[Updated] Exploring MSI’s Firmware Update for the Peak 1000 HDR Mode for Many of Their QD-OLED Monitors

January 22, 2025 Simon Baker

Share this content

Originally published 16 January 2025, last updated 21 January 2025

Index

Introduction

Recently MSI gave us early access to test their new firmware update for several of their QD-OLED monitors, an update that is designed to address issues with the brightness of the ‘Peak 1000’ HDR mode that has plagued many OLED monitors in the market to date. We’ll explain the history to this issue in a moment, but those familiar with modern QD-OLED monitors will probably understand that these HDR modes were darker than intended in real usage, to the point where they were not as bright overall as other modes on the monitor which were not even able to reach anywhere near the panels “peak brightness” potential.

Updated: Several manufacturers have tried to overcome this through different approaches and “fixes”, but none have managed to achieve what we would consider an accurate fix. When we originally published this article, our initial findings pointed towards a proper fix from MSI for this problem, although it seems there are still challenges and issues that need to be addressed unfortunately. We’ve updated this article now since our initial testing, as with the firmware now available for some of their QD-OLED monitors, we’ve received some very useful user feedback, and we’ve been speaking with Tim at Monitors Unboxed this week about his own testing as well. We’ll explain what has happened below and where things are up to.

If you already understand the background to this issue, things like EOTF tracking, Peak 1000 vs True Black 400 etc then you can skip past the next couple of sections and go straight to our results. If you’re not familiar with the background and what causes these problems in the first place then we’ve provided some detail below for you. It’s probably well worth a quick recap anyway.

Original performance of MSI QD-OLED monitors

MSI OLED monitors have two HDR modes available to select between in the OSD menu. We’ve previously published a detailed study about why this issue arises and the differences between common ‘Peak 1000’ (P1000) and ‘True Black 400’ (TB400) modes. You can read that here if you want loads more information.

For the below testing we will use the MSI MPG 321URX, their 32″ 4K 240Hz QD-OLED monitor.

Peak White Luminance Measurements

*From the MSI MPG 321URX, current firmware* *(FW.012)*

The simple peak white luminance measurements for the two HDR modes are shown above. You can see they are both basically the same for larger APL, but for <10% APL the P1000 mode can reach much brighter and gets up to the panel’s full luminance potential of over 1000 nits. Based on these measurements alone you might conclude the P1000 mode is brighter than the TB400 mode.

However, this isn’t providing a full view of what is happening in HDR and doesn’t tell you how the two modes perform in terms of brightness overall in real usage and in real scenes beyond just measuring white test patterns. For that we need…

EOTF Tracking Measurements

*EOTF tracking examples from MPG 321URX in Peak 1000 mode (FW.012)*

These results above were taken from the 32″ 4K 240Hz MSI MPG 321URX using the current FW.012 in its ‘Peak 1000’ mode for HDR (review here).

*EOTF tracking examples from MPG 341CQPX in Peak 1000 mode*

And as another example, these results above are from the 34” 240Hz ultrawide MPG 341CQPX model (review here).

Both screens show a similar behaviour which actually impacts all of the current MSI QD-OLED monitors and it is the same across other models too from many other manufacturers. In the Peak 1000 HDR mode, the EOTF tracking is generally nice and accurate for a 10% APL window size test and smaller. 10% APL is actually the industry standard for HDR measurements which is why in many standard testing scenarios this issue is not immediately evident. Using just a 10% APL test only really provides part of the picture when it comes to evaluating HDR performance unfortunately.

As you can see from the above EOTF examples, as the APL % area increases the brightness rolls off significantly for mid to light grey shades as seen by the dip in the grey measurement line for the monitor in the middle of the graph. This causes those mid to light grey shades to be a lot darker than intended. In real HDR content that causes darker overall scenes than intended.

*EOTF tracking examples from MPG 321URX (FW.012) in True Black 400 mode*

In contrast, above is the EOTF tracking from the 32″ MSI MPG 321URX (FW.012) again, but this time in the other TB400 mode. Here you can see the EOTF tracking is more consistent and accurate at all APLs with a far less drastic dip in brightness of grey shades. As we examined in a lot more detail in our previous article, this EOTF tracking variance is what causes the TB400 mode to look visually brighter in many HDR scenes than the P1000 mode, as the EOTF tracking is much more accurate and there’s not a major dip in the graph like there is in that mode.

Average Greyscale Measurements

We can also “convert” these EOTF measurements and graphs in to another approach which is to measure the luminance of the greyscale for mid to light grey shades, those are the shades which are impacted by the poor EOTF tracking and the dip in the graphs above. Those mid to light grey shades are what tend to impact “brighter” HDR scenes as well, and so have a direct impact on the overall experienced brightness of the content.

We can plot the average of those luminance measurements for the greyscale in both modes like this:

You can see again here in a more simple comparison graph that the luminance of the greyscale is higher in the TB400 mode than the P1000 mode, again demonstrating why that mode looks visually brighter in many situations.

How Other Manufacturers Have Approached a “Fix”

This issue has generally applied across all the different QD-OLED monitor panels and sizes in the market. For the purposes of this article we’ll compare the 32″ 4K 240Hz QD-OLED monitors offered by a number of leading manufacturers.

Asus ROG Swift PG32UCDM

Asus handled the issue for their ROG Swift PG32UCDM 32″ monitor by introducing a new menu option that allowed users to select a ‘dynamic brightness boost’ mode in the menu. We haven’t actually tested the updated Asus PG32UCDM firmware ourselves, but our friends over at Monitors Unboxed have and we’ve included their results below for comparison and discussion.

Asus’ newly added mode artificially brightens the screen by adjusting the EOTF for the smaller APL areas, making mid-to-light grey shades now quite a lot brighter than intended, in order to ensure that larger APL areas were then not too dark. That was how Asus addressed the issue with their screen.

Gigabyte AORUS FO32U2P

Samsung Odyssey G80SD

We saw this similar approach to HDR EOTF tracking from Gigabyte even when they first released their AORUS FO32U2P, and from Samsung with their Odyssey G80SD too. These modes were provided in the menus on their screens too.

Artificially inflating brightness

All these manufacturers tackled the problem by artificially brightening in the image and it’s resulted in what we’d call a “brighter-than-intended” behaviour, instead of the previous “darker-than-intended” behaviour. You can see in all cases that the EOTF tracking is pretty poor, especially for smaller APL where the brightness is often a lot higher than intended. This approach does help ensure that as the APL increases, the brightness doesn’t roll off for light to mid grey shades as drastically as before, and the perceived dimming isn’t as aggressive. However the trade-off is the over-brightening at lower APL.

This over-brightening causes scenes that are overall darker (low APL %) to be a lot brighter than intended, often impacting shadow detail and darkness of the scene. It also causes loss of tonal values in mid and light grey shades, getting crushed to white in many cases. You lose subtle detail in brighter scenes and brighter content quite often. The modes are “brighter” visually for sure, but they aren’t “accurate” unfortunately.

So far all manufacturers have tried to find a balance between overly bright low APL scenes, and overly dark high APL scenes. It has seemed like a sacrifice has to be made somewhere in these modes.

Some manufacturers only offer the “darker” mode

Some other manufacturers have left their screens only with the “darker-than-intended” P1000 mode, most notably the Dell Alienware AW3225QF, for which Dell told us that this was the “intended configuration by Samsung Display” (the panel manufacturer) and that they wouldn’t be providing any update for their monitor. Somewhat strange when everyone else is trying to offer alternative modes. We have reached out to Dell about the situation and our contacts have passed on information to the Alienware product team for consideration. At this time though, we don’t know if anything will be offered for their QD-OLED monitors.

Until now MSI have not made any changes either although we know they have been working on this for a long time, in fact we reported these issues to them back at the end of March 2024.

Some monitors offer users a choice

Some manufacturers have offered both modes – the original Peak 1000 mode which is darker than intended, and then an alternative mode that is brighter-than-intended. This is how Asus tried to fix the issue on their PG32UCDM by adding the ‘dynamic brightness boost’ mode. It’s also how Samsung Electronics handled it on their G80SD with two modes available. We aren’t knocking their approaches, they obviously made these available a lot quicker than it has taken MSI to provide an update and they’ve helped provide a noticeably brighter HDR experience which many people will prefer and value.

The problem is, it’s a somewhat compromised solution really, but it’s better to have a choice than only offer only one mode for sure. We expect most average users to prefer the brighter mode for a more impactful HDR experience, but it’s not without its issues related to accuracy and EOTF tracking.

Our initial testing of the new MSI HDR firmware

When we received the new firmware for testing on the MSI MPG 321URX we had expected the MSI HDR performance to be updated in a similar way to how Asus handled it on their PG32UCDM. Given all these major display manufacturers have all struggled with this HDR mode and decided just to offer users the choice between “too dark” and “too bright”, we assumed it was not possible to fix it properly and truly offer accurate EOTF tracking for all APL. We expected MSI to just release an update to offer the same. However, here’s the new EOTF tracking measurements following the same approach as before:

*MSI MPG 321URX with new firmware update*

Above are the new EOTF measurements in the Peak 1000 mode using the new firmware. These standard measurements suggest that the screen is accurate now for all APL which we were frankly amazed by. This suggested that they have solved the roll-off problems from before which was causing the screen to be too dark in HDR and now offers accurate EOTF tracking in all APL scenes. But just as importantly, it had done this without having to make the screen artificially bright, and is therefore accurate for all APL areas tested.

Peak white luminance remains as high as it was previously, so no change there. This mode continues to be able to offer the highest peak highlights in low APL scenes reaching up to >1000 nits, while now avoiding aggressive dimming for larger APL scenes.

In these tests you can see the massive difference this makes to the average greyscale luminance in this Peak 1000 mode compared to the old firmware performance. Average greyscale luminance, a proxy for real-scene brightness, is almost twice as high in some scenes and larger APL’s.

In these tests the P1000 mode is now overall brighter than the TB400 mode too, including at larger APL – represented here by measurements of the average greyscale luminance again. It’s not significantly brighter of course, but it’s not supposed to be when the EOTF tracking is accurate. These tests suggest that there should no longer be situations where the TB400 mode looks brighter in real use. This was our original conclusion to the testing.

Problems unfortunately remain (updated 21 Jan 2025)

With the new firmware available for a selection of MSI’s OLED monitors we’ve started to see plenty of user feedback about the update, and unfortunately this has pointed towards some remaining problems with this P1000 mode and that’s led us to update this article now. We also spoke to Tim at Monitors Unboxed this week who has been doing his own testing on the similar MAG 321UPX model (the MPG 321URX firmware has not been publicly released yet). He has since published a very insightful video based on some further testing which you can find here.

The problem seems to be that in real-world HDR content, in games and videos, most users are not seeing an improvement in brightness compared with the previous firmware. TB400 remains visually brighter in many scenarios as before. This does seem to be a little mixed, as some users are reporting improvements although there’s many different use cases and configurations and it’s hard to pin down exactly what is going on. There’s also several models which have received this firmware update now (list later on), and we can’t fully rule out differences between them at this stage. It’s likely the same remaining problems apply across the range, but user reports are conflicting.

What is happening?

Based on the extensive testing from Monitors Unboxed, it appears that the results from standard test pattern and measurement approaches is for some reason not translating to actual real-world HDR content. In brighter scenes (higher APL), dimming is still happening like it was before, and this results in the TB400 mode looking visually brighter still in those situations.

Tim found the same accuracy for EOTF measurements that I did and those results are correct. It appears that in various test pattern scenarios (standard APL area, constant APL, custom APL) everything looks correct and significantly improved. However, in real world usage the ABL dimming is still as aggressive as before, and there seems to be very little change to real scene brightness. Tim and several users also noticed strange dimming behaviour in Windows desktop in some situations (i.e. SDR applications within HDR mode) which is not a scenario we tested. That use case is very often a mess anyway, and not a configuration we would recommend for many reasons, but the observations there do seem to hint at the wider problem with the new firmware.

Tim’s hypotheses is that the screen is in some way “scanning” the content being displayed, and then dynamically adjusting the EOTF to try and compensate for panel dimming. In test patterns, including several atypical methods, it looks like this works as intended. But in real-world content, the approach is not sufficient to compensate for the panel dimming and can’t dynamically adjust the EOTF properly. It seems like a sensible hypothesis although we do not have any official word from MSI on how things were set up, or what has been introduced. The inner workings of their firmware update are not known at this stage.

What went wrong with our testing?

We’ll hold our hands up here, we should have carried out more thorough testing at the time, and I apologise that we missed this. We’ve been investigating this whole brightness issue on OLED monitors for a long time now, right back to March 2024, and it’s gone through several revisions and updates during that time. Our aim has always been to provide as much insight and testing data as we can for our readers, to help understand what the issues are and try and identify a root cause for it. From there our aim has been to create repeatable testing approaches that allow us to identify problems and compare HDR performance as best we can. HDR testing is a very complicated area and very little of it is standardised or agreed on across the industry.

We have created a range of different testing approaches over the last 10 months that have, until now, been very effective at identifying issues and measuring performance. It goes beyond simple white luminance measurements of old – testing EOTF performance extensively in a range of APL scenarios, considering greyscale luminance, luminance accuracy and a range of other criteria. We also use a variety of test patterns to try and avoid scenarios where specific exact APL areas can be “tricked”.

Even a “constant APL” test pattern like this example showed accurate performance

We had no reason to doubt this approach as it’s proved effective for a while now, and this is actually the first case we’ve seen where EOTF measurements for some reason don’t translate to real-world experience. As Tim said in his video, he saw the same performance and initially drew the same conclusions I did. Tim also used “constant APL” test patterns, which are another approach to measurements we’ve explored in the past but didn’t specifically use here, which can identify issues in EOTF performance, but didn’t pick up any oddities here.

Our mistake unfortunately was not triple checking how these measurements would cross over to real world content and we should have done. That’s a mistake we won’t be making again, and we’ll definitely be looking to see how we can expand our HDR testing further. This whole area is very complex and there’s a lot of different approaches possible. The trick is to try and find something that is repeatable, allows comparisons properly and reliably, and is easy enough for an average reader to understand. We’ll continue to investigate that, but if anyone has any suggestions or has experience in this area they’d like to share please do get in touch.

We’ve fed this all back to MSI who are investigating further

I personally don’t believe this is a deliberate attempt by MSI to “trick” test patterns or benchmarks. I don’t think you’d spend 10 months investigating it and working on it, only to try and fool people. Especially when that would very quickly be identified anyway as it has been here, although admittedly we should have picked up on this ourselves and for that I apologise.

I expect that the tuning and development has been carried out using standard test pattern approaches, which is the most practical way to measure and tune performance during development in a repeatable and standardised way. It should be represenative of real performance, and as we said earlier this is the first time we’ve ever seen the measurements not translate in to real-world performance. This seems like an unexpected oversight. However, the adjustments made to the HDR tuning, the ABL compensation, the brightness algorithm, or whatever you want to call it are not being realised in real-world content outside of those test patterns.

As soon as we became aware of these remaining issues, we fed everything back to MSI – thank you to those who provided constructive and thorough personal testing too. We’ve confirmed that this has all been fed back to the product development teams for further investigation and they are looking in to it and what can be offered through a further update and fix. We remain hopeful that something can be done here, a change has already been made and was working well during specific situations, and we hope this can be adjusted to allow correct operation in other situations, namely real-world HDR content. We will of course provide further updates as soon as we have any.

Although this initial firmware update hasn’t fully addressed the problems with HDR on these screens, we know MSI are committed to providing updates and improving these monitors and I think that’s been evident by the wide range of improvements and updates they’ve offered in the last year in various areas, from colour modes to input lag improvements. They are clearly investing time, money and effort to improve these products and address user issues and that’s great to see. Let’s hope they can make some adjustments here and utilise the work already completed to offer HDR improvements.

The delay with some models is unrelated

On a side note we can put any conspiracy theories to bed here about the delays with the flagship MPG 321URX and MPG 271QRX models. We had already asked MSI about this and have been told that there are some minor bugs with MSI OLED Care that need to be addressed which was causing the additional couple of weeks delay with some models.

HDR Demo and Test Video

Test and demo the HDR on your display using our handy compilation, highlighting black depth, contrast and peak brightness capabilities.
[View here]

Details and FAQ about the New MSI Firmware

[Updated] Which screens will receive this new firmware update?

The new firmware is already available for the following QD-OLED monitors:

MPG 341CQPX – 34″ ultrawide 240Hz model (firmware link)
MPG 491CQPX – 49″ super ultrawide 240Hz model (firmware link)
MAG 321UPX – 32″ 4K 240Hz (firmware link)
MAG 271QPX – 27″ 360Hz model (firmware link)
MPG 321CURX – 32″ curved 4K 240Hz model (firmware link)

These models are still pending. According to a post from MSI on Reddit the update for these models should be available in late Jan / early Feb. That date may change now with the additional findings:

MEG 342C QD-OLED
MAG 321UP
MAG 321CUP
MAG 271QPX QD-OLED E2

We are waiting to confirm availability for these models with MSI. We were originally told they would be available on 17 Jan, but that has not happened. The MSI rep on Reddit didn’t include these models at all. They should still be coming at some point.

MPG 271QRX – 27″ 360Hz model
MPG 321URX – 32″ 4K 240Hz model
MPG 491CQP – 49″ super ultrawide 144Hz model

Other models notable by their absence we are waiting to confirm with MSI on:

MAG 341CQP – 34″ ultrawide 175Hz model

This firmware release will also apparently optimize the OSD user interface of the MSI OLED Care features. The position of messages will shift to bottom side they tell us. The boot-up logo will also be updated.


Browser Alerts	Follow us on X (Twitter)	Subscribe	Support Us