work flow question with vegas 9e, virtual dub deshaker and cineform

W

william boehm

Member
i am new at Vegas 9, and finally successfully installed deshaker and virtual dub which recognize my cineform files. my question is retaining the best image quality of my wildlife footage, all shot in 60i.

1. do i start with virtual dub and deshake some of the grab footage where a tripod couldnt be set up in time (whales from a boat, sea otters from a kayak, birds while hiking etc), then edit in vegas, or

2. edit in vegas then deshake footage that needs to be stabilized.

3. do those of you experienced deinterlace cineform to 30p and edit, or keep it in 60i for color correction, brightness contrast, very little slo mo.

thanks in advance to those experience who are willing to share information to newbies. bill
 
william boehm said:
i am new at Vegas 9, and finally successfully installed deshaker and virtual dub which recognize my cineform files. my question is retaining the best image quality of my wildlife footage, all shot in 60i.

1. do i start with virtual dub and deshake some of the grab footage where a tripod couldnt be set up in time (whales from a boat, sea otters from a kayak, birds while hiking etc), then edit in vegas, or

2. edit in vegas then deshake footage that needs to be stabilized.

3. do those of you experienced deinterlace cineform to 30p and edit, or keep it in 60i for color correction, brightness contrast, very little slo mo.

thanks in advance to those experience who are willing to share information to newbies. bill
Click to expand...


If you have lots of rough stuff to cut out, you may want to pre-edit so you don't have to waste time "deshaking" it all (it can be very slow)

Uncompressed would be the "best image quality" for everything. Cineform is lossy even on the highest setting; it depends on how much you are willing to tolerate. It's good enough for most people.

Deinterlacing will degrade image quality, you would only deinterlace if you are exporting to web or progressive displays

What kind of 60i footage? AVCHD? HDV? DV-AVI? vdub doesn't handle interlaced MPEG2 correctly (chroma gets associated with the wrong field) , so you should use avisynth scripts for input instead of natively importing

Whether you deinterlace or not depends on your export goals (is this for blu-ray, dvd , web etc....)
 
thankyou for such a quick response. i shot hdv with a canon xha1, and this year hust started shooting a canon xfs100 vixia in avchd. both are transferred into cineform through neoscene 4, which i just upgraded to neoscene 5.04 which now recognizes my cfhd avi files.

so only some of my footage is shot from a boat or kayak, or sometimes a grab shot of a bear or bird in the backcountry. i tripod wherever i can.

it seems to make sense to rough edit first in vegus, cut clips to lengths and footage. but the next step is the question. do i deshake first then edit in vegus, or edit in vegas, put it into a cineform file again, deshake what i need? just trying to minimize as best i can IQ loss. thanks again in advance. bill
 
do i deshake first then edit in vegus, or edit in vegas, put it into a cineform file again, deshake what i need?
Click to expand...

If you have lots to cut out, I would cut out the "usless parts" first. Deshaking can be slow, and there is no need to deshake footage that you will cut out later anyway

You can edit in vdub as well (simple trims) , so there is no need to do the rough cut part in vegas. Moreover, the more times you import/export there are compression losses as well as colorspace losses (RGB=>YUY2). So you might use HDLink to convert your clips to cineform, then do rough cuts in vdub and deshake => export out stablilized cineform intermedate => import into vegas

Cineform is good for rendering multiple generations without much loss , but uncompressed will be better. Only if you "pixel peep" will you be able to see the difference
 
well i am very new..becuase i wasnt sure what you meant about using HD link? so each time i simply cut, not edit, but cut frames out i am changing the IQ? Again thanks for educating me. bill
 
well i am very new..becuase i wasnt sure what you meant about using HD link? so each time i simply cut, not edit, but cut frames out i am changing the IQ?
Click to expand...

HDLink is the conversion utility that comes with cineform (or maybe it doesn't come with the neoscene version?) How are you converting to cineform?

It's tough to give a simple straight answer:

If you cut in the uncompressed domain on a I-frame sequence, you're are not losing any quality. If you convert to cineform first, it's all I-frame and you can make cuts anywhere, but you still have losses upon export (both compression lossess and colorspace losses, but again these are minimal and neglible for most people)

Your native HDV and AVCHD videos are Long GOP, so whenever you make a cut that is not preceding an I-frame, you need to re-encode at least the entire GOP. If you use uncompressed for everything, there is no quality loss except for the colorspace conversion and using uncompressed would give you the best quality

You can cut and not re-encode with vdub hence NO quality loss from that cineform generation if you use "direct stream copy mode". BUT whenever you use an internal filter in vdub (like deshaker), it requires re-encoding, so you will incur a generation loss (There is also an RGB conversion)

Vegas can smart render HDV and sometimes AVCHD (it's very buggy in my experience) - Smart rendering is allowing pass though of untouched segments, and re-encoding only around the I-frame cutsite (things like effects will require the entire section be re-encoded and quality loss incurred). So one can make an argument to do the cuts first in vegas if you can get smart rendering to work properly

That might not be helping, but ask if it didn't make sense
 
Last edited:
i can grasp uncompressed vs compressed..but not necessarily the hd link concept that is in cineform.

i transferred all hdv tapes in the canon to a hard drive via the canon xha1 and imported through cineform..the highest quality. i dont remember the process since i simply borrowed a camera (sold mine last year) and transferred about 70 one hour tapes. so everything is in a cineform avi file. i have played with vegas, snapping ends and portions off that are then auto saved by default into an .veg file. i still of course retained the sample cineform avi' transferred in neoscene 4...again done in february of this year.

i am still a little foggy about your concept. i understand now that even just cutting the cineform avi files in vegas compresses and decompresses the cineform files. so. to minimize IQ loss..do i stay in cineform and cut the 60i footage..then save again and no loss? or are you suggesting cut in virtual dub in another codec..which your suggesting is lossless..then cut? i probably need to be led by the hand through this. thanks again for taking the time to help.b ill
 
i understand cuttng in 'direct streamm copy mode in virtual dub, butdoes vdub convert to lossless from the cineform files, then reconverts to lossless after the cuts and saving the shortened clip?
 
william boehm said:
i understand cuttng in 'direct streamm copy mode in virtual dub, butdoes vdub convert to lossless from the cineform files, then reconverts to lossless after the cuts and saving the shortened clip?
Click to expand...

Don't get hung up on HDLink, it's just a conversion tool.

OK, so your initial format is cineform (i.e everything has already been converted to cineform avi?) ?

Then your best course of action is to use vdub in direct stream copy mode for cutting. There is no quality loss

Vegas won't stream copy cineform, it will decompress to RGB, then you will need to re-compress with whatever export format. Even if you use uncompressed you will have incurred a YUY2=>RGB=>YUY2 conversion loss . Conversely, in vdub, without using filters, you can stay in the same colorspace.

BUT, if you use deshaker (or almost any other internal vdub filter, some work in YUV colorspace), vdub does an RGB conversion anyway (most vdub filters work in RGB colorspace), so if you export cineform from that , you will incur both RGB=>YUY2 colorspace loss and cineform 2nd generation compression losses. If you export RGB uncompressed (because you have RGB now from using a vdub filter) , you will not incur additional compression lossess other than the colorspace conversion back to YUY2 (which is cineform 4:2:2 ; now there are cineform 4:4:4 RGB varieties, but I suspect you aren't using them, they are more expensive)

Vegas will work in RGB mode when you import either cineform or RGB uncompressed AVI

SUMMARY: Since you already have cineform AVI as your starting point , you can do you rough edits in vdub and use deshaker. Even on clips that you don't want to use a filter or deshaker on on you can cut in "direct stream copy mode" , and lose no quality (from that cineform generation) by using vdub . By direct stream copy mode, this means "cuts only" editing i.e. linear editing and no overlays or multiple tracks like vegas

If you don't have cineform registered, you cannot export cineform (cineform has a free DEcoder, but you need to pay for the ENcoder) . In this case you could use uncompressed RGB from vdub or a lossless compression format (e.g. lagarith, huffyuv). If you use direct stream copy mode, this way stay in YUY2 colorspace , so the uncompressed format would be YUY2 (like the input cineform file), not RGB. If you use an internal vdub filter or deshaker, it will be RGB

Hope that helps :)
 
Last edited:
i think i have it...and thanks again for the explanation and time. i will then cut in virtual dub, but then cuts only clips will again be saved in the same cineform avi file by using the streaming copy mode. Right? If not then it is put into a lossless code that takes more GB space? what code..sorry but it needs to be laid out for me step by step

then those clips to be deshaken in a filter, using cineform avi will incur some loss because it becomes uncompressed the recompressed into an RGB color code before i final edit in vegas?
 
william boehm said:
i think i have it...and thanks again for the explanation and time. i will then cut in virtual dub, but then cuts only clips will again be saved in the same cineform avi file by using the streaming copy mode. Right?
Click to expand...

Correct. Video=>direct stream copy , Audio=>direct stream copy . Technically it's not "cutting", it's really stream copying clips into another cineform AVI (your original Cineform file is unaltered). No quality loss when done this way because cineform is all I-frame (Everything is a keyframe). It's basically passing though identical video with cuts

If not then it is put into a lossless code that takes more GB space? what code..sorry but it needs to be laid out for me step by step
Click to expand...

If you are using a filter like deshaker, video will be set to full processing mode instead of direct stream copy mode. This means vdub is operating in RGB colorspace. All you have to do is set video=>compresssion and select either no compression (top of the list) , or if you want you can install a losselss codec like lagarith or huffyuv and it should show up on the compression list.


then those clips to be deshaken in a filter, using cineform avi will incur some loss because it becomes uncompressed the recompressed into an RGB color code before i final edit in vegas?
Click to expand...

Yes. Because your input Cineform files are YUV colorspace (YUY2 4:2:2) , and filtering in vdub will operate in RGB colorspace. Colorspace conversions cause slight quality loss (there are rounding erros) . They are not visible to normal folk who don't pixel peep, but they accumulate after multiple generations, so avoid colorspace conversion when you can.

If you export uncompressed RGB, instead of cineform 4:2:2 , you will incur less loss (because 1 less colorspace conversion, and no compression loss)
 
Last edited:
again thankyou for your time translating. one final question to you is what would you export the deshaken virtual dub filter codec as to use in the next step of editing in Vegas ..? bill
 
I did receive a reply from david newmann with cineform just a few minutes ago...quote "Vegas supports smart render, lossless move of compressed sample from input to output file. VirtualDub can do the same. The multi-generation quality is so good, you need not worry." PDR ...whadya think of this?
 
Vegas does support smart render, but only of DV-AVI, HDV, (and sometimes AVCHD, it's quite buggy).

I can't comment on vegas smart rendering cineform - I suspect it cannot and there will be at least a YUY2=>RGB=>YUY2 conversion.

I don't use cineform in my workflows anymore so I cannot veryify. I can tell you with 100% certainty that it did not smart render cineform a few years back. You can do these tests yourself. Vegas well tell you if it's smart rendering or not

Note "smart rendering" isn't necessarily the same thing as "direct stream copying". Stream copying implies cuts only editing and passing though of segments. Smart Rendering might include other things like effects etc... The smart rendering part means re-encoding only those segments which have be altered (e.g. a crossfade) , other sections are passed through untouched

The multi generation quality is good, but you can definitely tell the loss even after 1 generation. People who do FX work can easily tell, they zoom in and "pixel peep". Ordinary viewers probably cannot tell the difference
 
thanks pdr..you have really helped clarify a complex issue for me. i hope others new to video editing will peruse this information. bill
 
I did a quick search, there were many posts indicating that vegas does indeed smart render cineform. If this is true with recent builds, then you could use vegas to do the rough cuts without loss

When you export out vegas it should be saying something like "no recompression required" along those lines. Let us know if it works out
 
You must log in or register to reply here.
Back
Top