdisc is film or video, how do you know when to IVTC or QTGMC?

You examine the video frame by frame (be sure you editor isn't "helping" you by deinterlacing on the fly), and field by field. True interlaced video (suitable for QTGMC) changes with every field and every frame looks interlaced (when there's motion). Telecined film does not change with every field and some frames may look progressive.

[clashradio]

If I have a multi-camera shoot using miniDV do i qtgmc each .avi file or deinterlace the final edited project?

[manono]

If I have a multi-camera shoot using miniDV do i qtgmc each .avi file or deinterlace the final edited project?

You don't have to filter them all individually. You can easily join them together to make a single video. You can even select the parts you want to keep at the same time:

A=AVISource("Video1.avi")
B=AVISource("Video2.avi")
C=AVISource("Video3.avi")
D=AVISource("Video4.avi")

A ++ B ++ C ++ D#Join the videos into a single video, keeping the audio synch

Trim(0,1000) ++ Trim(3000,4000) ++ Trim(6500,7500) ++ Trim(9500,10000)#an example of how to select the parts you want to keep

QTGMC()#bob the parts you're keeping

If you mention the other kinds of editing you're doing, you might find most or all of it can easily be done in AviSynth as well.

[clashradio]

[QUOTE=jagabo;2525777]

1) Shouldn't I see a 3:2 repeat pattern in the Let It Be file since it's normal telecined film?

Yes. Beware that many players/editors will automatically deinterlace DV for display. So you may not see it in your editor.

Just curious; if Vegas is automatically deinterlacing for display, then I shouldn't see the comb effect/scan lines correct? Since I do see these odd/even lines that probably means Vegas is not auto deinterlacing?

[jagabo]

if Vegas is automatically deinterlacing for display, then I shouldn't see the comb effect/scan lines correct?

That is correct. But be aware that it may treat different sources differently. It may know that DV is interlaced and automatically deinterlace the display, but huffyuv has no way of flagging it's interlaced so Vegas may not deinterlace for display. I don't use Vegas so I don't know what it does.

Since I do see these odd/even lines that probably means Vegas is not auto deinterlacing?

Yes, if you see combing it's obviously not deinterlacing that video.

[clashradio]

Since I do see these odd/even lines that probably means Vegas is not auto deinterlacing?

Yes, if you see combing it's obviously not deinterlacing that video.[/QUOTE]

That's what's confusing is I am seeing the combing, but not the 3:2 pattern.

[jagabo]

I am seeing the combing, but not the 3:2 pattern.

Are you talking about the Let It Be clip in post #40? I see a very obvious 3 progressive, 2 interlaced frames, repeating. Starting with IPPPIIPPPII...

Also, if you bob it you see a 3:2 repeat pattern except for the initial orphaned field: 01112233344...

Note that 3:2 isn't the only possible pattern. Some videos may use 2:3:3:2 (viewed as frames you see four progressive frames, one interlaced frame, repeating). To go from 24 fps to 25 fps (film to PAL video) a program may use 2:2:2:2:2:2:2:2:2:2:2:3, ie, 12 frames becomes 25 fields, which means 24 frames becomes 50 fields.

[clashradio]

Yes, the Let It Be clip;

Does orphaned field mean the fact that I captured a random frame/field and not the beginning of the movie?

When I scroll in my Timeline frame-by-frame I don't see any frames that are duplicating UNTIL I add Yadif. Then I get a 1,2,3,4,4 etc.

But before I add Yadif I do see interlacing on two consecutive frames: 1,2,3,4i,5i etc. But the frames that have interlacing are not duplicate.

Does the 3:2 pattern mean the frames repeat in that order?, or does it mean those frames will have interlacing?

[jagabo]

Yes, the Let It Be clip;

Does orphaned field mean the fact that I captured a random frame/field and not the beginning of the movie?

Yes. A video doesn't have to start with two fields from the same film frame. Film is often telecined, then edited as video without any consideration as to whether the cuts are made on progressive or interlaced frames.

When I scroll in my Timeline frame-by-frame I don't see any frames that are duplicating

You don't see duplicate FRAMES, you see duplicate FIELDS. Go back and read that long PM I sent you where I described how film becomes video. A quick recap:

To convert 24 frame per second film to 60 field per second the film frames need to be displayed for the duration of 2.5 fields (24 * 2.5 = 60). This is achieved by alternating the display of film frames for the duration of 2 or 3 fields, an average of 2.5 fields. If you literally displayed the film frames for 2.5 fields you would end up with a visible tears in some fields.

So, for four film frames

0, 1, 2, 3

We perform 2:3 pulldown:

0t 0b 1t 1b 1t 2b 2t 3b 3t 3b (t = top field , b = bottom field, top field displayed first, 4 film frames becomes 10 fields)

Notice that the first film frame, is displayed for 2 fields, the second film frame is displayed for 3 fields, the third film frame is displayed for 2 fields, the fourth for 3 fields.

When those fields are captured and digitized pairs of fields are woven together into video frames:

0t/0b, 1t/1b, 1t/2b, 2t/3b, 3t/3b (top field to be displayed first)

Note that the order of the fields is still the same and the sequence always alternates between top and bottom fields. When both fields come from the same film frame you have a progressive video frame. When the fields come from different film frames you have interlaced video frames. Removing the t, b and / notation:

00, 11, 12, 23, 33
 P,  P,  I,  I,  P

Every video frame looks different (when there is motion), you don't see repeated video frames. It's the fields that are repeated.

Does the 3:2 pattern mean the frames repeat in that order?, or does it mean those frames will have interlacing?

It means the film frames are displayed for the duration of 2 or 3 video fields. You can start with a 2, or you can start with a 3 -- you just have to alternate between the two to maintain an average of 2.5. Once the video has been edited (or captured starting at some random point) you don't know which the telecine started with.

[clashradio]

jagabo,

Would it be possible to run this 1-second clip into AVISynth for either IVTC or QTGMC? I'm curious to see how much better it can do vs. Yadif

[jagabo]

That video isn't properly interlaced. Someone has resized it vertically and the two fields are no longer separate. No deinterlacer is going to get you back to clean frames. It also looks like it's been through an PAL/NTSC conversion which adds another layer of problems. Those could be addressed if the fields weren't screwed up.

[clashradio]

I think those problems must be due to me. I took a section of the raw .avi capture and re-rendered that one second clip so the file wouldn't be too big. Plus I did resize it to crop out the dirty edges. I'll recapture that section of the concert and re-upload the unaltered clip.

But that leads me to another question; if I only use AVISynth for either IVTC or QTGMC depending on the source footage, render the footage to uncompressed .avi, and then drop that file in Vegas for editing, if I resize it to crop the rough edges, since that footage is now Progressive, it won't alter anything that was fixed by AVISynth correct? as long as I set the Field Order to None and Deinterlace Method to None (I know you're not extremely familiar w/Vegas).

I don't know what I did that make it look like a PAL to NTSC conversion? The LD is a Japanese disc, so it's NTSC and I captured at 480i 29.97. But in the Project Properties section I did change the Field Order to None when I went to render. Maybe I should have set that to Upper Filed First.

Thanks for taking a look at it.

[jagabo]

But that leads me to another question; if I only use AVISynth for either IVTC or QTGMC depending on the source footage, render the footage to uncompressed .avi, and then drop that file in Vegas for editing, if I resize it to crop the rough edges, since that footage is now Progressive, it won't alter anything that was fixed by AVISynth correct? as long as I set the Field Order to None and Deinterlace Method to None (I know you're not extremely familiar w/Vegas).

Correct. Of course, you could just do the crop and resize (or add borders) in AviSynth too.

[manono]

Correct. Of course, you could just do the crop and resize (or add borders) in AviSynth too.

That was my immediate thought as well - what's being done in Vegas that can't just as easily (and perhaps better) be done in AviSynth?

...render the footage to uncompressed .avi...

Use a lossless AVI codec, not uncompressed.

[clashradio]

Correct. Of course, you could just do the crop and resize (or add borders) in AviSynth too.

That was my immediate thought as well - what's being done in Vegas that can't just as easily (and perhaps better) be done in AviSynth?

Use a lossless AVI codec, not uncompressed.

Well I'm scared to death to learn AVISynth. I'm gotten used to Vegas. I originally planned on using AVISynth only for IVTC/QTGMC, but if it does a better job at other functions I'll try to figure that it too.

At the moment I can't remember if Adobe captures lossless or uncompressed. But I capture at YUV 10-bit 4:2:2 so that should be lossless? (if you're wondering why I use Adobe for capturing my Motu HD Express needs Adobe for capturing, but I like editing in Vegas). I suppose if I have to download AVISynth I might as well use VirtualDub for capturing? as long as it recognizes the Motu?

[clashradio]

jagabo,

Here is that same clip newly captured. When you get a chance if you can run it through AVISynth?

[manono]

I originally planned on using AVISynth only for IVTC/QTGMC, but if it does a better job at other functions I'll try to figure that it too.

if you can actually get QTGMC going in AviSYnth, most other filters (not all) are child's play by comparison, especially crop and resize. Maybe for color work Vegas might be useful, but not for simple procedures.

But I capture at YUV 10-bit 4:2:2 so that should be lossless?

Did you read jagabo's earlier comment on that question?

Try installing UT Video Codec. It supports 10 bit 4:2:2 YUV. It may show up in your software.

Do we know your intended final format? The chances are good it'll be 8-bit.

Your video sample is field-blended from a PAL source. Only AviSynth can unblend it to its original 25fps.