Iris tries too hard

I’m using TVAI to upscale 4:3 DVCAM to 16:9 HD. Iris does a great job, but with one flaw. All of my footage features airliners and they have titles on their fuselages and registration numbers. If the aircraft is sufficiently far distant that the lettering concerned is not clear enough, Iris makes it up. That ‘scribble’ is then visible in the upscaled video.

I’ve tried 0% and 100% recover detail, but with no success in curing this habit. I should add that I’m not a video expert, being self-taught. Does anyone have a solution to Iris doing this? Or, what’s the view on the best alternative to Iris for this kind of process?

Are you needing to deinterlace the video? Can try deinterlacing with Proteus or Dione models and then upscale with a different model like Gaia, Artemis or Theia in a second pass.

Hi Kyle

Many thanks for your reply. Yes, I do need to de-interlace. When converting from 4:3 to 16:9, I upscale to 1920 x 1440 as that then allows me to centre the subject within a 1920 x 1080 edit. My feeling is that Proteus has the edge over Dione. I’ll use that just to de-interlace and then upscale to HD @ 1920 x 1440 with one of the others you mention. Does that make sense, or is there a better approach?

Best regards

Howard

Try that and let me know how the outcome looks and then we can make some changes based on what you do not like about the preview or export. Start with Gaia and then we can look at some of the other options. There are some good tips in this forum post by other users for working with older video formats for upscaling and getting improved results. Not all of them will apply but they can be good starting points.

Iris and Proteus (even in auto mode) do antialias, sharpen and disolve large pixel block. It’s a more invasiv method than Gaia model. This brings you some amazing results but I also noticed this doesn’t work if there aren’t enough pixels for objects you have on wide view. It also sharpens details that are far away that should not be sharpened then looks bad. Would be ideal scene based different models are used, but I think that will be difficult to implement.

Apologies for the delaying in replying, but I’ve been on holiday. (I’ll get back to Kyle soon also.) Just to check, when you say This brings you some amazing results… are you refering to Gaia or to Iris/Proteus?

Many thanks.

Howard

Sorry I’ve been so long in taking this forward, but I wanted to explore this as far as possible and include the results in some of the programmes I’m working on, before reporting back.

I enlisted the help of David Clarke of dvctraining.co.uk and we have both used various presets and variations within those presets, to try to preserve the amazing upscaling available in Topaz without creating the artifacts generated by indistinct text in the original video.

We have not succeeded, but there are some work-rounds that will save affected footage, albeit with some minor reduction in overall quality.

Firstly, we found that de-interlacing outside of Topaz enabled greater improvements in the subsequent processing of defective footage. We used Staxrip, which has the advantage of shutting down your system when it’s finished (Topaz please note).

Iris produces marginally sharper images than Proteus, in my view, but can create ‘plasticky’ effects on large areas of block colour. Nevertheless, I would use Iris auto with 40% detail recovery as my first pass over 4:3 SD being converted to 16:9 HD. Different devices and TVs reveal any defects to different degrees. I didn’t think this issue was a major problem until watching what I thought was good, processed, footage on a home cinema/giant TV. Any minor text defects look awful on large screens.

However, using Iris manual, zeroing every parameter and then gradually increasing them, I found that with everything on zero except for Improve detail at 5, Sharpen at 9 and Recover detail at 100, most problem shots can have their faulty text changed from clear scribble to a slightly less clear image that does display the words in question. Inserted into a programme that is mainly composed of the excellent HD video created by Topaz, the diminution in overall quality is not going to be noticed by the ordinary viewer.

If that Iris Text Fix preset still doesn’t solve the problem (and it won’t with very small text), then using Proteus with the same parameters will do so, but the overall quality will be slightly less.

As I’ve said before, I’m not an expert (but David Clarke is) and so this issue may be down to my incompetence. If that’s not the case, then Topaz needs to work on this. I have a particular problem because my footage is all about airliners that all have text on them, but car number plates, adverting hoardings in the background, all sorts of things are going to cause problems for others too.

BTW is there a way of getting the Topaz clip suffix to be the same as the preset used, so it would automatically say Iris Text Fix instead of ‘Iris2’?

All reconstruction models (Iris, Proteus and Rhea) have this problem on at a further distance. Pixel blocks are heavily modified. This causes “unrest” when objects moves and of course all sorts of things are “recreated” such as lettering

You can minimize the effect using “Sharpen” and “Anti-Alias” only and avoid “Deblur”, “Improve Details” and “Dehalo”(<-which is the worsed). Rhea has this problem too, but much less than Iris and Proteus.

Gaia, Artemis and Teiha work differently and this problem does not exist

Many thanks for your helpful response. I’m using Rhea at the moment and agree it’s better, as it doesn’t try to interpret text with swirls, etc. I will definitely try the other presets you mention.

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.