HDV --> SD DVD Workflow (DebugMode FrameServer, AviSynth)

Report · Jul 02, 2008

While the old thread contains a wealth of great information, I've decided to break away from it. It's too old, too long, too confusing... In summary:

Many users are aware that Premiere/AME/Encore does a "less than stellar" job of converting 1080i HDV to interlaced SD DVD. I (and others) have tried every conceivable combination of options and found none that provided "professional quality" conversions. This is largely due to the way that Premiere handles scaling of interlaced material ... Example #1 ... Example #2 ... and Premiere's failure to convert from the HDV color matrix (Rec.709) to the SD specification (Rec.601).

In an effort to solve these problems, I developed a workflow (using several third-party freeware tools) that achieves results far superior to that of Premiere and/or Encore by themselves.

I have written a guide for this workflow, including step-by-step instructions and links to download all the required tools.

This page also contains a link to a more advanced option for doing these conversions (using mostly the same tools). Please do not attempt the advanced version unless you have already implemented my "basic" workflow!

The page and linked files will be updated frequently. It is a work in progress, but should already provide excellent quality. I look forward to comments and suggestions from this community (as always).

Enjoy! -- Dan

Report · Nov 02, 2010

Welcome to the forum.

You can use the CS4 workflow from my hd2sd tutorial with CS5. HDSLR footage works just fine.

-Jeff

Report · Nov 25, 2010

I'm sorry guys. I'm a bit off the topic ... Here:
http://forums.adobe.com/thread/727162
you can help me to publish a 64-bit version of DebugMode.
And many of your tricks finally will come to an end. And you`ll can simply export video from CS5 with a frame-server ...

Report · Dec 08, 2010

Hi Jeff - I tried the approach outlined in your CS4 video for 720p60 content and the result is a sped-up video. Was one of the steps wrong? Like the one that says to bring the 59.94 fps Intermediate video onto a new sequence that is at NTSC DV, 29.97 fps?

I'm going to try it again...but I thought I followed the tutorial word-for-word the first 3 times I tried it, so I'm wondering if something wasn't right.

Thanks in advance.

Report · Dec 08, 2010

For CS5, none of the CS4 workarounds are necessary. Just export a

native 720p60 sequence to MPEG2 DVD from Pr using Max Render Quality and

you should be good to go. Pr automagically takes care of the scaling

and interlacing consecutive frames. You shouldn't set your clips in the

720p60 to ICF any more - Pr handles that on export.

Let me know if you continue to have issues with the CS5 export.

-Jeff

Report · Dec 09, 2010

Ok - well that's great news! Thanks Jeff.

One thing though....what settings do I need to have to ensure Pr does the best conversion? Specifically Field Order which defaults to Lower Frame. Should that be set to Progressive? And Pixel Aspect Ratio?

Format: MPEG2-DVD

Preset: NTSC Widescreen High Quality

Quality: 5

Frame Rate: 29.97

Display Format: Drop Frame

Field Order: Lower Frame (or Progressive?)

Pixel Aspect Ratio: Widescreen 16:9 (1.212)

Render at Maximum Depth: UNCHECKED

Bit Rate Encoding: VBR, 2-pass

Minimum Bit Rate: 2.8 Mbps

Target Bit Rate: 7 Mbps

Maximum Bit Rate: 8 Mbps

M Frames: 3

N Frames: 15

Use Maximum Render Quality: CHECKED

----------

Also - for this particular project, it's a green screen video, so I'll need to do some compositing in After Effects. Is the best workflow to start in AE using the source video (720p60), then export from there matching the source settings and then bring it back to Pr to assemble the timeline before exporting to MPEG-2...?

Report · Dec 10, 2010

I've tried just about every possible combination in the Adobe world and I am getting decent results when I export using the above settings (from my last post) with the Field Order set to Progressive...with one significant problem. The actor's glasses don't look very good on a TV at all. The frames on the glasses still seem to be artifacting...for lack of a better term. Something just doesn't look right. I don't know if this may be due to using an older DVD player connected with component cables or if I'm still missing something in my Export settings...? It seems to be ok on a computer monitor...

Since I'm going from 59.94 fps to 29.97 fps, do I need to have Frame Blending checked?

Also wondering if there's any benefit of checking the Render at Maximum Depth box...? Although I don't think that would help this issue.

I'm really itching to try the hd2sd method, but since it doesn't work with CS5, I guess that's not an option...

Any advice would be greatly appreciated!

Report · Dec 10, 2010

do I need to have Frame Blending checked?

No, you *don't* want to check Frame Blending. You'll get horrible results.

if there's any benefit of checking the Render at Maximum Depth box...?

Read here:

http://forums.adobe.com/message/3331847#3331847

I'm really itching to try the hd2sd method, but since it doesn't work with CS5

Why do you say that? It works just fine with CS5. What problems have you encountered?

-Jeff

Report · Dec 10, 2010

Thanks again for the reply. At first I found that the debugmode framesever doesn't work with CS5, although I've seen that someone is developing a solution for that now. BUT...

I just watched your latest tutorial that says you can do it without using the debugmode frameserver by exporting a Lagarith AVI file directly from Premiere. I guess that was the step I was stuck on. Sorry about that. I've been reading so many forums on this topic, I think I'm losing it.

I will try that now and see what happens!

Thanks for all of your help and dedication Jeff. You are a huge asset to the entire video community.

-Brian

Report · Dec 10, 2010

Ok - I went through and following all of the steps you recommend for 720p60. I used DGMPGDec and then brought the d2v into Virtual Dub. Now I have my 720x480 Lagarith AVI. This doesn't look good on a computer screen, not that that matters since it's not the final video...unless it does and something is already wrong...

My question now is, how do I do the next export to MPEG-2?

Do I start a new Pr sequence using DV-NTSC Widescreen or Standard? If Widescreen do I then Interpret Footage to get it to fill the frame?

And on export I'm now choosing Interlaced? UFF?

Man this is frustrating to get right the first time...

Report · Dec 10, 2010

Author and burn your converted video to DVD and play it back on various playback devices: computer, SD CRT, HDTV. If it looks good on all 3, or just on the TVs, then you know you've got it right.

If it turns out wrong, there are a couple of hd2sd parmeters that you can add and tweak to ge the proper results.

-Jeff

Report · Dec 10, 2010

Hi Jeff ---

It somewhat interesting that I would be asking you about this, but it seems you have more recent and relevant experience with the process than I do! I notice that you are using/recommending Lagarith exports as opposed to high-bitrate MPEG2. I think this is a very good idea and I am a big fan of the Lagarith codec.

However, I have noticed some issues (sporadic frame corruption) with Lagarith decoding when using MT in AviSynth. I am wondering if you have experienced this also and/or experimented/found a solution. One possible solution (one that seems to work but I have not tested extensively) is to use MT mode 6, like:

SetMTMode(6)

AviSource("lagarith_hd.avi")

SetMTMode(2)

hd2sd()

I also just noticed that HC Enc has a new version available that performs YUY2 -> YV12 conversion. I need to find some time to play with. If it works as it should, then I would suggest always using OutputColorSpace="YUY2" when outputting from hd2sd() and letting HC Enc handle the YV12 conversion.

BTW – I am finally now using CS5 on my awesome new Core i7 950 with 24GB of RAM. Ahhhhhh… it's about time!

Report · Dec 10, 2010

Hi Dan,

Nice to have you back!

What version of Lagarith? Since updating to 1.3.20, I haven't seen any corrupted frames. I also use UT a lot now, and I've never seen corrupted frames with that codec. Caveat Emptor anyway: If you have a UT source file, run it through hd2sd and export a UT avi file from VirtualDub, UT will bring VirtualDub crashing down around your ears. No problems when UT is either the source or the output codec, only when it is both.

Congrats on your new system!

-Jeff

Report · Dec 10, 2010

Very interesting… So, we can export from Premiere to an HD UT file, then process that in Avisynth using SetMTMode(2) and save it out as Lagarith AVI, correct?

Not to get terribly off-topic here, but I am really enjoying Premiere CS5 so far on my new machine. I have discovered that (at least in SD resolution) Premiere handles editing of Lagarith RGB files extremely well. I am quite pleased by this. So, for example, I compress my hd2sd() output to LAGS AVI files, which I then edit losslessly in Premiere. The process works very well for me.

Report · Dec 10, 2010

Very interesting… So, we can export from Premiere to an HD UT file, then process that in Avisynth using SetMTMode(2) and save it out as Lagarith AVI, correct?

Yes. And you can actually use UT as both the source and the output codec *if* you do:

SetMTMode(3,0)

DirectShowSource("MyUTfile.avi")

SetMTMode(2,0)

hd2sd()

But that's getting beyond the scope of my tutorials...

It's worth pursuing UT workflows since UT encodes out of Pr at least 3x - 4x faster than Lagarith.

Be sure to check out the new Debugmode Frameserver thread. The 64-bit version for CS5 is in development, and several of us have donated to the programmer so that he can make it happen. Once it's done, Satish says he'll host and distribute it. We had to go to an independent developer because Satish had neither the time nor the desire to do it.

-Jeff

Report · Sep 26, 2010

The link from the original post 'a guide for this workflow' is no longer working, and appears to be exactly what I need. Is there a new link, or has this HDV-->SD DVD Workflow been abandoned?

Thx,

Dave

Report · Sep 26, 2010

Try the tutorials here:

BDVS Tutorials

-Jeff

Report · Sep 28, 2010

I have now done intensive testing using DVCPRO HD 60i originals using Adobe Premiere Pro CS 5. Having followed the tutorials and the workflows suggested, for footage that has a lot of pans and zooms which just loves to create artifacts when going to SD, I did not find this procedure satisfactory at all (although it seems to be fine for general use). Possibly there is some script or filter that would improve this but I do not know of it.

WHAT I DID:

I tried using DVCPRHD60i media starting in both a standard timeline preset for DVCHPROHD 1080I60 with 1.5 PAR and I also tried it in a 1920x1080 XDCAM EX 1080i60(HQ) to start out with square pixels.

In both cases, following Jeff's tutorials and Dan's Hd2SD filter, I created MPEG2 I frame (matching original frame sizes in each case) m2v files. I then created the .d2v files with DGIndex. Using AVISynth with Virtual Dub created interlaced avi file (using logarith codec) with basic avs script for interlaced. I then put the new avi file in a dv ntsc timeline in pr pro 5 and exported as mpeg 2dvd to encore.

However,

WHAT I FOUND TO WORK BETTER:

What I found to be far more satisfactory and which did NOT reproduce motion artifacts in 95% of the time, was the following workflow:

Import original dvcpro 1280x1080 dvcprohd 60i footage directly into Adobe Media Encoder without going via any timeline.

Format: Uncompressed Microsoft AVI

Video settings: 720x480

29.97 fps

progressive (yes! progressive! interlaced looked like crap in the end)

aspect D1/DVNTSC

video codec: v210

uncheck both render max depth and render maximum quality (did not find significant improvement in quality with either except for longer rendering times.)

Believe it or not, the key and most important aspect is the progressive setting in the above step. Crazy, but true.

Then import to Pr Pro CS5 and drop into DV NTSC timeline and edit away.Simple.

Right click on media in timeline and use field options > FLICKER REMOVAL as needed which might as well be 99% of the time.

(Note: using flicker removal on the original dvcopro footage didn't produce same results. Needs to be the avi file)

One has to be very careful here, for although the footage may look beautiful in the timeline in premiere pro, the moment you make your m2v file and import to Encore it will look like hell. So flicker removal is needed most of the time and yes, it will soften the footage slightly, but that is a given. And only use the flicker removal by rightclicking.

When done with your edit, export timeline to AME encoder using the MEPG2-DVD format, NTSC HIGH QUALITY preset, leaving everything else the same except use

QUALITY 5

NTSC

720X480

VBR, 2 PASS

TARGET BIT RATE 7

MAX BIT RATE 8

Other settings remain at 29;.97 fps, drop frame, lower field order, 4x3 standard PAR, NTSC.

Sorry, I tested this very extensively with a lot of different footage and the above does a better result than what I found with the hd2sd steps. I am not saying this is the best, nor am I saying that HD2SD doesn't work, but I repeatedly get DVCPRO HD footage and it has creatred a lot of heartache for me in the past and the above is what works for me. And anything better would be great.

I am using Windows 64bit pro, 18GB ram.

Report · Sep 29, 2010

That's an interesting workaround! You must have spent a LOT of time on this.

You took your AVI from VirtualDub into PPro and then exported it for Encore. Did you try either HC Encoder or TMPGenc to create an MPEG for Encore, rather than taking the AVI back into PPro? I found HC Encoder did wonders for the quality. Have not tried TMPGenc but Ann says it's even better.

But if your AME - PPro - AME - Encore workflow does it for you, then stick with it!

Report · Oct 17, 2010

Dan,

Does your sharpen parameter use the built in AviSynth sharpening filter? I have heard that there are excellent sharpening filters available, such as Limited Sharpen. Do you know if it is worth the trouble? How are external filters used with HD2SD in AviSynth?

Thanks.

Report · Oct 17, 2010

1.Hi --

Interesting question about the sharpening. Yes, it uses AviSynth's built-in sharpening filter. This is certainly not very sophisticated sharpening – and it can increase noise in the video. However, after much testing, I decided that it is "good enough" for most purposes and – for its ills – produces rather natural-looking, moderate sharpening very quickly.

Some things to consider when using it:

1.) In an effort to reduce artifacts, all sharpening is performed after deinterlacing (if applicable) but before scaling. In my opinion, this is really the best way to perform sharpening during the conversion process. If you would like to change this behavior, you can set SharpenPost=true (it is false by default), which will perform sharpening after the video has been scaled to SD. This can be advantageous when the default "false" option does not provide obvious enough sharpening. Rule of thumb: If your video requires a very high sharpness (such as Sharpness=0.75 or greater), consider adding SharpenPost=true and trying a lower value, such as Sharpness=0.3 (as Post-Sharpening will be more obvious at lower values than Pre-Sharpening).

2.) If you are sharpening the video, you should probably use some noise redcution also. Take note of the different NR options and how the affect the "stacking" of processing. (e.g: Setting NR=5 will apply noise reduction to the source [so it goes NR->Sharpening->Scaling], while NR=3 applies it to the output [sharpening->scaling->NR]).

3.) I did not use LimitedSharpen or any other "fancy" sharpening in my function. I tried a few of these out, but I found them either to be painfully slow or ineffective or both. It would also result in way too many optional paramters and make things more difficult for me to troubleshoot between different revisions, multithreading, etc. If you would like to try any of these, you'll need to do it manually.I suggest the using post-sharpening in these cases. Although it can make some artifacts more apparent, it is much easier to accomplish and processing will be much faster. Assuming progressive output:

hdsd(OutputFieldRate=30)

LimitedSharpen()

For interlaced output, you'll need to output 60Hz progressive (or 50p for PAL) and then re-interlace after sharpening:

hd2sd(OutputFieldRate=60, interlaced_out=false)

LimitedSharpen()

# Output BFF [change to 4,0,1 for TFF]

SeparateFields().SelectEvery(4,1,2).Weave()

Report · Oct 21, 2010

Hi guys, not been on in a while, but I've re-read the whole thread, and I cant seem to find an answer to what I'm hoping is a simple question.

I've been using the basic hd2sd method for well over a year, and I LOVE IT!

My workflow was:

Editing on a quad core PC (xp) 3 gig ram, plenty of HD space

HDV (1440*1080i (25, pal land) source footage on PPRo CS3

export using framserver with this script:

AviSource("l:\signpost.avi")
ColorMatrix(mode="Rec.709->Rec.601", clamp=false)
ConvertToRGB32()
SmoothDeinterlace(tff=true,doublerate=true,lacethresh=24,staticthresh=35,staticavg=50,edgethresh=35,blend=false,showlace=false)
Trim(1,0)

# This assumes NTSC. Change the next line to Spline36Resize(704,576) for PAL format
Spline36Resize(704,576)
AddBorders(8,0,8,0)
ConvertToYUY2()

AssumeTFF().SeparateFields()

# This will create lower-field first output. Change to SelectEvery(4,0,3) for upper field first
SelectEvery(4,1,2)
Weave()
Limiter(16, 235, 16, 240)

# uncomment the next line only if you are using HC Encoder or QuEnc
ConvertToYV12(interlaced=true)

# uncomment the next line only if you are using TMPEG
# ConvertToRGB32(interlaced=true)

(I think this is the older 'basic' workflow, but its the only one I could get to work, and it works sooooo good!)

The used HC encoder to create DVD ready m2v file (and a brilliant m2v file it was)

.

Now I've change my cameras to the new canon dslrs (love them) so my source footage is now 1920*1080p (25p) - I convert the native dslr .mov files to cineform avi's if that makes a difference in the process of HD2SD?

So can I ask what settings i need to change in eiother / or my script or HC encoder now that my source footage is 1920*1080p (25p) so I can continue making lovely looking DVDs?

Thanks again for all the help, and apologies if this has been covered already,.

Cheers

James

Report · Jan 11, 2011

Hi, Dan! This link: http://invertedhorn.axspace.com/hdv2dvd_basic.html in your first post is not work, can you update it?

PS. Sorry for my english.

Report · Jan 11, 2011

http://3dvp.com/hd2sd_sd2hd.zip

Report · Jan 11, 2011

Big thanks, Jaff!!!

Report · Jan 20, 2011

I've been using hd2sd for awhile now with great results but I have run into a source format that's not working perfect. Maybe you can help.

It appears to be DVCPRO50, but it's HD so does that make it DVCPRO HD? This is how it shows in MediaInfo:

General
Format                           : MPEG-4
Format_Commercial_IfAny          : DVCPRO 50
Format profile                   : QuickTime
Codec ID                         : qt
File size                        : 35.0 GiB
Duration                         : 1h 48mn
Overall bit rate                 : 46.0 Mbps
Encoded date                     : UTC 2011-01-10 20:29:09
Tagged date                      : UTC 2011-01-10 20:50:50
©TSC                             : 23976
©TSZ                             : 1000

Video

ID                               : 1
Format                           : DV
Format_Commercial_IfAny          : DVCPRO 50
Codec ID                         : dvhp
Duration                         : 1h 48mn
Bit rate mode                    : Constant
Bit rate                         : 48.8 Mbps
Width                            : 936 pixels
Original width                   : 960 pixels
Height                           : 702 pixels
Original height                  : 720 pixels
Display aspect ratio             : 16:9
Frame rate mode                  : Constant
Frame rate                       : 23.976 fps
Original frame rate              : 29.970 fps
Standard                         : NTSC
Color space                      : YUV
Chroma subsampling               : 4:2:2
Bit depth                        : 8 bits
Scan type                        : Interlaced
Compression mode                 : Lossy
Bits/(Pixel*Frame)               : 3.100
Stream size                      : 34.9 GiB (100%)
Language                         : English
Encoded date                     : UTC 2011-01-10 20:29:09
Tagged date                      : UTC 2011-01-10 20:50:50

I don't really understand "frame rate" vs "original frame rate". I thought this was 720p24 but after running through HD2SD (using interlaced=false) the audio I was given does not match up perfectly, then i looked at it in mediainfo and I see it says it is interlaced. When i run using interlaced=true, the resulting file has a frame rate of 47.95, which I can't say i've ever seen. So given above, how would you process this file using hd2sd?

thanks

Ryan

Adobe Community

HDV --> SD DVD Workflow (DebugMode FrameServer, AviSynth)