RetroTink 4K All But Officially Announced ....

[bobrocks95]

or you get that sharp and it's a flickering mess on the checkerboard pattern and any grayish colors.

What do you mean? You get it sharp but its too sharp?

If you've manually set phase before with an OSSC or another Tink product, there's a point when you're out of phase where you get rapidly moving horizontal lines of interference that jump around the screen. FBX's OSSC sampling videos show it off.

There's a sweet spot for no flickering and a sweet spot for maximum sharpness- most consoles have those lined up but Dreamcast seems to be pick one.

[EnragedWhale]

Atmos is supported via Dolby Digital Plus. It is not supported via Dolby TrueHD or MAT.

The PS5 can output Atmos in PCM. Will this be reduced to DD+ when run through the Tink then?

Can’t say I dug too deep but my Denon still displayed Dolby Atmos on the PS5 dashboard for whatever that matters.

[Guspaz]

The RT4K can't convert anything, it can only pass through whatever digital audio it receives untouched. If you're doing MAT over PCM then the RT4K might pass that through successfully, since the RT4K can pass up to 7.1 PCM. It can also do Dolby Digital Plus. It can't do Dolby TrueHD, which is the Atmos format you'll find on blurays. Streaming, I believe, tends to use DD+ atmos due to the lower bandwidth requirements.

[Konsolkongen]

With PS5 the AVR says “Atmos/PCM”:
https://cdn.discordapp.com/attachments/ ... 171778dc8&

When using streaming apps on the TV the AVR says “Atmos/DD+”. Not sure if there is any difference, but you would think that PCM could have increased audio quality?

It doesn’t really matter regardless. For the odd 2D game I would want to rescale through the RT I would never hear the difference

[Ricdeau]

With PS5 the AVR says “Atmos/PCM”:
https://cdn.discordapp.com/attachments/ ... 171778dc8&

When using streaming apps on the TV the AVR says “Atmos/DD+”. Not sure if there is any difference, but you would think that PCM could have increased audio quality?

It doesn’t really matter regardless. For the odd 2D game I would want to rescale through the RT I would never hear the difference

PCM will be the preferrable option when available because it's a raw signal (lossless) that your receiver or audio handling device has to decode. DD+ is a compressed signal (lossy) to lower bandwidth requirements. Both PCM and DD+ can carry Atmos as Guspaz pointed out, and what you are seeing sounds completely expected.

[SavagePencil]

New firmware dropped: https://consolemods.org/wiki/AV:RetroTI ... sion_1.0.4

Patch notes:
Fixed hidden '.' file bug introduced when using the SD card with Mac OS X.
MiSTer DV1 detected and parsed
Can be disabled in HDMI RX menu, default on
Automatically tunes the crops (if core provides info) and takes control of the decimation factor
New HDMI modes entries added for MiSTer DVI (240p, 288p, 480i, 576i and unknown/generic)
https://github.com/MiSTer-devel/Main_MiSTer/issues
Added 32 kHz audio sample rate detection (diagnostics page needs to be fixed, since it still displays 44.1 kHz)
Various FPGA optimizations and added the ability to override DE with DV1 info

[orange808]

The new firmware likes to crash while trying to dial in tate rotation with analog signals. I don't recall being able to lock up the machine on the previous firmware. I have to "hard reboot" the machine by disconnecting the power supply to recover and I'm uncomfortable with further testing.

-------

I'm also not having any luck with 1080p/60Hz inputs over HDMI and 1080p/240Hz output. The machine works fine in other output modes and the scaler syncs up with the display fine. I can navigate the menus, but there's no processed video. (Just menus on a black screen.) Is there a limitation for the input signal with 1080p/240Hz?

[DejahThoris]

New firmware dropped: https://consolemods.org/wiki/AV:RetroTI ... sion_1.0.4

Patch notes:
Fixed hidden '.' file bug introduced when using the SD card with Mac OS X.
MiSTer DV1 detected and parsed
Can be disabled in HDMI RX menu, default on
Automatically tunes the crops (if core provides info) and takes control of the decimation factor
New HDMI modes entries added for MiSTer DVI (240p, 288p, 480i, 576i and unknown/generic)
https://github.com/MiSTer-devel/Main_MiSTer/issues
Added 32 kHz audio sample rate detection (diagnostics page needs to be fixed, since it still displays 44.1 kHz)
Various FPGA optimizations and added the ability to override DE with DV1 info

As someone who hasn't run directvideo in any capacity previously (have used an analog io, mistercade, etc previously), can I just go HDMI into the RT4K and it will just work, or do I need to use an adapter like if I were to use it with anything else?

[Guspaz]

The new firmware likes to crash while trying to dial in tate rotation with analog signals. I don't recall being able to lock up the machine on the previous firmware. I have to "hard reboot" the machine by disconnecting the power supply to recover and I'm uncomfortable with further testing.

What power supply are you using? The power draw of the RT4K isn't constant (either from firmware release to firmware release, or in operation) and it's possible that your power supply is sufficient under normal circumstances but doing the rotation stuff pushes it over the edge. 5V2A is the minimum recommendation. Most TV USB ports can't provide enough power.

I'm also not having any luck with 1080p/60Hz inputs over HDMI and 1080p/240Hz output. The machine works fine in other output modes and the scaler syncs up with the display fine. I can navigate the menus, but there's no processed video. (Just menus on a black screen.) Is there a limitation for the input signal with 1080p/240Hz?

It probably exceeds the available memory bandwidth since it would need to read out the 1080p60 framebuffer at 4x rate. Though it may also just be pushing some other aspect of the scaler past the breaking point (while the OSD itself isn't dependent on the scaler). I've seen people using 1080p240 output successfully with lower resolution input sources, have you had any luck with those? As in, try 240p/480p/720p scaled to 1080p240 and see where it breaks.

You can check an estimate of the current memory bandwidth on the system status screen on page one, under "SDRAM Load".

[orange808]

What power supply are you using? The power draw of the RT4K isn't constant (either from firmware release to firmware release, or in operation) and it's possible that your power supply is sufficient under normal circumstances but doing the rotation stuff pushes it over the edge. 5V2A is the minimum recommendation. Most TV USB ports can't provide enough power.

Yes. I checked the power requirements on the wiki and I (made sure to) use a 5V2A plug.

It probably exceeds the available memory bandwidth since it would need to read out the 1080p60 framebuffer at 4x rate. Though it may also just be pushing some other aspect of the scaler past the breaking point (while the OSD itself isn't dependent on the scaler). I've seen people using 1080p240 output successfully with lower resolution input sources, have you had any luck with those? As in, try 240p/480p/720p scaled to 1080p240 and see where it breaks.

You can check an estimate of the current memory bandwidth on the system status screen on page one, under "SDRAM Load".

Yes. I can get 1080p/240Hz output with lower resolution inputs and it looks very nice. I'll give 720p input a look when I get some time.

It's a shame that 1080p input (maybe?) is too much, although nothing can be done about hardware limitations. I get it.

240Hz output with the black frame insertion (1 frame and 1 blur) is remarkably clear on my beamer--after I finally got HDR dialed in to compensate for the loss of brightness.

I did try out MiSTer with the new firmware and HDMI input. Although, I only looked at DoDonPachi in tate. Works well enough, but I fear for newbies that attempt it.

I also lag tested 240Hz output with the Time Sleuth. 240Hz black frame insertion appears to buffer ~12ms before it begins to send the image up the wire. The scan out takes ~4ms and the total frame scans out (bottom number) at just under 17ms. In terms of lag on a fast display, 240Hz black frame insertion can be essentially free. The total scanout time is virtually identical to connecting a native 60Hz source to my display. 1080p/60 input BFI with 240Hz output without adding lag (on some displays) would have been a marketable feature.

1080p 60Hz input and 1080p 120Hz output works, but the motion clarity isn't as good; should be only 4ms of latency (added to "bottom number") for 120Hz 1080p BFI.

[Guspaz]

120 Hz (and increasingly 144 Hz, which isn't helpful for 60 Hz BFI) is a more practical limit since that's what you'll find on modern TVs, high refresh rate monitors (that is, 180+ Hz) are a lot less common than 4K120 TVs. It's unfortunate that the RT4K can't push 4K120, but the cost increase of going HDMI 2.1 would have been just too much. There's a possibility (as in, it's not necessarily possible, but has not yet been proven impossible) of 4K120 in the future with 4:2:0, but it would probably compromise CRT/overlay effects that did anything with chroma. In the mean time, 1440p120 is the best we've got. Something like 1920x2160p120 or 3840x1080p120 might be possible, though I've not tried them. I can see one of those or the other working better with certain types of mask/overlay/scanline effects than 1440p. You can maybe even push that resolution a bit higher using CVT-RBv2 timings to keep under the 600 MHz pixel clock limit, but I also haven't tested those.

[ZellSF]

New firmware dropped: https://consolemods.org/wiki/AV:RetroTI ... sion_1.0.4

Patch notes:
Fixed hidden '.' file bug introduced when using the SD card with Mac OS X.
MiSTer DV1 detected and parsed
Can be disabled in HDMI RX menu, default on
Automatically tunes the crops (if core provides info) and takes control of the decimation factor
New HDMI modes entries added for MiSTer DVI (240p, 288p, 480i, 576i and unknown/generic)
https://github.com/MiSTer-devel/Main_MiSTer/issues
Added 32 kHz audio sample rate detection (diagnostics page needs to be fixed, since it still displays 44.1 kHz)
Various FPGA optimizations and added the ability to override DE with DV1 info

As someone who hasn't run directvideo in any capacity previously (have used an analog io, mistercade, etc previously), can I just go HDMI into the RT4K and it will just work, or do I need to use an adapter like if I were to use it with anything else?

Yes, just HDMI from MiSTer directly to RetroTink4K, enable direct_video in mister.ini

Auto cropping/decimation is new, it might not work optimally right away. But worst case scenario you just set manual cropping/decimation settings.

[DejahThoris]

Yes, just HDMI from MiSTer directly to RetroTink4K, enable direct_video in mister.ini

Auto cropping/decimation is new, it might not work optimally right away. But worst case scenario you just set manual cropping/decimation settings.

Thank you so much! Will check it out this weekend.

[Josh128]

240Hz output with the black frame insertion (1 frame and 1 blur) is remarkably clear on my beamer--after I finally got HDR dialed in to compensate for the loss of brightness.

What kind of display is your "beamer"?

[Lord of Pirates]

240Hz output with the black frame insertion (1 frame and 1 blur) is remarkably clear on my beamer--after I finally got HDR dialed in to compensate for the loss of brightness.

What kind of display is your "beamer"?

Presumably a projector.

[SavagePencil]

Not one but TWO minor rev drops (same day, no less)!

https://consolemods.org/wiki/AV:RetroTI ... sion_1.0.6

Version 1.0.6
Originally released January 6th, 2024
Click here to download
Patch notes:
Interpolation, Scanline, LCD and Mask effects automatically rotate CW or CCW when RoTaTe is on
Forced PAR and allow DAR to OFF for MiSTer Generic DV1
Compensated for decimation change when MiSTer menu is open in 1:1 PAR mode
Unlocked Auto-crop controls in menu for HDMI sources
New LCD effects
Second mono LCD pattern with adjustable horizontal size
GBR and RGB patterns
Fixed RoTaTe CW crop bug
1440p120 changed to CVT-rb timing
4K50 and 1080p50 changed to CEA-861 with extra blanks instead of using their 60 Hz versions
Changed HDCP Alert

Version 1.0.5
Originally released January 6th, 2024
Click here to download
Patch notes:
Enabled auto-crop for HDMI sources
AUX1 - vertical trim only
AUX2 - crop active and stretch to 4:3
AUX3 - crop active and stretch to 16:9
Adjusted auto-crop algorithm
1080p -> 1440p120 CVT-rb works now

[kamiboy]

While searching for HDMI switches I fell upon this little thing, and others like it:

https://www.aliexpress.com/item/1005005 ... Redirect=y

It seems to have EDID dip switches and capable of outputting to two different displays at the same time. Now, correct me if I am wrong, but would that not sort of imply that they would have to circumvent HDCP for that to work?

I've seen simple HDMI splitter devices that take one HDMI input and send the signal to two different displays, with EDID dip switches allowing you to change the signal to, say, 1080p for one of the displays. That can only work if HDCP is circumvented, right? Since it would require the input signal to be converted for one of the outputs.

If so, other than perhaps suspect build quality, this does seem ideal. It strip HDCP for the PS3, and switch between all my analogue devices, maybe? It is bordering on being cheap enough to be worth a shot, even if it does not do what I expect it to.

Does anyone know how RT4K emulates the analogue DAC for the Analogue systems? My main concern is that this switch might somehow disrupt that process, in which case it would be rendered unfit for its purpose.

[Guspaz]

While searching for HDMI switches I fell upon this little thing, and others like it:

https://www.aliexpress.com/item/1005005 ... Redirect=y

It seems to have EDID dip switches and capable of outputting to two different displays at the same time. Now, correct me if I am wrong, but would that not sort of imply that they would have to circumvent HDCP for that to work?

I've seen simple HDMI splitter devices that take one HDMI input and send the signal to two different displays, with EDID dip switches allowing you to change the signal to, say, 1080p for one of the displays. That can only work if HDCP is circumvented, right? Since it would require the input signal to be converted for one of the outputs.

If so, other than perhaps suspect build quality, this does seem ideal. It strip HDCP for the PS3, and switch between all my analogue devices, maybe? It is bordering on being cheap enough to be worth a shot, even if it does not do what I expect it to.

Does anyone know how RT4K emulates the analogue DAC for the Analogue systems? My main concern is that this switch might somehow disrupt that process, in which case it would be rendered unfit for its purpose.

It sends a specific EDID to the Analogue console. If the switch interferes with that EDID in any way (like modifying or caching a previous one) then it will stop it from working.

[D]

Shame I will have to throw out the original SD card with the custom RetroTINK label. Oh well.

You can peel the sticker off fairly easily and slap it over another. I am kind of hoping Mike's able to offer replacements in the future after getting screwed by the supplier, but it's not a huge deal.

This is what I did. Grabbed an X-acto knife and removed labels from both my SD cards and moved the Tink label to my 32GB Sandisk. It looks mint. Not that it matters at all since you can't see the card once it's in the Tink, but it's nice to have the label just so I always remember which card is for the Tink.

My sd card stopped being able to save/edit on it. Both on the rt4 and also on my pc. So Now I use another fat32 sd card. Luckily that is the only thing that failed on my rt4k

[kamiboy]

While searching for HDMI switches I fell upon this little thing, and others like it:

https://www.aliexpress.com/item/1005005 ... Redirect=y

It seems to have EDID dip switches and capable of outputting to two different displays at the same time. Now, correct me if I am wrong, but would that not sort of imply that they would have to circumvent HDCP for that to work?

I've seen simple HDMI splitter devices that take one HDMI input and send the signal to two different displays, with EDID dip switches allowing you to change the signal to, say, 1080p for one of the displays. That can only work if HDCP is circumvented, right? Since it would require the input signal to be converted for one of the outputs.

If so, other than perhaps suspect build quality, this does seem ideal. It strip HDCP for the PS3, and switch between all my analogue devices, maybe? It is bordering on being cheap enough to be worth a shot, even if it does not do what I expect it to.

Does anyone know how RT4K emulates the analogue DAC for the Analogue systems? My main concern is that this switch might somehow disrupt that process, in which case it would be rendered unfit for its purpose.

It sends a specific EDID to the Analogue console. If the switch interferes with that EDID in any way (like modifying or caching a previous one) then it will stop it from working.

That is what I was afraid of. I think on the other simple splitter device there is a dip switch option to just leave the EDID alone, so that might be solution. I guess the only way to know is to buy and find out.

[orange808]

Seems to be a tug of war over on the wiki with bleeding edge firmwares. What's happening with that?

[SquidHominid]

Seems to be a tug of war over on the wiki with bleeding edge firmwares. What's happening with that?

It's been Mike's policy to keep bleeding-edge ('experimental') firmwares limited to the Discord since he started releasing such firmwares in early 2022. Experimental firmwares are, additionally, NOT intended to make their way to the wiki; the RT5X firmware page on the wiki consistently reflects this. Due to a series of miscommunications, the first few post-release firmwares made their way onto the wiki, and this was spotted and at least partially addressed today after someone started making undesired edits to the RT4K firmwares page.

There are still some bleeding-edge firmwares on the page, but I assume those will be deleted once we get an official, non-experimental update to post beyond 1.0 RC26.

[Guspaz]

It should be noted that the Discord is public and anybody can download the experimental firmwares if they want, but they're not always as polished as the stable firmware that gets pushed to websites.

[SavagePencil]

Ah jeez. I guess that’s one of the tradeoffs of using a public wiki as a distro site

[KnightNZ]

Ah jeez. I guess that’s one of the tradeoffs of using a public wiki as a distro site

It has the word "Experimental" on it pretty clearly and as such not intended for mass consumption. People should realise what they're potentially getting themselves in for, although Mike's firmwares are really if ever truly problematic.

[SavagePencil]

Ah jeez. I guess that’s one of the tradeoffs of using a public wiki as a distro site

It has the word "Experimental" on it pretty clearly and as such not intended for mass consumption. People should realise what they're potentially getting themselves in for, although Mike's firmwares are really if ever truly problematic.

It did not have “experimental” on the wiki anywhere at the time (see the page history when those versions were being posted). Seems they’ve squared it away now.

[SquidHominid]

Ah jeez. I guess that’s one of the tradeoffs of using a public wiki as a distro site

It has the word "Experimental" on it pretty clearly and as such not intended for mass consumption. People should realise what they're potentially getting themselves in for, although Mike's firmwares are really if ever truly problematic.

It did not have “experimental” on the wiki anywhere at the time (see the page history when those versions were being posted). Seems they’ve squared it away now.

The original RT5X firmware channel was named '#experimental-firmware'. The term 'experimental' briefly got lost in the server reorganization due to an oversight.

[orange808]

Can someone share the latest firmware? I strongly disagree with Discord's predatory data harvesting "verification" methods.

They asked for my number and I gave them a fake one. Must be a business run by boys, because they still didn't get the hint.

[Guspaz]

This is very much still a work in progress (by which I mean, literally as we speak, files changed as recently as a few minutes ago), but:

https://retrotink-llc.github.io/firmware/

[SquidHominid]

Can someone share the latest firmware? I strongly disagree with Discord's predatory data harvesting "verification" methods.

They asked for my number and I gave them a fake one. Must be a business run by boys, because they still didn't get the hint.

The latest firmware (1.1.0) is a release firmware, and can be found both on the wiki as well as on the new firmware repository website Guspaz linked. I am unsure if Mike plans to make a firmware update blog post.