[joevt]

It only takes a few seconds. There are three parts to the firmware. Each one is done separately according to the instructions provided.

[joevt]

Another simpler idea is to use a MST Hub between the converter and the display – the hope is that the converter will accept the MST Hub, and the MST Hub will transmit the DisplayPort to one (or all) of it’s DisplayPort outputs – this requires two things:
1) the DisplayPort input of the MST Hub must accept a non-MST signal from the converter. I think this is likely as StarTech says it’s MST hubs work with macOS which doesn’t support MST for multiple displays (but in this case the input is mirrored to all outputs of the MST Hub).
2) The MST Hub must in some way present to the DisplayPort source (the converter) a DisplayPort device that is different than the G-SYNC display in a way that makes the converter not vomit. I’m not sure about this one. I think it’s a necessary feature of the hub to modify the DisplayPort information, as the hub is able to take an MST input and output to multiple SST outputs. Basically, it takes multiple SST displays and presents to the computer a chain of MST displays. I could be wrong about this. Maybe SST is the same as MST with display count = 1?

[joevt]

That statement does not explain why it doesn’t work. G-SYNC monitors can work with non-Nvidia DisplayPort sources (without using G-SYNC’s variable refresh rate functionality), so why can’t G-SYNC monitors work with this converter? Maybe the eDP chip doesn’t like the DisplayPort of the G_SYNC monitors (but it does work at lower resolutions or refresh rates?). Maybe it can be fixed in firmware.

I wonder if piping the DisplayPort output of the converter into a Thunderbolt add-in card, then extracting the DisplayPort using a Thunderbolt dock or Thunderbolt 3 to Dual DisplayPort adapter would have an effect?

[joevt]

ost, how do you know the pixel rate (MHz) is faithfully duplicated from the HDMI input to the DisplayPort output? Onscreen menus of displays usually only show the vertical refresh rate (Hz). You might be able to use DDC/CI to query the horizontal refresh rate (kHz).

I think partylink is saying that the pixel rate is faithfully duplicated, or at least the converter will not take a 533 MHz input and output 594 MHz with extra pixels for the horizontal blanking?

It would be nice to know for sure how the converter handles various pixel rates (as required to support various resolutions and timings).

[joevt]

MCMG is referring to the problem in the previous post where the optomedia/gofanco/SIIG converters aren’t able to output video to a g-sync display.

The g-sync display is being used as a DisplayPort display without g-sync features. The current speculation is that the problem is due to the fact that the g-sync display does not allow 594 MHz pixel clock timing (has that been confirmed using custom resolution settings on a PC?) and for some unknown reason, the converter will not output the required 533 MHz signal. Either the converter won’t output 533 MHz when receiving a 533 MHz HDMI input (can be tested using custom resolution settings on a PC), or the PS4 won’t provide a 533 MHz HDMI input.

[joevt]

My point wasn’t about being able to play with the settings. It was about being able to get a 4K 60 Hz 533 MHz signal to work. There is currently no explanation why I can create a 4K 60 Hz 533 MHz signal to work on my display, but you can’t get it to work with your display unless the converter is doing something inexplicable like taking 533 MHz and outputting 594 MHz. Can they confirm this behavior? Does output clock not always equal input clock? Is output clock usually not equal to input clock (i.e. input clock may have a continuous range but output clock can only be certain discreet values)?

If you’re saying you can get a 4K 60 Hz 533 MHz signal to work on your display using a PC but not with a PS4 then that would be an important distinction worth noting here.

[joevt]

I can create custom resolutions on a PC with various timings and they’ll work on my Dell P2415Q with or without the adapter. Are they saying that a 533MHz signal from HDMI is converted to 594MHz by the adapter? It doesn’t do that with other timings from 45 Hz to 60 Hz. Well, to be clear, the display only shows the resolution and refresh rate, not the pixel clock so I can’t be 100% sure that some kind of pixel clock conversion isn’t happening. If the pixel clock is being increased, then it would have to add pixels to the blanking areas to maintain the refresh rate. I believe the converter has diagnostic information available via a USB serial console that could help describe what is happening, but I don’t know the necessary commands.

[joevt]

Yeah, it only changed the EDID on the HDMI side. If your display doesn’t work with the adapter from a computer (where you can create any custom resolution) then your display won’t work with anything until they change the DisplayPort side of the firmware.

With the adapter, If your display can work from a computer using the timing specified in the adapter’s EDID but not from a game console then maybe this new firmware will help (but only if the game console honors the information in the EDID).

[joevt]

I have a new firmware 0731 from optomedia that replaces the 1920×1080 EDID with an EDID that has a 3840×2160 60Hz timing with only a pixel clock of 533 MHz. Nothing else has changed. Send me a message if you want to try it out. It works with my Dell P2415Q but the Dell works with all the previous firmwares so I don’t know if this new one will fix anything.

[joevt]

That G-SYNC comment in the SIIG specs is new. G-SYNC monitors can work with non-Nvidia DisplayPort outputs, so there’s no reason why a G-SYNC monitor can’t work with this converter. Can you get full 4K 60 Hz from Intel or AMD graphics?

The term “due to mismatched timings” contains no information. Why can’t they make the timing match? The converter supports almost a continuous range of resolutions and refresh rates. I have discovered a range of refresh rates that inexplicably don’t work at 3840×2160 (in the 35Hz to 44Hz range) but I had no problem with 533MHz 60Hz timing. I don’t know if the DisplayPort output timing matches exactly the HDMI input timing of the converter because the display does not report pixel clock, only horizontal and vertical refresh rate.

The updated firmware (there’s only one so far) for the SIIG has a EDID with 533 MHz as primary timing, and 594 MHz as secondary, but I don’t know if the PS4 Pro will choose the primary which is why I suggested that they should put the 533 MHz in a separate EDID and if that doesn’t work then you need to tell Sony that their PS4 is not using the EDID correctly.

[joevt]

The BL3200PT is only 2560 x 1440, so a cheaper HDMI 1.4 to DisplayPort adapter will be much less expensive and still get you full refresh rate (I’m not sure about HDCP support).

ARTA was not able to get a CMV-535 adapter (similar to the SIIG converter as both have the same firmware so they both probably use the same hardware) to work with a BLP3201PT at more than 2560 x 1440 @ 44 Hz but they are very different displays (the 3201PT is 3840 x 2160).

[joevt]

I’ve received the firmware update from optomedia. It is the 0608 version (June 8), identical to the gofanco and SIIG update (except the instructions are in a pdf instead of an excel file). They also gave me the 0312 version (March 12). I haven’t noticed any improvements since I haven’t encountered the problems others were having (my Dell supports 594, and I don’t have an HDCP 2.2 source to test).

They told me about the change ARTA found with the 3840 x 2160 60Hz EDID, where they added the 533 as a preferred timing and 594 as the second timing. I’ve confirmed that the EDID is now as ARTA described. The three other EDID’s have changed slightly or not at all.

Archer333 said the update didn’t work with the LG 31MU97. Maybe the 533.25 is not close enough to the LG’s 533.28 (not just the pixel clock, but also all the other parameters)? I’ve suggested to optomedia that it would be better if the 533.25 MHz was in a separate EDID than the 594 to be sure the game console doesn’t choose the 594 timing. I’ve also suggested that they make the EDID editable.

“Serial Port Monitor” shows the COM commands for updating the firmware of the DisplayPort and HDMI ICs. I couldn’t get USBTrace to look at the HID commands for updating the firmware of the main device because you have to power cycle the adapter.

[joevt]

I think it would be better if they put the 533 in a different EDID than the one containing 594 – like maybe into the one containing only up 1080p.

You have a GTX 1080 reference card so there should be nothing wrong with your HDMI 2.0 port, and you’ve already tested that with 5xx+ MHz.

I’ve looked at the firmware updates, and converted the hex files to binary files. The main binary seems to have the EDID’s but there is some unknown compression where duplicate groups of bytes seem to occur only once so you need to know how to decompress it before you can edit it. I think the converter’s USB serial connection has diagnostics that could be informative, but we don’t have the software to use it. There are many diagnostic strings in the HDMI and DisplayPort parts of the firmware. I have a terminal program that works with USB serial devices, but you need the commands to enable the communication. All I can get is repeated ; or = at 57600 bps (I’m not even sure if that’s the correct speed).

[joevt]

I believe the problem is with the converter, and not the display.

I don’t think the TMDS clock is the problem. As I stated before, the TMDS clock mentioned in the output of moninfo.exe has a bug where it shows a default max TMDS clock of 165 MHz when the HDMI VSDB does not have enough bytes to actually specify a max TMDS clock.

I looked at the EDID of your BenQ BL3201PT again, trying to figure out where the 178 MHz might be coming from. The CE video identifier (VIC) that is marked as Native is 1920x1080p at 60 Hz. This has a CVT or GTF pixel clock of approximately 173 MHz (5 less than 178). Maybe that’s a coincidence, because VIC #16 is supposed to be 148.5 MHz (CEA/EIA-861, not any of GTF, CVT, and CVT-RB; the latest standard is called CTA-861-G).

Did you test a direct DisplayPort (without converter) connection to the BenQ BL3201PT with the same  resolutions and timings? This will prove that the limitation is with the converter and not your display. This problem is different than the 533/594 MHz problem that other displays have (because you’re trying 2560 x 1440), but your display may have that problem as well if your display can’t do 3840 x 2160 594 MHz without the converter).

I was able to do 2560 x 1440 up to 73 Hz (296 MHz) using the CMV-535 with Dell P2415Q. 74 Hz did not work (300 MHz). Using a direct DisplayPort connection, I can go up to 86 Hz, 350.81 MHz (but not 87 Hz, 355.06 MHz) which exceeds my display’s max refresh rate of 76 Hz. I’m not sure why they differ. I previously reported my 3840 x 2160 tests where all refresh rates from 30 to 60 Hz worked with a direct DisplayPort connection, but the converter couldn’t do the range between 35Hz and 44Hz (but did the ranges 30 to 34 and 45 to 60Hz). Very strange.

[joevt]

Another interesting thing: The screen shots in the SIIG and gofanco firmware update instructions have the model number CV-524 that was mentioned previously before (advertised by gomax) and which optomedia says is an earlier version of their CMV-535. I’m not sure which Taiwan company is the original manufacturer. I don’t know if the CV-524, SIIG/gofanco, and CMV-535 can all use the same firmware. I’m still waiting to hear back from optomedia about their firmware update.

[Author]

Posts