Sharpening vs Detail: What Retro Scaler Sharpening Filters Actually Do

The first thing to unlearn about retro upscaling is that a scaler is supposed to make the image sharper. It is not. An upscaler is a format converter. It takes a 240p or 480i analogue signal and turns it into a digital frame your flat panel can show. That is the whole job. Crispness is a filter you add on top of that conversion, and it is doing something entirely different from detail. When you push the sharpening slider on your OSSC or RetroTINK 5X-Pro, you are not recovering information that was lost in the analogue-to-digital sampling. You are asking the scaler to invent edge definition where none exists in the source data. The distinction between sharpening vs detail retro upscaling is the difference between a filter that fakes edge definition and pixel data that was actually captured from the console. A CRT never needed this fake crispness, because the electron beam was already doing something no LCD can do.

Why Pixel Art Dies on Modern Panels Without Help

The reason people reach for sharpening in the first place is that a 240p image on a 4K OLED without any processing looks soft and blurry. That is not a failure of the scaler. It is a mismatch between the source resolution and the panel's native resolution. A 240p image is roughly 320 by 240 pixels. Stretched to 3840 pixels wide, each source pixel becomes a block of roughly 15 by 15 screen pixels. The panel has to do something with that block, and if it simply scales it, you get fat, soft-edged pixels that look like they are covered in Vaseline. But the answer is not to sharpen. The answer is to let a proper scaler like the RetroTINK 5X-Pro or RetroTINK 4K handle the scaling with integer math, so that each source pixel becomes an even, predictable block of output pixels. That is detail preservation. It looks clean. It does not look like a CRT, but it looks honest.

The Ringing Halo Is the Signature Failure of Aggressive Filters

The ringing artifact is not subtle, and it is the first thing you will notice when you accidentally leave a scaler's sharpening filter on. Take a SNES sprite with a dark outline, say a green Koopa shell on a light background. With sharpening at 0 on the RetroTINK 5X-Pro, the edge of the shell is a clean transition from dark green to the background colour. Turn the crispness filter up to 7 or 8, and you will see a bright white line appear just outside the dark edge, and a slightly darker line just inside it. That is the overshoot and undershoot. On a moving sprite, this halo becomes a shimmering edge that crawls across the background, and it is the single most common reason a retro setup on a modern TV looks 'digital' in a bad way. FirebrandX, who publishes calibration profiles for the RetroTINK and OSSC, has documented this exact behaviour in his testing patterns, including the dead pixel test and the zone plate pattern. His profiles set sharpness to 0 because he measured that any positive setting degrades the image more than it helps.

Why the Halo Was Never Present on a CRT

The reason a CRT never showed this halo comes down to physics. A CRT's electron beam has a finite spot size, and the intensity profile of that spot is roughly Gaussian. When the beam sweeps across a dark-to-light transition, the phosphor does not turn on and off instantly. It ramps up over a small distance, and that ramp is what creates the natural softness. Pixel artists knew this. They used the limited bandwidth of the video signal to their advantage, and they often relied on the beam's spread to blend colours that the console could not output simultaneously. A Sony PVM-20L5 showing a 240p image shows a sprite edge as a subtle ramp from the sprite colour to the background colour over about one scanline. The eye interprets this ramp as a sharp edge with no visible halo. A modern scaler showing the same sprite with a sharpening filter at 8 replaces that ramp with an abrupt step and adds a bright fringe. The fringe is not detail. It is a bug in the filter's math, and it cannot be fixed by turning the crispness down slightly. It has to be turned off or set to a level so low that the filter is effectively transparent.

What the OSSC Will Never Do

The OSSC will not invent detail. It will not make a blurry composite signal look like RGB. It will not add texture to a flat background. And critically, it will not sharpen. If your goal is to make the image look 'crisper' in the way a phone camera's HDR mode does, the OSSC is the wrong tool. It is a scaler, not a filter. It takes a signal that is already clean and presents it at a higher resolution without degrading it. The RetroTINK 5X-Pro is also a scaler first, but it offers a crispness filter as an option for people who want a softer image to be artificially enhanced. The OSSC has no such option, and that is not a feature it is missing. It is a design decision that the retro community has validated.

The Gamma and Contrast Trap

Part of the sharpening trap is that it interacts with the scaler's contrast and gamma settings. A sharpening filter increases local contrast at edges, but it does not increase global contrast. If you set the RetroTINK 5X-Pro's sharpness to 4 and then boost contrast to compensate, you end up with an image that has crushed blacks and blown-out whites, which makes the ringing halo even more visible. The correct approach is to set sharpness to 0, set contrast to the scaler's calibrated range, and let the panel handle the rest. The RetroTINK 4K has a detail enhancement feature called Texture Adaptive Sharpening, but it is separate from edge sharpening and is designed to improve the appearance of textured regions like grass or fabric without touching edges. It is a more sophisticated tool, but its default is also 0, and it should stay there unless you have a specific reason to change it.

How to Test for Chroma Issues Before You Buy

To know whether your scaler is dropping chroma, use a test pattern with thin red and blue vertical lines. If the lines blend into a purple smudge, your signal path is subsampled. This is a different test from a sharpness test, and it is the one that will show you the difference between a scaler that is doing its job and a scaler that is adding artifacts through the digital pipeline. The FirebrandX calibration profiles include this pattern. Download them. It is the single most useful thing you can do to diagnose a soft image.

The EDID Problem

The EDID is the data your TV sends to the scaler that describes what resolutions and colour spaces it supports. Some capture cards, and some budget TVs, report an EDID that says 4:4:4 is supported when it is not, or vice versa. The result is a signal that is negotiated incorrectly, and the scaler sends a format the panel then has to convert again. This is a known failure mode with the RetroTINK 5X-Pro, where some capture cards report the 1080p output as 1080i, causing a deinterlacing artefact. The fix is to check the scaler's on-screen display for what signal it is actually sending, and to use a panel that you have verified supports 4:4:4 at the resolution you have chosen. This is not a niche problem. It is the difference between a pixel-perfect image and a smeared mess.

What the mClassic Is Good For

The mClassic is useful for one specific scenario: a modern console like a Switch that outputs 1080p and needs a small boost to look better on a 4K TV. It is not a substitute for a proper scaler like the RetroTINK or OSSC, and it will not make a 240p game look like it does on a CRT. The mClassic's claimed 1440p output is a 1440p frame with a 1080p active image area that has been upscaled, not a native 1440p render. The gap is a full resolution tier. If your goal is to play NES or SNES games, the mClassic is the wrong tool. If you are playing a modern game that needs a slight crispness boost, it might help, but the artifact it introduces is real.

The One Test You Can Do at Home

You do not need a test pattern generator to see the artifact. Load a game with a sprite that has a dark outline and move it against a light background. If you see a white halo trailing the sprite, the scaler's sharpening is too high. If you see a dark shadow ahead of the sprite, that is undershoot. Both are sharpening artifacts, and both indicate that the filter is on. The RetroTINK 4K's Adaptive Sharpness filter, which is separate from its Detail Gain, will show this halo at a width of 1 output pixel at default strength. That is why the default is off. The RetroTINK-4K's Texture Adaptive Sharpening, by contrast, affects textured regions only and does not create halos on edges. It is a rare example of a detail filter that does not cause the ringing artifact. But it is still an enhancement, and a purist will leave it off.

The One Way to Get Truly Zero Lag

The only way to get true zero lag is to use a CRT. A Sony PVM-20L5 has zero processing lag because it is an analogue device. A scaler cannot compete with that. But if you need a modern screen, the OSSC is the closest you can get because it does not use a framebuffer. It has a theoretical minimum of one scanline of latency, which is the time it takes the scaler to read the input and output the result. That is not a compromise. That is a physical limit.

Does sharpening add detail to a retro game?

No. Sharpening adds edge contrast, which creates a halo effect. It does not create new pixel data. The only way to add real detail is to use a scaler that can process the original signal at its native resolution and output it without degradation, which means using a high-quality scaler with 4:4:4 chroma and a sharpness setting of 0.

Why does my OSSC look soft on my OLED?

An OLED has a different pixel structure than a CRT, and the OSSC is doing no processing beyond line multiplication. If you want more crispness, you are using the wrong tool. The OSSC is for purity. If you want a softer, more CRT-like image, use a scanline filter.

What is the best sharpness setting for the RetroTINK 5X-Pro?

The default of 0 is correct. FirebrandX's calibration profiles confirm that any positive setting adds a measurable artifact. If you want more perceived crispness, increase the resolution to 1440p and enable 4:4:4 chroma.

Is the mClassic worth it for retro gaming?

No. The mClassic is for modern consoles. It adds a visible halo and does not handle 240p correctly. For retro gaming, a RetroTINK or OSSC is the right choice.

Why does my 4K TV still look blurry at 1080p?

Your TV is scaling 1080p to 4K internally, and that scaler is adding artifacts. Set your scaler to output native 1080p and let the TV do nothing, or use the RetroTINK 5X's 1440p mode to avoid the TV's scaler entirely.

What is the difference between edge sharpening and detail gain?

Edge sharpening increases contrast at edges and creates halos. Detail gain increases micro-contrast in textured regions and does not affect edges. The RetroTINK 4K has both, but the detail gain is off by default because it is an enhancement.

Can I fix chroma subsampling with a sharpening filter?

No. Chroma subsampling removes colour information that cannot be recovered. The only fix is to use a screen and scaler that support 4:4:4.

Set your scaler's sharpening to 0, turn off the TV's own sharpness control, and enable the TV's game mode. Then download the FirebrandX calibration profiles and run the zone plate pattern to verify that you are seeing no halo. If you see a halo, lower the sharpness until it disappears. If you see no halo, you are done. The image you are looking at is the truth of what your console outputs. It is the image the pixel artists intended you to see, before a filter added a ghost. Do not be tempted to turn it up. The artifact is not a style choice. It is a defect, and it will not make the game look better. It will make it look like it is being emulated badly. When you have done this, you will understand why the community calls the default settings 'reference'. The image is clean because the scaler is not lying.

• RetroTINK-5X Pro Sharpness Range: 0-10, default 0
• RetroTINK-4K Adaptive Sharpness Range: 0-100, default 0
• mClassic Halo Width: 1-2 pixels at 1080p to 1440p