PixelFX Morph

[PaTaito]

Just pre-ordered mine. Cannot wait! That transparent magneta morph case on their website pics is utterly gorgeous too, so I hope they release different shells at some point.

Say what you like about the pixelfx twitter comments, but the price they are releasing the Morph at surely gives some credence to the comments. They are in a better position than most to know wholesale pricing on large quantities of parts.

Chriz2600, Woozle, and Dan Kunz have done a hell of a lot for us hobbyists over the last 7 years or so, and from my experience the aftersales on their gear is fantastic...Chriz2600 went WAY out of his way to help with a small ps1digital issue I had last year.

[BuckoA51]

Say what you like about the pixelfx twitter comments, but the price they are releasing the Morph at surely gives some credence to the comments

Ok let's nip this in the bud. the comment you refer to was in bad taste even for a competitor to say, and hopefully Dan will think harder before he posts anything like that again. The full comment also said that it was IF RetroTINK were using a cheaper part, and now we know the final specs of the RT4K we know they are not.

Put it another way, you can certainly make the argument that the RT4k is expensive, but it's not because there's an unreasonable mark up on it, considering the parts in it. Is its FPGA overkill? Time will tell I suppose!

[H6rdc0re]

Say what you like about the pixelfx twitter comments, but the price they are releasing the Morph at surely gives some credence to the comments

Ok let's nip this in the bud. the comment you refer to was in bad taste even for a competitor to say, and hopefully Dan will think harder before he posts anything like that again. The full comment also said that it was IF RetroTINK were using a cheaper part, and now we know the final specs of the RT4K we know they are not.

Put it another way, you can certainly make the argument that the RT4k is expensive, but it's not because there's an unreasonable mark up on it, considering the parts in it. Is its FPGA overkill? Time will tell I suppose!

Perhaps with joining the MARS team he'll get experience with the Efinix family of FPGAs. Surely switching to the Trion TI180 should be Mike's first priority. This will allow him to lower costs and maybe even the price for consumers all with him having a higher profit margin.

[SavagePencil]

Surely switching to the Trion TI180 should be Mike's first priority.

Do you offer lessons from you armchair?

[orange808]

Say what you like about the pixelfx twitter comments, but the price they are releasing the Morph at surely gives some credence to the comments

Ok let's nip this in the bud. the comment you refer to was in bad taste even for a competitor to say, and hopefully Dan will think harder before he posts anything like that again. The full comment also said that it was IF RetroTINK were using a cheaper part, and now we know the final specs of the RT4K we know they are not.

Put it another way, you can certainly make the argument that the RT4k is expensive, but it's not because there's an unreasonable mark up on it, considering the parts in it. Is its FPGA overkill? Time will tell I suppose!

Perhaps with joining the MARS team he'll get experience with the Efinix family of FPGAs. Surely switching to the Trion TI180 should be Mike's first priority. This will allow him to lower costs and maybe even the price for consumers all with him having a higher profit margin.

Well, maybe MARS should adopt copyleft like many other devs do.

Looks like I poked a hole in your little boat.

[Lopenator]

if you ordered the morph on Black Friday sale ita a drag that you gotta pay shipping twice if you want the ADC. Halfway ruins the deal, no?

[bobrocks95]

if you ordered the morph on Black Friday sale ita a drag that you gotta pay shipping twice if you want the ADC. Halfway ruins the deal, no?

For US shoppers probably still a deal (how much was shipping?) though anyone in another country yeah not worth it most likely.

[Lopenator]

if you ordered the morph on Black Friday sale ita a drag that you gotta pay shipping twice if you want the ADC. Halfway ruins the deal, no?

For US shoppers probably still a deal (how much was shipping?) though anyone in another country yeah not worth it most likely.

shipping was 9.45$. I'm US

[H6rdc0re]

Surely switching to the Trion TI180 should be Mike's first priority.

Do you offer lessons from you armchair?

Better FPGA for $200 less is a bad idea, no?

[Guspaz]

There is no such FPGA as the Trion TI180. The highest-end Trion is the T120. You may be referring to the Titanium Ti180.

[H6rdc0re]

There is no such FPGA as the Trion TI180. The highest-end Trion is the T120. You may be referring to the Titanium Ti180.

Yep it's the Titan or Titanium Ti180. The confusement was because there's a Trion Ti120 and a Titanium Ti120.

Never the less $60-70 for a Ti180 vs $270 for the Cyclone V used in the RT4K.

[orange808]

There is no such FPGA as the Trion TI180. The highest-end Trion is the T120. You may be referring to the Titanium Ti180.

Yep it's the Titan or Titanium Ti180. The confusement was because there's a Trion Ti120 and a Titanium Ti120.

Never the less $60-70 for a Ti180 vs $270 for the Cyclone V used in the RT4K.

You have no idea if Mike licensed any IP for his project. There's an entire ecosystem built around these platforms. Well, around the major platforms. I have no idea what's available for budget offerings. So, we really don't know everything that went into the design decisions.

AFAIK, verilog can be used on every FPGA platform right now. So, I don't understand what exactly a dev is going to "learn" about designing for any particular platform--outside of identifying idiosyncrasies and gotchas. You implied something about a teaching moment for a dev that doesn't really need one.

Last time I heard, even the OSSC Pro was using some licenced bits for deinterlacing. Although, that may have changed.

[Jefferson]

- Auto analog resolution detection/switching (huge for PS1/Saturn and other consoles with tons of resolutions between different screens)

This is the real killer feature of the RT4K for me. I haven't heard anything about the Morph getting a similar feature (and I get ignored on their Discord server when I ask).

[bobrocks95]

- Auto analog resolution detection/switching (huge for PS1/Saturn and other consoles with tons of resolutions between different screens)

This is the real killer feature of the RT4K for me. I haven't heard anything about the Morph getting a similar feature (and I get ignored on their Discord server when I ask).

I agree, though some other people weren't impressed at all. For me this changes my plans and I can save a lot of money not getting several consoles HDMI-modded where I'm already happy with their analog output. Not having to look up and worry about what resolution the game I'm playing is in to get optimal sampling is huge imo.

[Unseen]

AFAIK, verilog can be used on every FPGA platform right now. So, I don't understand what exactly a dev is going to "learn" about designing for any particular platform--outside of identifying idiosyncrasies and gotchas.

Every FPGA contains vendor-specific hardware, which may vary in features and with few rare exceptions cannot be described in generic Verilog. The synthesis tools tend to be good enough to recognize certain styles of memories and multiplications and map them to hardware memory blocks and hardware multipliers. But there are other important pieces like a clock synthesizer that cannot be described in generic synthesizable Verilog, so for a non-trivial design you have to use vendor-specific features.

[BuckoA51]

This is the real killer feature of the RT4K for me. I haven't heard anything about the Morph getting a similar feature (and I get ignored on their Discord server when I ask).

It's a planned feature, Woozle already posted a video of auto crop/scale to show it could be done based on signal/image characteristics. It'll work two ways on Morph, for consoles that support GameID you can have per game settings, otherwise it's planned to work on signal detection.

However do not get me wrong, it's not in the test firmware I have at the moment, while on the RT4k it seems to be pretty much done, polished and fully working.

[VEGETA]

I believe ossc is using intel scaler ip, but no info on what retrotink is using. since he is using intel fpgas then it could be that too... unless he has done all that from scratch which is not a good or painless thing to do. no info on morph scaler ip too.

if you check you find all fpga vendors offer scaler ips which do have de-interlacing and motion-adaptive de-interlacing options, didn't check if they have inverse telecine though.


thus fpga scaler developer of such products can just use the provided ip, and pair it with his own firmware development for his MCU and analog video ADC as well as HDMI in and out... which is not an easy thing to do! I believe the ADC step is the most important of all.

can anyone shed a light on this fpga ip thing?

[orange808]

AFAIK, verilog can be used on every FPGA platform right now. So, I don't understand what exactly a dev is going to "learn" about designing for any particular platform--outside of identifying idiosyncrasies and gotchas.

Every FPGA contains vendor-specific hardware, which may vary in features and with few rare exceptions cannot be described in generic Verilog. The synthesis tools tend to be good enough to recognize certain styles of memories and multiplications and map them to hardware memory blocks and hardware multipliers. But there are other important pieces like a clock synthesizer that cannot be described in generic synthesizable Verilog, so for a non-trivial design you have to use vendor-specific features.

Indeed. Each vendor platform is called a "platform" for a reason, no? In addition to vendor specific IP that I referenced.

You''re describing a general reality of software development of all kinds--and it's similar to procedural coding in that way. No effective and useful programming language of any kind has ever managed to abstract away the hardware (directly around the API) and I strongly dislike every programming language that has ever really tried.

Absolutely. Dealing with specific hardware is at the heart of most any nontrivial coding project of any kind.

[Extrems]

can anyone shed a light on this fpga ip thing?

I think you'll quickly realize it's not Intel VIP from the feature set. Efinix has also no equivalent.

[VEGETA]

can anyone shed a light on this fpga ip thing?

I think you'll quickly realize it's not Intel VIP from the feature set. Efinix has also no equivalent.

do you mean retrotink 4k that is not intel ip or what exactly? if not, then did he design it from scratch?

what do you mean by effinix has no equivalent?

[orange808]

Is any of that really our business?

I interpreted an earlier post in the thread as: a suggestion that Mike needs lots of help and schooling to understand how to develop properly and sell products cheaper. It wasn't said directly, but it was implied.

That's bollocks. First, there were devs suggesting the machine was a cash grab. Next, there's someone suggesting the design is inefficient and the dev needs to learn how to do it right

It's all fucking bollocks.

[Jefferson]

This is the real killer feature of the RT4K for me. I haven't heard anything about the Morph getting a similar feature (and I get ignored on their Discord server when I ask).

It's a planned feature, Woozle already posted a video of auto crop/scale to show it could be done based on signal/image characteristics. It'll work two ways on Morph, for consoles that support GameID you can have per game settings, otherwise it's planned to work on signal detection.

However do not get me wrong, it's not in the test firmware I have at the moment, while on the RT4k it seems to be pretty much done, polished and fully working.

Ya I saw Woozle's demo video. I couldn't tell if that was the end of their work on it, or if there was more coming with signal detection to profile mapping. I feel like things don't get communicated that well. A lot of the games I want to play have resolution switching (PS1/Saturn), so I don't know how helpful GameID would be for them.

[Guspaz]

It's a planned feature, Woozle already posted a video of auto crop/scale to show it could be done based on signal/image characteristics. It'll work two ways on Morph, for consoles that support GameID you can have per game settings, otherwise it's planned to work on signal detection.

It's not the same problem to solve, as cropping doesn't require detecting the resolution, it can be done by counting pixels/scanlines with a threshold (as to if a column or row is close enough to black). GameID can help slightly, as it can narrow down the potential set of resolutions that you need to look for (since a game-specific profile could say which resolutions a game might use), but does not by itself solve the problem since games can use multiple resolutions, and many consoles will cause nearly all games to be multi-resolution on account of using an uncommon resolution for their splash screen (like the PS1 or GCN).

[BuckoA51]

Yes please do not think I'm saying this feature will be in the launch firmware, clearly there's still work to be done on it before it's ready I would not expect to see it on the Morph's launch firmware (though auto crop might be).

[SavagePencil]

Any more footage of the Morph in action? I only see a handful of videos here. Dunno who Lo0twig is, but he needs to check his color calibration

[768peeistrash]

The morph so far seems like a barebones extension of the firmware on PixelFX's digital products, RT4k seems much more robust.

The morph price is compelling, but when you add it all up I think it's worth ponying up for 4k. it's been rigorously tested for a year by tons of devs and such.

unless you're only after digital only 4k scaling, then the HDMI only morph makes more sense I think .. I dunno

[TooBeaucoup]

The morph so far seems like a barebones extension of the firmware on PixelFX's digital products, RT4k seems much more robust.

The morph price is compelling, but when you add it all up I think it's worth ponying up for 4k. it's been rigorously tested for a year by tons of devs and such.

unless you're only after digital only 4k scaling, then the HDMI only morph makes more sense I think .. I dunno

The Tink4k is going to age so much better in the long run. I mean, if you're serious about retro games and advanced scalers, which I would think anyone already spending $300+ dollars is, you should just save up extra and get the Tink4k. Mike is such a mad man. The value of his software and frontend is worth the extra price, alone. The Tink5x has gotten 12 official firmware updates, not to mention a ton of Beta firmware in between that have all been available for people. The amount of features and tweaks he has added since launch is insane, and I can't imagine Pixel ever matching it.

That's certainly not to shame anyone who picks the Morph, but when you look at the support the Tink5x got, there's just no other choice for me when it comes to my future scaler choice. It's not even close.

[tongshadow]

Pixel FX's HDMI mods also received updates and new firmware, so it's not fair to assume the Morph wont as well.

[Sargon]

I have no doubt that the Tink4K will be an amazing product with outstanding support.

With that said, I don't think it is fair to speculate that the "Tink4k is going to age so much better in the long run." If anything, I think the gap between the Tink and the Morph may be wider at release and then narrow as the Morph software matures.

[Josh128]

I have no doubt that the Tink4K will be an amazing product with outstanding support.

If anything, I think the gap between the Tink and the Morph may be wider at release and then narrow as the Morph software matures.

Theres no evidence to support that thought though, while we know from past experience that Mike has provided stellar support and updates on the 5X, and has progressively built on that knowledge, which logically should continue with the 4K. We have zero reason to believe the Morph will close the gap on their very first rodeo while Mike continues to improve. If Mike was to do nothing, yeah-- but we all know Mike is not going to "do nothing".