Algolith Dragonfly Reviews - Page 3  

Not only but specially. The Dragonfly has a good motion detection. Everything you shoot at it should play smooth. Smoother than the competition. But his that enought to buy a Dragonfly over and HD2+? Not sure only your testing could draw a conclusion. Personnaly I think a combo Mosquito + HD2+ would be more beneficial than let say the Dragonfly alone. Specialy for watching SD material.

Bruno

What is the effect of using your satellite box watching SDTV shows, but using the STB HDTV 1080i outputs to go into the Algolith as opposed to the 480i source going into the algolith, using a S-video to component converter, or directly outof SD component outputs if the DSS STB has those? Does the DSS STB internal 480 to 1080 upscaling causes the algolith to not properly recognise the mosquito artifacts? I know using an outboard scaler to do the same 480 to 1080, if one did, needs to be done after the algolith, not before.
Thanks

Also, when the DSS boxes switch soon from MPEG2 to MPEG4, how will that effect Algolith mosquito reduction processes????

Damn good question Mark. This thing could be obeselete before it hits the street. What about HD-DVD?

Mark,
maybe I'm missing something but I don't see why the fact to switch to MPEG4 should change anything to the Mosquito processing.
The MPEG processing is done and the Mosquito receives an uncompressed video signal (in analog or digital) that needs to be cleaned.

Unless you're implying that MPEG4 is producing a better image?

Hi
I actually have no idea if the Algolith needs something based on MPEG2 rather than MPEG4. Just curious if that will make a difference. Supposedly the MPEG4 will give better quality video compared to the MPEG2, but is that just false claims by DSS providers, or true statement based on technology differences?
Thanks
mark

I believe someone posted that Mosquito specifically looks for the 8x8 blocks and, therefore, works best with unscaled input. As far as MPEG4 vs MPEG2, I read that MPEG4 is also block-based. I could be wrong.

Hi Wilson,
So a DSS STB tuned to a SD channel such as CNN, but outputting component slected as 1080i, would not be a good way to deal with SDTV with algolith?

Mark, it's better to pass SDTV into Mosquito unscaled. I wish my HDTivo had a native output mode....

The funny thing is with my pioneer plasma and Simmetry scaler, i prefer the SDTV scaled by the DSS STB as 1080i component into the Simmetry over the S-video of the DSS STB into Simmetry then scaled to the NR of plasma by Simmetry. So I might not like how the SDTV looks based on the nonalgolith set up as just described.

Mark - Have you tried 480i via component vs. S-Video into your scaler?

Hi Josh,
My DSS STB only has S-video 480i, no 480i component outputs.

Any updates on when this puppy is shipping?

Lon

Some time in the Fall, last I heard.

what happen to dragonfly????
no news still now :(
someone have new information for their shipping ,they have been very enthusiasmic at the show at new york and after that nothing!!!!

It's not supposed to be too far off.

Patience in this industry is a must...

I heard that it would not ship until September

Is anyone at all disappointed that aloglith chose not to include the full mosquito dnr functions in the dragonfly? I would have been happy to pay for the one box solution...or are there reasons for not doing so?

Seems to have been computing power problems. The mosquito hardware is only fast enough to do the mosquito work and the dragonfly hardware is only fast enough to do the deinterlacing/scaling work. Of course they could try to put both boards into one big device and kind of connect them internally, but that wouldn't help neither cost nor size. I guess we'll see a combined box somewhen in the future, but not too soon. Let them just get out with the DragonFly first. Then maybe they'll start working on a combined device.

One thing that recently crossed my mind is this: I've read in a Mosquito review somewhere (don't remember where) that the Mosquito worked better on progressive sources than on interlaced sources. On the other hand Algolith sais, the Mosquito doesn't like scaled sources. So that means the optimal order would be:

480i
-> DragonFly deinterlacing -> 480p
-> Mosquito noise reduction -> 480p
-> DragonFly scaling -> native display resolution (progressive)

However, this is only possible by either buying two DragonFly devices or by combining it all into one big device and optimizing the work flow internally. So a combined device could have advantages in PQ, too.

One thing that recently crossed my mind is this: I've read in a Mosquito review somewhere (don't remember where) that the Mosquito worked better on progressive sources than on interlaced sources. On the other hand Algolith sais, the Mosquito doesn't like scaled sources. So that means the optimal order would be:

480i
-> DragonFly deinterlacing -> 480p
-> Mosquito noise reduction -> 480p
-> DragonFly scaling -> native display resolution (progressive)

However, this is only possible by either buying two DragonFly devices or by combining it all into one big device and optimizing the work flow internally. So a combined device could have advantages in PQ, too.[/quote]

If this is the case then it appears that aloglith never intended for both devices to be a package solution.....i guess you'll just have to pick your poison. I have to say that the iscan looks like a better route to go....

A better route to go compared to what? The iscan does not do any noise reduction. So you can't compare DragonFly + Mosquito with the iscan. What you can compare is the iscan and the DragonFly alone. The iscan will be cheaper, then DragonFly will probably win in PQ. You can use the Mosquito in combination with either the iscan and the DragonFly or you can use the Mosquito without any scaler/deinterlacer, too.

Before we all get too concerned about this, I don't believe that the Mosquito necessarily works better with progressive sources, only that there was an "issue" discovered with some interlaced sources that did not occur with a progressive source. A firmware update should correct this issue and make it moot. My guess is that this will be addressed by Algolith soon.

Obviously, for SD sources, given the lauded de-interlacing/scaling ability of the DragonFly/Realta, for best results, I think we would all like to input an unprocessed 480i signal.

Just my $.02

Alan

Oh, good to hear.

Just thought about it again: I think if the source is encoded progressively, then the Mosquito might work better if it gets progressive signals, because only that way it will see the whole MPEG mosquito "area" in one piece. In interlaced mode it will only see half of it. However, if the source is encoded in interlaced mode, the mosquito might fare better with interlaced signals. At least that's how I'd imagine it should work.

But honestly, I don't know enough the internals, so I guess I should better stop guessing... :p

I guess I had a naive dream of plugging all of my video sources into one processor and feeding my plasma with it. I assumed most of the good scalers WERE performing noise reduction in order to improve SD feeds....well, i should have known salvation will never come in one box! :)

I think the Mosquito does work slightly better on progressive sources. That means, that it should receive 480p or 576p.

I tried it on a DVD with a Faroudja deinterlacer and it works quite nicely.

On a day to day basis, I only use it for interlaced sources. It works wonderfully.

Hi dk358,

Rest assured that there are manufacturers aware of the potential of the Realta chip who fully realize that many of us prefer "all in one box" solutions. The use of multiple chips in a single unit will enable this to come about. However, since these chips are largely dependent upon the programming on-board, each manufacturer sporting an HQV chip will actually be selling quite different image processing.

So if you really want a one box solution, there will probably be one on the horizon. But if you really want Mosquito noise reduction and DragonFly scaling, then you'll have to wait until Algolith decides to deliver it all the same box. Meanwhile, you can go double boxes with Alan and David's cool powerbuy, or wait until there's more choice on the market possibly including single boxed multichip solutions.

Cheers,

I agree. Put my order in yesterday. Hoping the Mosquito will be here t'morrow.
Javry

It's not just the compression scheme being being block based that matters, its the encoding scheme that counts more so.


According to Algolith, the Mosquito is designed to specifically process MPEG2 material (see page 4 of the manual). How well it will perform on MPEG4 is to be seen - mainly because MPEG2 is DCT based and MPEG4 is wavelet based.


If the Mosquito is expecting MPEG2 decompressed data, flagging suspect blocks, then is doing its funky processing then great. However if MPEG4 was the source, then the nature of the artifacts have changed. There might be enough differences in the artifacts (Gibbs, quantization, et. al.) that the processor may perform sub-par.

Without knowing what type of filtering they are doing - the effects, improvement or DE-provement, are not known until wavelet like data is used. I imagine a dedicated AVS enthusiast could do some experimentation if they had some high quality input data, a Mosquito, a HTPC, and some decent compression software.

I suspect processing MP4 will help only minimally and HOPE that it does not additively add to any degredations.

Is Algolith Dragonfly the only processor based on HQV?

Dave

For now, yes. Dennon used the chip already in one of their DVD players. Lumagen was rumored to be working on a scaler based on the chip. JVC was also as well as a replacement for the rebadged Faroudja 1010 scaler they bundle with their HD2K. They have a pro rear projection monitor based on the HD2K that uses a Realta based scaler already. I thought I remembered hearing an audio processor was going to use it, but I can't remember the details (or if it's even true). I may be missing some others. More are sure to come, but the DragonFly looks like it's going to be the first available general purpose scaler based on the chip.