Tech Explained: Memory LCD

Posted under Tech Explained on August 23rd, 2026 by

Memory LCDs have been around for a while. It came into my radar only recently. I thought it deserved an dedicated explainer. It sits right in the middle of a TFT LCD and e-paper. Let’s discuss that today.

A Memory-in-Pixel LCD is still an LCD, but each pixel contains a small memory circuit. In Sharp’s panels, its 1 bit/pixel. Once written, unchanged pixels do not need the same image streamed continuously. The panel is reflective too, so daylight replaces the backlight used by most TFTs.

Sharp Memory LCD Screens

Its really power efficient. Sharp’s 1.28-inch, 128 × 128 LS013B7DH03 specifies about 1µW typical while holding an unchanged image and 50µW when updating the full screen once per second. A similar 1.3-inch TFT module can use around 90mW just for its backlight. E-paper can approach zero between updates, but a typical 1.54-inch B/W panel uses about 26.4mW while refreshing and takes around 2s for a full refresh. Hence it kinda sits in middle.

You also get excellent sunlight readability like e-paper, but much faster response for changing numbers, menus, graphs and simple animation. Indoors it is less vivid than a backlit TFT, and panels are usually monochrome or some limited-colour.

The interface works well with small MCUs. A 128×128 monochrome frame buffer is only 2kB. Sharp’s panel uses a simple 3-wire serial interface, and you can rewrite only the lines that changed instead of pushing the entire frame.

There is one important difference from e-paper. The LCD still needs a tiny amount of power and periodic VCOM polarity inversion, so it cannot retain the image indefinitely after power is removed.

I think its sweet spot is an always-on battery device where the screen changes often, such as watches, meters and outdoor instruments. If the image changes once an hour, e-paper is hard to beat. If it changes every second and battery life matters, Memory LCD becomes very interesting. Cost remains a constraint. These panels are generally above commodity TFTs and closer to small e-paper modules. Do check them out for your future requirements.

If you liked the post, Share it with your friends!

Leave a Comment

Your email address will not be published. Required fields are marked *

Advanced Tech: AI Text Watermarking

Posted under Advanced Tech on August 16th, 2026 by

In an older post we looked at how SynthID hides a watermark inside AI-generated images. Today, let us look at text watermarking and how it works. This week Anthropic says Claude models will use a version of Google DeepMind’s SynthID-Text for all text it generates.

So an LLM writes one token at a time. At each step it usually has several reasonable choices. For eg. “The regulator stayed stable even when the load suddenly _” Here changed, increased or jumped can all fit the line.

SynthID based AI watermarking

SynthID-Text uses this freedom to leave a statistical signature. A secret key, together with the recent text, generates hidden scores for possible next tokens. The sampler chooses from sensible options, and over many steps those choices become correlated with the secret pattern. These scores change with context, so there is no fixed list of “watermark words”. The same token can support the watermark in one sentence and not in another.

A detector with the matching key can later recreate the expected scores and check whether the chosen tokens match that pattern more often than chance. One sentence gives little evidence. Longer passages give more choices to test, so confidence improves. This can be added to many LLMs at the sampling stage without retraining. Each provider can use its own key and configuration. Claude’s detector would therefore look for Claude’s pattern, while another provider may use a different key or method.

The method works best when the model has freedom in wording. Creative writing offers many choices. Short factual answers and code don’t have that many options. Heavy rewriting can also weaken the signal, which is why watermark stealing(deducing patterns from text) is an active research area.

If widely adopted, text watermarking could make AI involvement easier to verify later. I am assuming this will be heavily used in academia and school work. A watermark can indicate that a compatible model likely contributed to the text, without identifying the person or chat behind it.

If you liked the post, Share it with your friends!

Leave a Comment

Your email address will not be published. Required fields are marked *

1 2 3 … 142
Copyright © 2026 Amaldev. All Rights Reserved.