Pages: [1]   Go Down

Author Topic: Rendering twice.  (Read 9598 times)

stamper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5882
Rendering twice.
« on: June 25, 2013, 07:17:35 am »

I am a newbie to the movie making genre. My aim is to use PS CS6 for enhancing the colour - if necessary - in a movie file. One reason is that I know about using PS from shooting stills but it can't do everything. I have Sony Vegas which is probably superior but I prefer CS6. My question is can I use Sony Vegas to try and lessen the shake that is my files taken hand held and render the movie and then use PS CS6 to enhance colour, contrast etc without degrading the quality of the already rendered movie? In a nutshell it isn't like saving a jpeg twice which causes loss of quality?

Christopher Sanderson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2693
    • photopxl.com
Re: Rendering twice.
« Reply #1 on: June 25, 2013, 08:37:51 am »

...My question is can I use Sony Vegas to try and lessen the shake that is my files taken hand held and render the movie and then use PS CS6 to enhance colour, contrast etc without degrading the quality of the already rendered movie? In a nutshell it isn't like saving a jpeg twice which causes loss of quality?

As long as the initial render is not to a lossy format you will be fine.

IOW, use an intermediate high bitrate format (4;2;2 or 4;4;4) for the render out of Vegas.

Only use a lossy (H264?) format for the final finished render.

fredjeang2

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1376
Re: Rendering twice.
« Reply #2 on: June 25, 2013, 09:59:38 am »

To use a still imagery analogies

The Consumer cameras (like the Nikon dslr etc...) generate what would be equivalent of a bad jpeg (4 or 5 quality)
That's what you got.

You understand that you're not going to use a bad jepeg for further manipulations otherwise you'd immediatly degrade.

What Chris is saying, would be the equivalent to create a robust tiff from your bad jpeg. The Tiff created does not add
any quality. You can't invent what's not been recorded.
But what you can do is to stop the bleeding using a version robust enough to handle manipulations.

The bad jpeg is the: compressed codec for adquisition like AVCHD or for delivery like H264

the tiff would be for example a PRORES 444

Then you got what Chris calls "intermediate" wich is a jargon that means codecs that have been specially
studdied for post-production tasks. PRORES, DNxHD, Canopus HQ

There are generally robust enough to handle families ( copies from copies ) and generaly come in different flavours
according to your needs.

Some intermediates are at the same time adquisition and mastering, it's the case of PRORES or DNxHD for ex.
It means that you can use then to shoot, to edit, to color correct, to master
there are a bit like a 4x4 car.

So, for the final render, if it's for the web, you're going to want a lossy format because you need small files-good enough qualy
thus: H264 for ex

BUT you're not going to manipulate your images in H264 or AVCHD. It would be suicide.
So you need to use a high bitrate codec when any manipulation is necesary
« Last Edit: June 25, 2013, 10:10:43 am by fredjeang2 »
Logged

fredjeang2

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1376
Re: Rendering twice.
« Reply #3 on: June 25, 2013, 10:23:10 am »

So to resume simplifying

for the adquisition of your images = adquisition codecs :
- consummer cameras: AVCHD...
- profesional: Prores, Infinity, DnXhd, RAW video etc...

for the post-production = intermediate codecs :
- DnxHD, Prores, Canopus HQ etc... they come in different versions according to your needs
(for example if you need alfa chanel, if it's for editing only etc...)
- DPX etc...

for special effects = image sequences:
- like Open EXR, Tiff etc...

for mastering = looking for high-quality where you can make copies from in different flavours
- like uncompressed QT, PRORES higher versions, the list is long

For archiving you can use JPEG2000 for ex

Then you got the wrappers. Wrappers are the formats (file) that contains the codec. QuickTime is a wrapper or container.
So you can have a QT of H.264 or a QT of jpeg2000, a QT of his mother in law etc...

QT is a consumer and prety bad wrapper. MXF is a standart. But practicaly, QT is vastly used at any levels.
So the wrappers are your files (or file format)

It's resumed and simplified on purpose as you just come into video but you got the idea.
« Last Edit: June 25, 2013, 10:34:03 am by fredjeang2 »
Logged

Christopher Sanderson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2693
    • photopxl.com
Re: Rendering twice.
« Reply #4 on: June 25, 2013, 11:06:03 am »

...It's resumed and simplified on purpose as you just come into video but you got the idea.
;D

fredjeang2

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1376
Re: Rendering twice.
« Reply #5 on: June 25, 2013, 11:34:28 am »

;D
Gosh! I forgot the delivery ones !...
Logged

RFPhotography

  • Guest
Re: Rendering twice.
« Reply #6 on: June 26, 2013, 10:59:48 am »

Yes and no.

Yes, in that you can save out the file in a lossless intermediate format.  The Lagarith Lossless codec will do that.

No, in that there is no anti-shake filter in Vegas.  People have been wanting one for years but there isn't one.  Not even a third party plugin, as far as I'm aware.

There is an anti-shake plugin for After Effects.  There is also Virtualdub which has an anti-shake plugin available.  I haven't used Vdub in a while so am working a bit from memory but I believe it will allow you to use the Lagarith Lossless codec to output an AVI file (AVI is the only output format in Vdub).  Vdub is free, open source and takes a bit of getting used to.  Anti-shake filters aren't a panacea so don't expect miracles.

The problem you may have is that I don't know if PS will be able to recognise the Lagarith codec.
Logged

stamper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5882
Re: Rendering twice.
« Reply #7 on: June 26, 2013, 11:26:13 am »

Anti shake is probably the wrong term. I have the Sony Vegas Movie Studio HD Platinum 10.0 and it has a stabilise function that has different strengths depending on how bad you consider the shake to be. I tried it with a "mild" hand holding shake and it worked very well. However I think there is a loss of resolution and possibly cropping. As to a codec I am still working out which is "best". I am trying different clips and saving them to see the what kind of quality they save at. Still some learning to be done. :)

RFPhotography

  • Guest
Re: Rendering twice.
« Reply #8 on: June 26, 2013, 01:43:50 pm »

Holy crap!  They finally did it?  Wow.  Users had wanted it for years and it was never included.  I haven't upgraded since v9.  Sorry about that.
Logged

Sareesh Sudhakaran

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 546
    • The Indie Farm
Re: Rendering twice.
« Reply #9 on: June 28, 2013, 03:50:13 am »

However I think there is a loss of resolution and possibly cropping. As to a codec I am still working out which is "best". I am trying different clips and saving them to see the what kind of quality they save at. Still some learning to be done. :)

Cropping is inevitable with stabilization and tracking. (Look up match-moving and camera tracking).

As far as loss of resolution is concerned, if the initial video had motion blur (which it will), it will lose resolution - it's hard to say whether that is a perception thing or an actual 'thing' without studying the footage. Consumer-grade algorithms try to 'compensate' using available camera data in the file, and this isn't always a good thing.

Add some sharpening if you must.

As far as intermediary codecs are concerned, I don't recommend them at all. You could perfectly stabilize in After Effects with native footage, edited natively in Premiere Pro. Test, and see.
Logged
Get the Free Comprehensive Guide to Rigging ANY Camera - one guide to rig them all - DSLRs to the Arri Alexa.

fredjeang2

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1376
Re: Rendering twice.
« Reply #10 on: June 28, 2013, 07:14:00 am »

Cropping is inevitable with stabilization and tracking. (Look up match-moving and camera tracking).

As far as loss of resolution is concerned, if the initial video had motion blur (which it will), it will lose resolution - it's hard to say whether that is a perception thing or an actual 'thing' without studying the footage. Consumer-grade algorithms try to 'compensate' using available camera data in the file, and this isn't always a good thing.

Add some sharpening if you must.

As far as intermediary codecs are concerned, I don't recommend them at all. You could perfectly stabilize in After Effects with native footage, edited natively in Premiere Pro. Test, and see.

Yes, cropping is inevitable with stab-tracking.

As for intermediate codecs, it depends. They can be extremely usefull or extremely useless. It just depends.

The profesional editors here, and specially the onces involved in cinema,
never edit with the native material. It's not a question of power, those guys have all the computer power available, but stability
There is no magical ultimate workflow,
However, the practise is to cut with a low-bitrate intermediate and relink or conform in the end.

but also, there are intermediate codecs with a high-bitrate that are suitable for editing and manipulations and don't stress the software. Specially when
the original material is highly compressed like AVCHD. No serious editor cuts in AVCHD, unless it's small volume, fast delivery.
A good old DNxHD and you're good to go from the beginning to the end.
And that's why the broadcast are also so crazy about the minimal specs, exchange material, metadatas etc... because they don't have time the cinema editors have and they don't want hassles.

But using native material for image manipulation, and specialy for heavy color correction,  if the native material is not robust
it will only lead to problems at one point or another of the pipeline. Long volume editing, the same.

That's something Arri understands and why the Alexa in Prores configuration is so popular, because the adquisition and editing being the same, and robust, the workflow is
straightforward. You shoot already in a format that is suitable for post-production and can handle severe CC. No surprise.
And that has a cost, the price of the Alexa. Same, using intermediate in post has a cost: time, but most of the time it's worth.

Manufacturers have developped those for good reasons. If you're in Avid, you better cut in Avid Native media (not native support). If you're in Grass Valley, the same.
It's not a golden rule but it avoids lots of hassles.
Most editors have only a word in mind: stability. No fancy fast turnarrounds. No risk. Avid? dnxhd. Apple? prores. Grass Valley? Canopus HQ  etc...

Also, if you cumulate layers and complex fx within an editor in for ex AVCHD, the software will be prone to freeze easily. With the use of an intermediate it does not.
Then it's only a question of conforming to the higher-disponible media.
Now, if people wants to do special fx with low bitrate 4:2:0 AVCHDs or so...well, as we say: swim at your own risk.

So the use of intermediate is a question to be answered by everybody's priorities according to several parameters, but certainly not accesory.
« Last Edit: June 28, 2013, 08:22:22 am by fredjeang2 »
Logged

RFPhotography

  • Guest
Re: Rendering twice.
« Reply #11 on: June 28, 2013, 08:36:42 am »

Cropping is inevitable with stabilization and tracking. (Look up match-moving and camera tracking).

As far as loss of resolution is concerned, if the initial video had motion blur (which it will), it will lose resolution - it's hard to say whether that is a perception thing or an actual 'thing' without studying the footage. Consumer-grade algorithms try to 'compensate' using available camera data in the file, and this isn't always a good thing.

Add some sharpening if you must.

As far as intermediary codecs are concerned, I don't recommend them at all. You could perfectly stabilize in After Effects with native footage, edited natively in Premiere Pro. Test, and see.

Sareesh, so you're suggesting simply edit in the h.264 format that comes out of the camera? 
Logged

stamper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5882
Re: Rendering twice.
« Reply #12 on: June 28, 2013, 08:43:46 am »

Bob. I downloaded the Lagarith codec but in Sony Vegas it didn't show up. Two different sources I read stated that for intermediate rendering Quicktime was recommended but the file size was huge compared with the original size when I tried it. However I don't think it is something that will be stored in your computer long term?

Sareesh Sudhakaran

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 546
    • The Indie Farm
Re: Rendering twice.
« Reply #13 on: June 28, 2013, 10:48:29 pm »

Sareesh, so you're suggesting simply edit in the h.264 format that comes out of the camera? 

Yes, unequivocally.
Logged
Get the Free Comprehensive Guide to Rigging ANY Camera - one guide to rig them all - DSLRs to the Arri Alexa.

bcooter

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1520
Re: Rendering twice.
« Reply #14 on: June 29, 2013, 12:59:16 pm »

Not to state the obvious, but if you shooting video, don't want shake, practice deliberate movements, keep the camera off your face and better yet use some kind of tripod.

Movement is great, if it works with the story, hand held movement if not annoying can make a story seem more real, but shake, is just shake.

Some people carry a mono pod which to me is not any more or less cumbersome than a lightweight tripod and I find more limiting. 

There are a lot of anti shake solutions in NLE's and as plug ins, but they work in different ways and if your moving in to the subject as well as shaking the shake software gets fooled and parts of the image will slightly bulge.

It's a shame that all cameras don't shoot a minimum of a 4:2:2 10 bit codec but they don't, most shoot some flavor of h264, some better than others.

The Panasonic GH3 is very good, the Sony FS100 is not, somewhere in between falls the 5d2/3, and the Olympus omd and a truck load of other cameras that cost better $5,000, and $500.

Still the real answer to all of this is garbage in, garbage out, good in good out, great in great out.

Drop a proper 14 bit captured still image in a NLE and key frame it and move it around all over the place, then do the same with the same scene shot in a marginal h264 video file and look at the difference.  Unless it is flat lit and not moving, the difference is obvious.

That is the difference between a RED, and Arri and any prosumer camera.

IMO

BC

P.S.   the very best image stabilization I've seen in any camera, still or motion, regardless of costs is the Olympus OMD's in camera stabilization.  You have to really work to make it look bad (in regards to motion).
Logged

Sareesh Sudhakaran

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 546
    • The Indie Farm
Re: Rendering twice.
« Reply #15 on: June 29, 2013, 11:36:28 pm »

It's a shame that all cameras don't shoot a minimum of a 4:2:2 10 bit codec but they don't, most shoot some flavor of h264, some better than others.

I would gladly take 8-bit 4:4:4. It's roughly a 30% increase over the Canon C300's 8-bit 4:2:2 50 Mbps. Definitely doable - they should abolish chroma subsampling from acquisition. The same with 10-bit should be a 50% increase.

It's already being done to some extent with RAW, but would love to see a Prores 444 8-bit/10-bit without having to pay extra.

Shoot RAW+Prores at the same time, and use RAW only when you have to - very excited by the potential of the new BMCC 4K. Tomorrow is July.
Logged
Get the Free Comprehensive Guide to Rigging ANY Camera - one guide to rig them all - DSLRs to the Arri Alexa.
Pages: [1]   Go Up