Would there ever be a reason to capture "upconverted" video?

[yoda313]

So I was thinking would there ever be a reason to capture "upconverted" video?

I was thinking about this when watching a tv show dvd on my ps3 last night. Since it was connected via hdmi (source is widescreen regular def dvd fyi) it was being upconverted like usual. But I was wondering would there be any savings in processing power if the video was to be captured in its upconverted state?

Say I was to capture the video output via component cables from the ps3 (or xbox 360 - I have both and know both upconvert nicely) with my hauppauge hd pvr. Now I would be able to capture the signal in 1080i if I have the output from either console set to that level.

I guess my question is would there be any technical benefit to capturing the upscaled video?

I do understand that either the player or the tv itself can do upscaling. But if the initial source that is played is at this "upscaled" level would it be reupconverted again with say a wdtv media player or playing on a pc? Or even burned to a avchd and played again on the ps3 itself?

I am wondering if the video that was originally 720x480 and upscaled to 1080i would be reupscaled again upon playback. Also is there any processing power savings by playing the newly upconverted conversion? Lets say you have a weak pc that can't do high def video very well. Would it be easier on the pc to take a upconverted m2ts or mp4 file that is already 1080i and play that versus upconverting an original source 720x480? I do understand video card power is a factor in this equation. I guess I'm asking for some generalities, is the original video being upconverted take more power than to play the same video in a preupscaled file?

I have no need to use a captured upscaled video file since I have the original video in question. This is strictly theoretical stuff I'm wondering about.

[lordsmurf]

My opinion is "no".

[yoda313]

My opinion is "no".

Thanks.

Would you care to elaborate?

Would it be because the upscaling is done by the tv or the player anyway and that it doesn't matter what the source is as long as it is relatively "clean"?

[lordsmurf]

This is an old discussion for a new format. Some years ago, it was upscaling 16mm onto new 35mm masters.

You're a Robotech fan -- look for Carl Macek's discussion of Super Dimensional Cavalry Southern Cross (Robotech Masters saga) and how they had to upscale 16mm to 35mm. Many discussions ensued over that one situation alone. It's either in the old art books, or maybe some Robocon bootleg tapes. I think they even re-discussed it in some of the various extra content for Shadow Chronicles, either on-disc or in written pieces here and there.

In many cases, the upscaled version looks worse or even artificial compared to the original. Sometimes it adds noise. There's all kinds of issues.

In the case of digital video, upscalers improve with every generation, so don't lock yourself into what you have now. Stick to the original content, at the original size, and let technological "nature" take its course.

[jagabo]

Upscaling is almost always performed by the graphics card (this is one of the main points of using video overlay) and even the cheapest graphics card's these days have upscaling ability (although the quality varies). On a weak computer playing back an HD video will be more taxing than playing an SD video and letting the graphics chip upscale.

There can be reasons to capture upscaled SD. For example, analyzing the quality of different upscaling devices. Or, when a really good upscaler is used it may be better than your computer can do with the SD source.

[TBoneit]

I Zoom (Upscale) from my DVR to capture Letter-boxed SD content as 16:9 with no conversion needed.
It works for me and seems to me to be a reason to do so.

[yoda313]

@lordsmurf - thanks for the analog analogy. That does seem to make sense. Since the original is "smaller" than the larger format I can see how it can be inflated with errors or just overall softness compared to the original source.

Thanks to jagabo and tboneit as well.

One other question:

Is there technically any difference between the terms UPSCALE and UPCONVERT or are they essentially the same thing?

[edDV]

A reason to do this is to integrate SD material into an HD edit. Hardware "upconverters" differ in quality but they do the job 1x realtime. Quality software upconversion takes hours to days.

If you have s serious project with quality DV source, you can run the needed clips through a high end upconverter at a post house without spending much money. You can preview edit with placeholders so that only the needed clips are converted.

[edDV]

Is there technically any difference between the terms UPSCALE and UPCONVERT or are they essentially the same thing?

Similar. Upscale implies decompress to frames and rescale of each frame (or field pairs). Upconvert may imply additional processing such as noise reduction, deblocking or motion based processing.

[lordsmurf]

A reason to do this is to integrate SD material into an HD edit. Hardware "upconverters" differ in quality but they do the job 1x realtime. Quality software upconversion takes hours to days.

This was the issue with the Robotech 16mm>35mm conversion.

I often see upconvert and upscale used interchangeably -- beware.

[sphinx99]

Quality of upscaling and processing is a moving target and is improving over time. I watch SD material routed through a Onkyo 5007 receiver with a HQV Reon video processor. Its prowess with deinterlacing, upscaling, noise reduction, sharpening and color-balancing puts to shame every Virtualdub and AVISynth filter or script I've ever seen, used or written, and it does it in real time. As long as I get it the raw video in a format that the receiver recognizes as in need of processing, it's going to do a better job than I could with all this mucking about with software processing, and its replacement in my home five years out will do an even better job.

Not everyone is invested in a top-notch hardware processor yet, but ultimately these are chips that will eventually land in every television, projector and receiver out there. Having been immersed in outboard "home theater" processors of various kinds for some years now, I'm convinced that it does not make sense to process video too much, except for transportability or viewability. If you capture upconverted footage, you're benefiting from the algorithm quality of your current scaler, but missing out on the quality of your future scaler. I don't think that's a trade-off worth making, so I would suggest capturing video in the most raw, untouched format possible (if that's 480i or 480p, fine) and do scaling on playback with the best video processor you can currently afford. When the means opens up to pick up a better one, whether standalone or as part of a better receiver or display device, then go for it and your entire video library will benefit at once.

[yoda313]

Thanks everyone for all the insight. As I mentioned this was all out of curiosity and not for any project that I was planning on.