[Dontan]

Does anyone have opinions on these two scalers. Are they similar in quality? I am now contemplating getting the Optoma HD81 which comes with the Gennum scaler but originally was looking at other projectors and was going to get a Silicon Optix Realta based scaler. I know that the Realta has gotten a lot of positive buzz but I don't know much about Gennum products. Any help on this is appreciated.

 

[mark haflich]

They are both good (the chips) but like anything else each has strengths and weaknesses and depending on what's important to you, one could validly prefer one to the other. Some here like to babble and say, both "properly" deinterlace. Of course, properly doesn't mean much, especially with respect to deinterlacing video versus film. video deinterlacer design involves the chip alg writers making design choices and there is no such thing as the proper choice. Neither chip is the ultimate in scaling, both exhibiting some ringing. From reading your question and the generality of your question, I would think that you (and most viewers) would be quite pleased with either chip. Its only when you start A/B ing and looking for faults that most difference are discerned. If you just put it on and watch the content of what you are watching, both should make one quite happy.

 

[oferlaor]

Both are good and both have tadeoffs, I did a shootout that compared the HD81 with Gennum to the HQV Reon equipped mitsubishi HC5000.


Processing was excellent in both cases, I really didn't see much of a difference. There were some things that Gennum did better and some that HQV did better.


I tend to like the ABT deinterlacing most, these days, the VP50 does an excellent job.

 

[TomHuffman]

We did a head-to-head comparison of scalers equipped with both chips. It seemed like the deinterlacing of the Gennum was perhaps a little better. However, the biggest difference was that the noise reduction on the Realta unit was noticeably better. The Gennum unit had a hard time reducing noise without compromising detail.

 

[Steve Goff]

The new Gennum chip (the GF9450) should do noise reduction better than the older models.

 

[mark haflich]

I agree with Ofer. But I do not like the ABT scaler. It isn't bad, but the Lumagen is noticeably better. Since I watch a lot of sports HD at 720p upscaled to 1080p, I use the Lumagen a lot.

 

[Pete]

Apparently the folks at Lumagen decided to switch from Silicon Optics to Gennum in their new Radiance XD processor. The reason they cited was that Gennum's new VXP chip does temporal noise reduction in both standard definition and HD, whereas the SO chip only does it in standard definition.

 

[mark haflich]

According to Lumagen, the Radiance is being designed with a replaceable deinterlacer chip daughterboard. The idea is to not have the machine wed to any one specific deinterlacer/processor chip. Lumagen's press release states that the initial board produced would be for the new Gennum chip. The Realta could come later.


We can only guess at reasons for doing the Gennum first. Only Lumagen knows for sure.

Possible explanations might include the one previously mentioned, the fact that the Gennum chip is less costly than the Realta chip, the ease of getting the thing to work correctly, the support available from the chip manufacturer, the quality of the various deeinterlacing algs etc etc. My guess is that the Realta does somethings better than the Gennum and vice versa. Lumagen does much or what these chips can do better than the chips, such as scaling. One of the nice things about the Radiance is how much spare gated chip capacity will be built into the mother board allowing Lumagen to develop and use algs for such things as NR better than what the chips can do. If and until it does, Lumagen will be able to use whatever is in each chip. Sweet.

 

[oyfoo]



Quote:
Originally Posted by Pete /forum/post/0


Apparently the folks at Lumagen decided to switch from Silicon Optics to Gennum in their new Radiance XD processor. The reason they cited was that Gennum's new VXP chip does temporal noise reduction in both standard definition and HD, whereas the SO chip only does it in standard definition.

The low cost Silicon Optix Reon VX chip (only available since mid 2006) will only do TNR at SD res. The higher cost Realta (available since fall 2004) does TNR from SD to HD.


Here are the products currently announced using these chips AFAIK.


The projectors that use Realta are 3M (XGA 1xDLP), Toshiba (480p 1xDLP ), Yamaha (720p 1xDLP), NEC (2kx1k 3xDLP), Cinetron (1080p 3xLCOS), Digital Projection (1080p 1xDLP and 3xDLP with ext. VP), Epson (720p and 1080p 3xLCD with ext. VP), BenQ (both 720p and 1080p). I think there's only one projector announced with Reon VX , the Mitsubishi (1080p 3xLCD).


VPs with Realta come from JVC, NEC, Digital Projection, Calibre, Epson, maybe some others. None with Reon as far as I know.


LCD TVs with Realta come from Syntax and Proview.


DVD players from Denon (SD with both Realta and Reon versions), Toshiba (HD version with Reon), Samsung (Blu-Ray version also with Reon).


Nothing announced yet with the new Silicon Optix Geo chip, targeted for projection systems as well as LCD non-uniformity compensation.

 

[Steve Siener]

oyfoo,


Good info. Thanks for clarifying.


If Samsung gets it right in their new BD player it will be my next video purchase. Regarding the practice of reinterlacing 1080p24 disc content in some hi def players, I mused in the Samsung BD-P1200 thread: I wonder if some of these hi def players reinterlace so they can pass that data to a chip (like the Reon) and have it do 3:2 pulldown of 1080i24 source, meaning it's simpler for these players to reinterlace 1080p24 than it is for them to take 1080p24 and convert it directly to 1080p60. But I would hope that any player supporting 1080p24 output would pass it through unprocessed instead of taking the reinterlacing path.


The Geo seems promising for anyone who has a misconverged 3-panel display/projector.