HDV to SD. Best method?

[Steve Crisdale]

Originally Posted by Kevin Red

I haven't been using mpg2 compression yet. I've been trying to get a sharp uncompressed downrezzed video first. But Ill try that, Steve.

I found that the unsharp mask works well to bring out the dark parts and sharpness.

Don't forget; the whole theory of editing HDV is - the number of processes you can reduce, will increase the final output quality.

That's where the capabilities of Cineform's CFHD intermediate codec come in. Or, for a different take with the same fundamental philosophy of doing as little as possible to the m2t stream there's Gearshifts proxies...

Unless you're using one of these options, you're already degrading the data by rendering to SD. As much as it's "uncompressed", you are losing data straight away in the down-sizing process, and that process must still make "decisions" about which pixels to keep and which to turf...

To then re-compress again with the DVD level MPEG2 is more than double dipping into the pool of available pixel information!!!

Keep in mind, that anything you do to HDV is going to degrade quality - ANYTHING!! It's a bit like audio CD rips of audio CD rips... regardless of how good digital is; there's still data degredation with each action.

That's why it's best to do all your editing/effects/titles etc. in either Gearshift proxy or CFHD AVI format... and the last thing you do is write to DVD level MPEG2. If you don't write a "DVD compliant MPEG2", your DVD authoring application may even attempt to compress all over again - so just be aware of that!!

If it does want to recompress your resultant MPEG2, render your edited clips to CFHD AVI, and load them directly into your DVD authoring app instead. That will keep the source material high quality while allowing the DVD authoring app to create it's version of "DVD compliant"!!

[Kevin Red]

You're absolutely right, I use CFHD AVI intermidiate.

[Laurence Kingston]

Originally Posted by Kevin Red

I haven't been using mpg2 compression yet. I've been trying to get a sharp uncompressed downrezzed video first. But Ill try that, Steve.

I found that the unsharp mask works well to bring out the dark parts and sharpness.

This may be your problem. You should always render HDV directly to DVD complient MPEG2 instead of to DV codec format. The reason being that the colorspace is different. Rendering to MPEG2 and avoiding the DV codec stage altogether looks noticably better.

[Michael Liebergot]

Larry, quick question.

After capturing your HD video in either Cineform codec or proxy codec and working in Vegas with the final destination being MPEG2 for DVD.

Would you work strictly in the HD 1080i template or work in the 4:3 or 16:9 DV template that the final destination will be in.

Especially if I were to mix 2 cameras for the final video, one camera let's say PD170/VX2100 and the other being FX1/A1.

I would be using the HD camera is main and SD camera as cutaway 2nd 3rd camera.

[Paul Wags]

Reading the post, I have to agree with you there about HDV straight to DVD. I have done heaps of testing and the best I can come up with is HDV Cineform straight to DVD using Procoder.

The test is
HDV to DV to DVD
or HDV to DVD.

Go here and have a look for yourselves, I used some grainy underwater footage to test it on.
www.ningalooreefteach.com/HDV.htm
Download the big pics and zoom in with a graphics program.

Paul

[Laurence Kingston]

Originally Posted by Michael Liebergot

Larry, quick question.

After capturing your HD video in either Cineform codec or proxy codec and working in Vegas with the final destination being MPEG2 for DVD.

Would you work strictly in the HD 1080i template or work in the 4:3 or 16:9 DV template that the final destination will be in.

Especially if I were to mix 2 cameras for the final video, one camera let's say PD170/VX2100 and the other being FX1/A1.

I would be using the HD camera is main and SD camera as cutaway 2nd 3rd camera.

You can do it either way. I like to work in the format the footage is in though.

[Phil Hamilton]

Originally Posted by Steve Crisdale

The 'secret' to any good quality DVD master MPEG2 is to get the Bit-rate as close to maximum as you can get it!!

So here's what I have found to work pretty well... Render As> Select template> MPEG2> DVD Architect WS, then hit Custom> Project tab - Video quality - BEST> Video tab - Video Quality...slide to HIGH - Variable Bit-rate - Two pass - Max. 9,300 - Av. 8,000 - Min. 6,200 .

Steve - I read somewhere and my head is spinning looking at all of this - that you should never set the bitrate past 8,000. I notice here that you're modifying the template using two-pass (WHY?) and upping the bitrate. Thoughts? ph

[Steve Crisdale]

Originally Posted by Phil Hamilton

Steve - I read somewhere and my head is spinning looking at all of this - that you should never set the bitrate past 8,000. I notice here that you're modifying the template using two-pass (WHY?) and upping the bitrate. Thoughts? ph

I've seen the odd bit of guff along the lines of "I wouldn't go above 8,000"... and there's many an app that has 8,000 set as the Default Setting. Just because it's the Default Setting doesn't make it the Optimum Setting!!

Default is "computer speak" for "will work for even the most inept operator" setting - the "we hope there's no way this klutz can screw this up and try to blame us because our setting pushed their underpowered, below specification hardware beyond it's limits" kind of formula taken by those who've become immune to the litigation emboldened incompotents they are sometimes forced to deal with.

Same case for Dual/Two-pass or however your favourite encoder wants to label it... Default setting is usually Single Pass. However; the second pass is an opportunity (at the cost of longer encode times - with the gamble that it won't optimize any further) for the encode algorithm to further refine the compression; reducing artifacts etc. Sort of like a 'second bite of the cherry'.

If you're concerned about the resulting file size from using increased bit-rate with CBR, then use VBR - but always adjust the Default Settings for Max., Av. and Min. because they are inevitably set woefully low.

If you've got the quality and bitrate in the original - why throttle the quality of what you make with it by not maximizing the final products' bitrate also?

[Phil Hamilton]

Originally Posted by Steve Crisdale

I've seen the odd bit of guff along the lines of "I wouldn't go above 8,000"... and there's many an app that has 8,000 set as the Default Setting. Just because it's the Default Setting doesn't make it the Optimum Setting!!

Steve - I used your settings to render HDV 60i to SD DVD. The result looks pretty good - where I think I could tell the difference was in the lighting. With more light - as usual - the picture is pretty darn sharp. I did not use an unsharp mask as suggested but might. The only other question I have is in the property settings for the Vegas project. Since ulimately I am going to DVD 29.97 interlaced using the template should I set de-interlace method to NONE in the project properties? I always use blend and if I know I'm going to 24p then yes that would make sense. Or is Vegas smart enough to bypass this setting UNLESS you're going to progressive output? ph

[Steve Crisdale]

Originally Posted by Phil Hamilton

Is Vegas smart enough to bypass this setting UNLESS you're going to progressive output? ph

I haven't used a piece of software yet that was smart enough to over-ride an inappropriate setting!!

If you render an interlaced final - blend mode should be "None"...

Only ever use blend mode when heading to progressive output.

If you preview your rendered interlaced DVD level MPEG2 on computer screen, remember that you're viewing it on a NON-interlaced device, so use VLC if you're unsure of whether you have a player that can automatically account for interlaced footage!!

[Phil Hamilton]

Originally Posted by Steve Crisdale

I haven't used a piece of software yet that was smart enough to over-ride an inappropriate setting!!

If you render an interlaced final - blend mode should be "None"...

Only ever use blend mode when heading to progressive output.

Yes. I was so used to keeping the BLEND on because I was also playing with 24p - which this particular project is not. It looks good and as I work with other pieces I will be sure to leave the De-Interlace set to NONE. tks for your help. ph

[Fred Foronda]

I've been converting HDV to SD straight off the camera onto Vegas and it looks great. Does HDV to cineform codec then to DVD looks any better than doing it straight off the fx1/z1???

[Douglas Spotted Eagle]

Originally Posted by Fred Foronda

I've been converting HDV to SD straight off the camera onto Vegas and it looks great. Does HDV to cineform codec then to DVD looks any better than doing it straight off the fx1/z1???

Were you using any editor other than Vegas, I'd say convert at the camera.
But Vegas does a much better conversion than does the camera...so let Vegas do the conversion if you can manage it.

[Steve Crisdale]

Originally Posted by Fred Foronda

I've been converting HDV to SD straight off the camera onto Vegas and it looks great. Does HDV to cineform codec then to DVD looks any better than doing it straight off the fx1/z1???

As Douglas mentioned above, it's sorta agreed amongst those of us who've taken the time to do the tests and check the results... that Vegas does a better 'down-convert' than the camera's encoder.

If you want the opportunity to organise your clips into a form that would be almost impossible to shoot dirrectly to tape, and colour correct or maybe adjust sound levels, add 5.1 audio, background music or commentary, transitions, titles and just about anything that you'd consider an enhancement to the raw video; before going to DVD ... you'd need to work in Vegas with HDV clips with an intermediate editing format or proxies.

Bottom line. If you are happy with exactly what's on tape, you could get away with the camera downsampling - just be prepared to be disappointed with how much better HDV edited correctly before downconversion using either CFHD avi or Gearshift proxies looks.

I've also actually wondered... as you NTSC guys end up with marginally less data in SD than those of us in PAL lands - would downconverting 1080i 60 NTSC to PAL 720x576 25p for DVD production provide a marginally better DVD quality?

[Fred Foronda]

Originally Posted by Douglas Spotted Eagle

Were you using any editor other than Vegas, I'd say convert at the camera.
But Vegas does a much better conversion than does the camera...so let Vegas do the conversion if you can manage it.

Hmm thats a hard choice. For one, I like the scene detection if I convert from the fx1/z1.