I’ve had multiple emails now from different people working at US government agencies – Agisoft Metashape is no longer available nor allowed because it’s made by a Russian company. Below, I’ll provide a run through of Reality Capture as it stands today. The main competitor to Metashape is Reality Capture, now owned by Epic Games. ... Continue Reading →
‘[New] Outlook’ – everything that’s wrong with modern software
[FYI - This post is mainly just for me to keep track of my venting] The time has finally come - Mail and Calendar, my favourite email application in a long time, has now been completely replaced by 'new' Outlook, and my attempts to revert are met by an instant reinstallation of the naff new... Continue Reading →
[Academic Tech] Philips 40″ 5K2K Ultrawide Thunderbolt monitor
A shortish review here about the Monitor I've been using at work for about 6 months now. It's the Philips 40" curved monitor: I needed a new monitor to go with the new computers my team and I got set up with, and I very specifically wanted a very high-resolution monitor for multi-tasking. I wasn't... Continue Reading →
Dabbling with Linux again. And quickly giving up… again.
I tried installing linux again. Got frustrated, went back to windows.
Focus stacking and photogrammetry
I experimented with macro photogrammetry using a small fossil and a Nikon Z8 camera with MC105 VR S lens. I used focus-shift and focus-stack with shallow depth of field. Reality Capture produced a good model from the focus-stacked images, but using the lens at f51 also yielded a great result. Both methods have their benefits.
An Excellent Free and Open Source focus-stacking solution
I struggled with merging focus-shifted images using Photoshop and Affinity Photo but found success with Helicon Focus, though with some limitations. I then discovered focus-stack on GitHub, a simple command-line tool, which was comparable to Helicon Focus.
My experience training a local LLM (AI chatbot) on local data…
The user encountered challenges while attempting to use various methods to feed information into local Large Language Models (LLMs) via RAG (Retrieval-Augmented Generation). They explored methods such as Nvidia Chat with RTX, Ollama with Python scripts, and Ollama with Open-webui. Results varied, with some methods providing inaccurate or incomplete outputs. Comparatively, Microsoft Co-pilot, running GPT4-Turbo, significantly outperformed the local methods.
[Academic Tech] JBL Tour One M2 Noise Cancelling Headphones
*Note, I was sent these headphones for review by JBL nearly a year ago, but they've had precisely zero input into this review (obviously, given I haven't written it until now), and it doesn't affect what I'm writing at all. The only benefit I stand to gain is the amazon affiliate link I'll shove at... Continue Reading →
Reality Capture going free for everyone*
The other day, I saw that Capturing reality announced a new pricing plan for their photogrammetry software Reality Capture: https://twitter.com/peterfalkingham/status/1767573164614103112?s=20 Reality Captured wormed it's way into my heart as my favourite photogrammetry software. Initially, while I was blown away by the speed of Reality Capture, I had a bit of a rough first impression due... Continue Reading →
Beholder Vision Desktop
I reviewed Beholder Vision back in 2022, and then it was a web-based service, where you upload your photos, it produces a model, and you then download it. That service is still available, but Beholder have now released a desktop version. Best of all, it’s complete free for personal use. It’s technically ad-supported, though so... Continue Reading →