Ok I have to make an SD dvd and have always just gone from FCP straight to compressor. But others have said the imge is "muddy", and although I have been relatively happy with the finished DVD, I am sure there is room for improvement.

So a couple of weeks ago I bookmarked a thread on here which discribed how to do it. Fine, I thought, as soon as I make the DVd I will use this method. Here it is below.

-------------------------------------------------------

The key I have found is to make sure you start with progressive footage! It always looks bad when you go from Upper Field first HD to lower field first SD. If you have HDV in 1080i instead of progressive, then here is what you need to do:
1) Take your edited HD sequence in FCP.
2) Create as others have said an SD sequence, but MAKE SURE THE FIELD DOMINANCE IS SET TO NONE!!! It usually defaults to lower field first!!!
3) Drag your HD sequence to it, render and output.

If the SD sequence is set to lower field first then it will look HORRENDOUS! Try it out...make one lower field, drag an interlaced HD sequence into it, you'll see all sorts of weird jaggies and bad interlaced artifacts. Then click "Sequence settings" and then make it "none" for field dominence. You'll then need to click on the timeline somewhere and it'll "snap" into great clarity!

It took me a LONG time to figure this out! Now, I give you this knowledge for free!! :) Seriously, this may not be the issue your having, but it was for me!
-----------------------------------------
With thanks to Mr Fredericks.

So I have a problem. I have done al my editing in HD, I make a new SD sequence, then try and cut and paste the HD timeline onto the SD one. A popup says "for best performance your sequence and video should be set to the format of the clips you are editing. Change sequence settings to match the clip settings?

If I say yes the SD timeline turns into a HD timeline. If I say no, I get a popup, "operatin not allowed, you cannot nest sequences with different editing time bases".

Now I tried to just drag the "sequence1" onto the sd timeline. Thats what I get.

If I cut and past the edited timeline, it corrupts and the video and audio go slightly out of sinc by 0.1, and tiney gaps appear between clips. Guess that's not the way to do it.

So what is the right way? Where have I gone wrong?

Thanks in advance.


__________________

Sounds like an editing issue, not an EX issue.

Not sure if I can replicate your issue but having recently shot in HD (720) for a SD/NTSC project I allowed Compressor to do the downconversion before submitting to the client. I left the sequence in HD until that point.

Another way I would have tried is cutting the project on an HD timeline, opening up another timeline, pasting, selecting the sequence settings, choosing the appropriate SD format, rendering the sequence in the new SD format and then applying gfx, titles and other paraphenalia. Is this what you did?

Both of these seem to work fine for me. I have also used MPEG Streamclip to great effect when downconverting.

Jus.

It also sounds like your using FCP 5. Is that right? FCP 6 allows the use of varying codec/formats/framerates within a single sequence.

Jus.

I recommend that way:

1. Edit on a (interlaced) HD-timeline
2. export with Compressor to progressive SD with 50fps (PAL) or 59.94fps (NTSC) and ProRes-Codec.
3. Convert the result with Compressor to SD-interlaced and mpeg2-codec.

That works because Compressor proper converts interlaced->progressive(->rescale) and vice versa, but because of some kind of a bug it has problems with interlaced->rescale->interlaced.

But if you know from the very first that the final result will be SD, you should shoot with 720p/50 or 720p/60.

It also sounds like your using FCP 5. Is that right? FCP 6 allows the use of varying codec/formats/framerates within a single sequence.

Jus.

No 6.04. Very frustrating. I can cut and paste the footage but it messes it up, opens up fraction of a second gaps between clips etc.

I recommend that way:

1. Edit on a (interlaced) HD-timeline
2. export with Compressor to progressive SD with 50fps (PAL) or 59.94fps (NTSC) and ProRes-Codec.
3. Convert the result with Compressor to SD-interlaced and mpeg2-codec.

That works because Compressor proper converts interlaced->progressive(->rescale) and vice versa, but because of some kind of a bug it has problems with interlaced->rescale->interlaced.

But if you know from the very first that the final result will be SD, you should shoot with 720p/50 or 720p/60.


Hi Dom, I shoot 720P50. I am currently trying a new method to see if there's an improvement.

Do all editing in 720P50, export as a Prores 422 ( not hq), bring that into compressor and get Compressor to downconvert it to DVD.

Tried it before and got a bit of aliasing. Will know in about two hours.

Tried it before and got a bit of aliasing.
It's very important that you change the rescaling-algorithm in Compressor from bilinear to optimal. Then you'll get perfect results without any (noticeable) aliasing.

I think it's trying to tell you that you've got an HD sequence with edit time base, say, 23.98 and your SD sequence is set up at , say 29.97. Just match the SD edit time base to the one your using for the HD sequence and you should be fine.

Use these SD sequence settings:

Frame Size 720x480/ NTSC DV 3:2
Field Dom: None
Edit Time Base: 29.97
Compressor: DVCPRO50/NTSC
advanced: 29.97/Progressive/16:9

Just made a bunch of SD DVD's with these settings after dropping a default XDCam sequence into this SD sequence. You'll need to render before outputting, but you can skip Compressor and just output using QT, self-compressed optional.

The key I have found is to make sure you start with progressive footage!
1) Take your edited HD sequence in FCP.
2) Create as others have said an SD sequence, but MAKE SURE THE FIELD DOMINANCE IS SET TO NONE!!! It usually defaults to lower field first!!!
3) Drag your HD sequence to it, render and output.


Steve, that is the way to go. No copying, no pasting though. Just drag the HD Edit sequence into an SD sequence as if it were a simple clip. Render and you are done. Bingo.
To get a sharp SD video use ProRes422 or MotionJPEG A as your target codec for the SD sequence.

Or as Justin put it correctly: Output an HD video and downscale that using MPEGStreamClip. Very easy. Absolutely crisp results. Target codecs again ProRes422 or MotionJPEG A.

Hope this helps. pe.

I think it's trying to tell you that you've got an HD sequence with edit time base, say, 23.98 and your SD sequence is set up at , say 29.97. Just match the SD edit time base to the one your using for the HD sequence and you should be fine.

Use these SD sequence settings:

Frame Size 720x480/ NTSC DV 3:2
Field Dom: None
Edit Time Base: 29.97
Compressor: DVCPRO50/NTSC
advanced: 29.97/Progressive/16:9

Just made a bunch of SD DVD's with these settings after dropping a default XDCam sequence into this SD sequence. You'll need to render before outputting, but you can skip Compressor and just output using QT, self-compressed optional.


OK you sussed it. Being UK bound I shot 720P50, and I was trying to drop this onto an SD timeline of PAL 25 FPS. I changed it to 50 and the sequence dropped on the sd timeline with no problems. I will now see how this comes out.

Then I will try Dominiks method and change the algorithm in Compressor.

Thanks to all.

It's very important that you change the rescaling-algorithm in Compressor from bilinear to optimal. Then you'll get perfect results without any (noticeable) aliasing.

Can't find this anywhere in Compressor. Is it called something else? If not, WHERE IS IT!!!!

Cheers
Steve

Can't find this anywhere in Compressor. Is it called something else? If not, WHERE IS IT!!!!
I don't know how it's exactly called in the english version of FCP, but perhaps this picture helps:
http://www.dominik.ws/compressorrescaling.png

Then I will try Dominiks method
The method I posted was for interlaced footage. But your footage is 50p, so do this:

1. edit on a 720p/50-timeline in FCP
2. select export->compressor
3. in compressor choose your SD-mpeg2-preset
4. change the rescaling-quality and, if you want to retain the 50fps-information, select interlaced output (odd fields first) in the mpeg2-settings and frame-settings

or

1. edit on a 720p/50-timeline in FCP
2. select export->quicktime->ProRes
3. use the result as source in compressor
4. in compressor choose your SD-mpeg2-preset
5. change the rescaling-quality and, if you want to retain the 50fps-information, select interlaced output (odd fields first) in the mpeg2-settings and frame-settings

I don't know how it's exactly called in the english version of FCP, but perhaps this picture helps:
http://www.dominik.ws/compressorrescaling.png

OK Thanks Dominik.

Hi Steve, I've found that I get excellent results doing edits in the native XDCam EX timeline and then doing an HD export of the final edit directly out of FCS. I use (File/Export/QT Movie/"Current Settings & Make Movie Self Contained". Then I drop this movie into an DVDSP project set to SD settings. The results are great. I've tried the other route dropping the HD into an SD timeline and can't see any difference in the results. I've also tried using Compressor with no noticeable improvement. KISS applies, I suppose.

As for your problems editing with erratic results in FCS, I'd be certain that your timeline corresponds to the settings of your clips. To do this, as you no doubt know, you can let FCS do the settings by creating a new sequence and editing in a clip from your HD footage. If the settings aren't the same, a FCS dialog will come up asking if you want to reset them to the clip's settings, which I'm happy to do. I've been cutting EX1 footage on a MPB 2.2 core duo for several months now and haven't seen any glitches or erratic behavior similar to what you've mentioned, so I suspect it's something in your setups.

Hope this helps. As with so much of this stuff there are any number of ways to the same conclusion.

Hi John thanks for the info.

I have now made a dvd by editing in HD timeline, exporting to sd timeline and out via compressor. Looking at the results of comparing this method and going from a HD timeline straight to compressor is striking.

The DVD striaght from the HD timeline to compressor is slightly muddy and looks poor. Lots of mosquito noise around moving images.

The DVD which has gone from the HD timeline to SD timeline, rendered and out to Compressor looks very clean, much less mosquito noise and a cleaner all round image I am very happy with.

In the next day or so i will try the other method of making a QT movie and compare which I think is best.

The DVD striaght from the HD timeline to compressor is slightly muddy and looks poor. Lots of mosquito noise around moving images.

The DVD which has gone from the HD timeline to SD timeline, rendered and out to Compressor looks very clean, much less mosquito noise and a cleaner all round image I am very happy with.
This makes absolutely no sense. Are you sure the mpeg2-settings were each time the same during these tests? Were the output-resolutions each time 720x576? Progressive or interlaced output?

The recommended method - well, one that I adhere to is:

I edit in native HD (whatever flavor, XDCam Ex, ProResHQ are my main two). I use a second HDTV monitor using intensity card (to view the canvas window), so editing in SD is a downgrade for me. I then send out a Quicktime export, using the same settings as the timeline - video only. I make my .MOV a reference movie so FCP is not overtaxed. You can also export a self-contained movie for archiving later - but it's a wopping file size.
I then import the .MOV into compressor and process in SD - usually 5.5 to 6.0 bitrate. This goes to a .mv2 file which I use in DVDSP. I can batch process in Compressor and just let it run in the background - usually I process overnight, as I often work on 60 to 90 minute movies and I run a first version dual-MacPro.
The audio I send to Soundtrack Pro from FCP, where after mixing it will export out as AC3 in surround.

I get stunning results - on my 1080p 50inch TV - the SD looks great. This is the method that Apple techs recommend and one I have seen many times in seminars I have been to.

This makes absolutely no sense. Are you sure the mpeg2-settings were each time the same during these tests? Were the output-resolutions each time 720x576? Progressive or interlaced output?

Totaly the same. Why bother going through all this rigmarole if sending the footage directly from a HD timeline produces perfectly good dvds? ( Which it doesn't)

The purpose is to get the best loking dvd possible, and by dropping the hd footage on an sd timeline, render and then out, gives a much better looking dvd.

, and by dropping the hd footage on an sd timeline, render and then out, gives a much better looking dvd.
This can't be, because the mpeg2-encoder is the same in both cases and Compressors rescaler is at least as good as FCPs, so there must be something faulty with your testing.
Can you post short example-results of both workflows? Perhaps then I can tell what the matter is.

This can't be, because the mpeg2-encoder is the same in both cases and Compressors rescaler is at least as good as FCPs, so there must be something faulty with your testing.
Can you post short example-results of both workflows? Perhaps then I can tell what the matter is.

This is the exact method I used.
Outputting Standard Definition in FCP from the Sony XDCam EX1 (http://www.kenstone.net/fcp_homepage/ex1_sd_output_young.html)

The recommended method - well, one that I adhere to is:


I get stunning results - on my 1080p 50inch TV - the SD looks great. This is the method that Apple techs recommend and one I have seen many times in seminars I have been to.

Jon,would you workflow work for ntsc?