1440 * 1080 vs 1920 * 1080

[Peter Dunphy]

Uh-oh upon importing the asset into DVD Studio the footage has that horrible horizontal line interlaced look. Would this be as a result of using a ‘bigger resolution than 1440? Back to square one I fear :o/

[Peter Dunphy]

AS things stand I’m just going to run the entire sequence thru the compressor with the 1440 * 1080 setting chosen

[Michael Gissing]

The horizontal lines are most likely to be a field order problem, not anything to do with 1440 v 1920. FCP can handle the different pixel ratios internally. It is just a matter of having the distort values correct in the motion tab.

In the end you are using Compressor to create a standard def mpeg. Search this forum for field order and DVDSP posts. There must be at least three or four a week dealing with field order on DVD via Compressor.

[Peter Dunphy]

Thanks for your reply Michael – I chose ‘Progressive’ in my Compressor settings and it’s finally seemed to work okay. Will check out the threads you suggested :o)

[Michael Gissing]

Unfortunately, many chose progressive to solve their field order problem, but in doing so you are throwing away resolution. If you want a progressive look, that is another matter.

[Peter Dunphy]

Hi again Michael thanks for your valuable advice,

Below my signature is some relevant information for the reference of anyone reading this who has the same issues I obtained relating to other Creative Cow forum members’ issues by searching through this forum.

First of all, I have to admit I haven’t a clue about ‘field orders’ and interlacing despite reading a WIKI entry and another article about them, so please bear with me. I’m trying to get my head round it all but I’m learning as I go :o) I’m not sure what ‘field order’ my Canon XH-A1 films in, or if the field order changed once I converted the HD footage to AppleProRes422, and then to Quicktime, and then to M2v and ac3.

Any suggestions whatsoever about my below resolution predicament will be heartily welcomed with open arms and a pint of Guinness over here in Belfast.

In eventually choosing ‘Progressive’ in the Inspector in Compressor to avoid the horizontal banding I got, I would like to find out the way to get a better, clearer image on my final (DVD Studio Pro) DVD with better resolution than I achieve by choosing ‘Progressive’.

I will now describe what steps I followed to get my finished movie into DVD Studio Pro (albeit in ‘Progressive’ mode with resolution less than I should be getting) – perhaps your good self or anyone else here can tell me where I’m going wrong. It’s a promotional video I’ve made and I would love it to look as crystal clear as possible :o)

I exported my ProRes422 sequence (Originally HD footage from a Canon XH-A1 HD 1440 * 1080) from FCP timeline using the ‘Export’…’Quicktime Movie…’. The Setting was ‘Current Settings’ and Recompress All Frames was not* selected and Make Movie Self-Contained was* selected.

As a side note, to get my HD footage into a ProRes422 FCP sequence, I referred to the tutorial here that I found via ‘Google’:

https://edu.moviola.com/hdv_prores#new_easysetups

I then added the resultant Quicktime .mov file to Compressor, and chose the Best DVD Quality option. In the choices of field order both the bottom first and top first options caused the picture to have those horrible banding horizontal lines across it. Progressive seemed to work okay and I was able to use the resultant m2v and ac3 files to add to DVD Studio Pro.

The quality of this output seemed okay to me, but if it’s possible in any way to achieve better quality, I’d really appreciate any workflow tips so that I may go back and fix things.

Michael you mentioned “It is just a matter of having the distort values correct in the motion tab.” Which motion tab are you referring to? Perhaps it is something in DVD Studio Pro that I may have overlooked? Perhaps, as mentioned to someone else in this forum, “…you could also skip the compressor step, and just send a quicktime to DVD SP and let it compress it. ” ? Maybe that would be a safer step for me so I wouldn’t have to worry about any field order or resolution issues?

Some forum members have mentoned using ‘interlacers’ for the footage. I’ve looked up Canon Xh-A1 product details but can’t determine for definite what field order it uses. Anyway, I’m not sure whether the field order of the footage gets changed once I log and capture the HD footage into FCP as ProRes422 :o/

Lastly, from the forum I came across this “If you are editing DV then and only then should you convert the DVD mpeg2 files to DV codec files which are lower. All other PAL codecs are upper.” With regards to this, in my FCP timeline I had a combination of DV and HD footage, so I’m not sure what happens to the field order in that case. I had exported the entire ‘mixed’ sequence from FCP as a Quicktime file, and then ran the Quicktime file through Compressor to convert it all from a ProRes422 mixture of 1440*1080 (HD) and 1920*1080 (DV), to a ProRes422 1440*1080 sequence.

I’ve noticed the below Ken Stone link in this forum about exporting a HD timeline to Apple Pro Res 422 using Quicktime conversion, but i don’t know if it’s a better method than the ‘log and capture’ method where HD footage is immediately captured into FCP as ProRes422?:

https://forums.creativecow.net/thread/8/1005646#1013701

Again, any tips/suggestions will save me having to save up to get my hair stapled back in after all the clumps I’ve pulled out while trying to get my HD footage look the best it can via DVD SP :o)

All the best

Peter

——————–

Related info. I’ve sourced from this forum:

“It is possible you have field dominance settings wrong in making the DVD. In PAL everything in upper field dominance except DV codec.

Unless we have some basic info about how you are making your DVD, we can only guess at this stage. Check your work flow from the sequence settings through to the DVD burn process. The field switch will be in there somewhere.

Deinterlacing seems to be the mantra to fixing this common issue but it is throwing away resolution and potentially creating motion artefacts on movement. Creating a progressive look is another issue but must be thought about before shooting, not after.”

“In PAL (720 x 576) only DV is lower field. Convert the footage to the codec you are already editing with if this is to add to an existing sequence.

If you are starting a new sequence and this is the only footage then don’t chose DV unless that is all you can run on your machine. DV50 is better but I would go straight to uncompressed if you have a reasonable machine and decent drives. DVDs are mpeg2. Transcoding via any other codec will increase image loss. If you haven’t the drive space, then consider ProRes. ”

“Are you trying have your finished product scan progressively? If so, you should do the interlace to progressive conversion in Compressor.”

“If you are working with HDV, just stop “Export with Current Setting”.
You shouldn’t export to HDV unless you need to print to tape.
Haven’t you heard yet about the advantages of ProRess?
As Borjis points if you need to de-interlace use a good filter to do it. Default de-interlacers, in general just throw away half of the picture information. ”

“If you export a reference movie (not a self contained) movie and import that to Compressor, you can assure that your field order will be taken care of.”
“Generally when I have used Compressor, I NEVER use the “Send to Compressor” function and it always results in issues. Best to Export a Quicktime Movie from your timeline, you can do a reference movie, and then take that manually into Compressor.”
“I prefer to do as Walter suggest: Export a Self-contained movie and send it to Compressor or any other appilcation (I’m using BitVice for MPEG-2). When you send something to get a “Double pass” in Compressor, everything have to be rendered twice, even if your sequence have been fully rendered.”
“Instead of using the deplorable QT conversion menu, just export your HDV sequence as HD, and do the MPEG-2 down convert in Compressor.”

“-Export your FCP project to Compressor
-In Compressor, find all of the DVD options and choose the one you want. Drag the one you want into the batch window.
-Make sure you also drag a destination into the batch window, too. Compressor needs to know where to put the assets when it’s done. Using the desktop is fine.
-When Compressor is done, you should see two new files on your desktop: an .m2v(video) file and a .acm(audio) file.
-Launch DVDSP, and delete the two default assets that appear in the window.
-Drag your two new assets into the window.
-Right-click the icon and select “First Play”.
-Click the burn button (I forget the exact name). It should open your DVD tray. Throw in a dvd and off you go. ”

“you could also skip the compressor step, and just send a quicktime to DVD SP and let it compress it. “

[Michael Gissing]

Peter, the Guinness sounds very inviting. Pity I am in Tasmania, Australia.

HDV footage is usually interlaced. JVC cameras are the exception. Canon shoots interlaced and I assume you are 25 fps.Interesting link for the Easy setups. So you have converted all your Canon footage to ProRes 1440×1080. Viewing the files in your edit bin, it should show you your field dominance which should be upper.

Check that your footage in the timeline doesn’t have the shift fields filter and also check that your sequence is set to upper field in the sequence settings. If your edit sequence is ProRes 1920 x 1080 then the ProRes that you captured from the camera will have a value set in the distort tab. Double click a shot into your viewer and click on the motion tab. Under distort it should display 33.3. What this setting does is stretch the 1440 to 1920. If your sequence is 1440 then the distort will be zero on the camera footage and the other files that you made which were 1920 will show a distortion of -33.3 as they are scaled to the sequence. FCP usually does this automatically.

So the fact that you have 1440 and 1920 in the same timeline is usually not an issue because FCP corrects the aspect for you. So to trace the field issue, check the footage is upper, the sequence is upper and the resultant Qt file is upper (stay in the same codec as your sequence). In Compressor your mpeg encode should also be set to upper field. With all those set correctly, you shouldn’t have to deinterlace to fix your problem.

In the 25 frame world it is simple. All video except DV is upper.

[Michael Gissing]

I just looked again at the easy setups on offer from that web link. None are for 50i material so I hope you didn’t use one of those to capture your footage.

[Peter Dunphy]

Hi Michael,

Thanks for your reply, really appreciated. Well if I’m ever in Tasmania I’ll be sure to bring some proper Guinness over with me in my suitcase :o)

I’ve taken your suggestions on board and included my findings below. If you notice anything incorrect, or anything that can be tweaked to get better quality I’d really appreciate you flagging it up:

The Capture Preset i used was HDV-Apple ProRes 422
The Device Control Preset i used was HDV 1080i50 Firewire Basic
The setting for Edit to Tape/PTV Output Video was HDV (1440 * 1080) 50i

———-
The sequence settings are:

HD (1440×1080) (16:9)
Pixel Aspect Ratio: HD (1440×1080)
Field Dominance: Upper (Odd)
Editing Timebase: 25

Quicktime Video Settings: Compressor – Apple ProRes 422, Quality 100%

Video Processing: Render 10-bit material in high precision YUV

———–

I’ve now checked my ProRes 1440×1080 clips in my edit bin and can now see that the Field Dominance is indeed Upper (Odd).

On each of the HDV clips when I clicked ‘Remove attributes’ there was nothing available (no ‘filters’ option /shift field filter available to remove). However, on the DV clips (which I had expanded from 4:3 to 1440 * 1080 in Compressor), when I clicked on ‘Remove Attributes’ the ‘Distort’ option was available to remove.

In the motion tab for the DV clips the Aspect Ration shows as (plus) 33.33

In the motion tab for the HD clips the Aspect Ration shows as (zero) 0.

So, I’ve checked the footage is upper, the sequence is upper, but the Quicktime Movie Inspector shows the following information (doesn’t seem to mention whether the movie is Upper or not):

————

Format: Apple ProRes 422, 1440 x 1080, Millions, 16-bit Integer (Little Endian), Timecode…
Normal Size: 1920 x 1080 pixels
Current Size: 1920 x 1080 pixels (Actual)

————

At this point, the face that the footage size has grown from 1440×1080 to 1920×1080 concerns me. The size I want the footage to be is 1440 x1080, in keeping with my sequence settings. The bigger size could make things appear too ‘blown up’ (such as selecting View>Double Size) in a Quicktime Viewer Window! Well, maybe not as exaggerated as that, but hopefully you see my concern. I

*** …the resultant Qt file is upper (stay in the same codec as your sequence) ****

I’m not sure how to double-check in QT that the resultant QT file is Upper. However, I added the QT file to Compressor and noticed it states the Native Field Dominance is “Top First”.

An interesting response I read on another forum about another person’s issue was “upper and lower in the same timeline that is not good. The only thing that you could do is add a deinterlacing filter to one or the other. If you are scaling down the HD stuff to match the SD footage I would create a lower feild first SD timeline and then deinterlace the HD footage and scale it down to match the SD footage.”

mpeg encode should also be set to upper field

*** In the 25 frame world it is simple. All video except DV is upper. ***

For the DV clips in the sequence, should I perhaps add a shift field filter to them? the Item Properties for each of the DV clips states:

Field Dominance: None
Vid Rate 25 fps
Frame Size 1920 x 1080
Compressor Apple ProRes 422

Once I added the Quicktime Movie File to Compressor and added the DVD Best Quality 90 minutes setting, when I selected only the MPEG-2 Mbps 2-pass Setting for my movie and clicked on Inspector, I got the following information:

Name: MPEG-2 6.2Mbps 2-pass
Description: Fits up to 90 minutes of video with Dolby Digital audio at 192 Kbps or 60 minutes with AIFF audio on a DVD-5
File Extension: m2v
Estimated file size: 1.37 GB
Type: MPEG-2 video elementary stream
Usage:SD DVD
Video Encoder
Format: M2V
Width and Height: Automatic
Selected: 720 x 576
Pixel aspect ratio: PAL CCIR 601 (16:9)
Crop: None
Padding: None
Frame rate: (100% of source)
Selected: 25
Frame Controls Automatically selected:
Retiming: (Fast) Nearest Frame
Resize Filter: Linear Filter
Deinterlace Filter: Fast (Line Averaging)
Adaptive Details: On
Antialias: 0
Detail Level: 0
Field Output: Same as Source
Start timecode from source
Aspect ratio: Automatic
Selected 16:9
Field dominance: Automatic:
Selected Top first
Average data rate: 6.2 (Mbps)
2 Pass VBR enabled
Maximum data rate: 7.7 (Mbps)
High quality
Best motion estimation
Closed GOP Size: 12, Structure: IBBP
DVD Studio Pro meta-data enabled

—————–

In the “Encoder’ area of the Compressor Inspector it states:

Video Format: PAL
Frame Rate: 25
Aspect Ratio: 16:9
Field Dominance: Top First

——————

in The ‘Geometry’ area of the Compressor Inspector it states (albeit greyed-out):

Dimensions (encoded pixels)
Frame Size: 720 (Width) 576 (Height) ‘Automatic’ –
Pixel Aspect: 0.7031 PAL CCIR 601 (16:9)

—-

The resultant m2v file details are:

Type: MPEG elementary stream
Duration: 0:30:09
Data Size: 1.43 GB
Bit Rate: 6.79 Mbps
Video Tracks:
224 MPEG-2, 720 × 576, 16:9, 25 fps, 7.70 Mbps, upper field first
Audio Tracks:
128 AC3 2/0, 48 kHz, 192 kbps

———

It is this M2v file that I intend to send to DVD StudioPro.

—–

in DVD StudioPro I’m not too sure of which settings are best for my video, so in the meantime I’m opted for:

SD DVD Menus, Tracks, and Slideshows:
Display Mode: 16:9 Letterbox

HD DVD Menus, Tracks, and Slideshows:
1440 x1080i
Display Mode 16:9 Letterbox

Encoding:
Aspect Ratio 16:9
Field Order Top
Mode Two Pass VBR
Bit Rate 4.0 Mbps
Max Bit Rate 7.0 Mbps
Motion Estimation: Best
Method: Background Encoding selected

—–

How does this look? Previously I’d selected Field Order “Auto” in the ‘Encoding’ area of DVD StudioPro.

If there’s anyway I can boost quality and get a clearer image by modifying any step in my ‘workflow’ I will embrace it. Perhaps I could use the Bit Rate sliders?

Again, any ideas welcome.

All the best

Peter

[Peter Dunphy]

As an update, I went from Final Cut, to Quicktime, to Compressor, to m2v…ensuring that the field order each time was upper field first. When I played the m2v file in MPEG Streamclip I got the horizontal banding issue. When I checked the ‘stream info’ in Mpeg Streamclip it detailed:
——-

Duration: 0:30:09
Data Size: 1.37 GB
Bit Rate: 6.53 Mbps

Video Tracks:
224 MPEG-2, 720 × 576, 16:9, 25 fps, 7.70 Mbps, upper field first

Audio Tracks:

Stream Files:
NEWTESTm2v.m2v (1.30 GB)

———

So I’m thinking that perhaps the ‘corruption’ (horizontal banding) occurs as a result of the DV and HDV being mixed into the same timeline? :o/