AAR: Similarity

Almost A Review: Find Similar Audio Files

      Over the years, I’ve built up a collection of music files on my computer. I suspect most people have, to some degree. And while I have tried to keep things organized, with no duplication, I know there are doubles in my collection.

      Finding 100% identical files is easy, there are a number of tools that do just that, regardless of files types. But the same audio can be packaged in a variety of ways. Differing formats, compression styles, bitrates, or something as simple as tags. Complicating things, song are often enough released (or re-released) by multiple groups over the years. Do I want the live performance, the studio recording, or both? Listening to each song, multiple times, can be a tedious process. Especially when I just want to enjoy the music.

      On and off over the years, I’ve looked for a tool that would automate this process. A tool that would find any song that I have multiples of, for whatever reason. While I’ve found some that sort of did the job, and a number that made the claim, I hadn’t seen anything that actually worked (at least well enough to use). But that has now changed.



Similarity

      While the name may not be the best, as it will show up in too many other situations, this program actually does the task.

      It will take a folder, or multiple folders, and compare all the audio files in them. The details of how it’s comparing are adjustable, but the default settings are good enough. When it’s done, it gives you a list of audio files that are ‘similar’. Then you can go from there.

      Trust the tool, and let it delete what it found. Or listen to them, and decide for yourself. The latter is what I do, as sometimes the reported doubles aren’t quiet doubles. For example, I have some songs with voices, and then the voiceless instrumental version of the same song. Similarity decides those are close enough to be duplicates. Which does make sense, even if I still want both version.



Buy Now (or not)

      There is a ‘premium‘ version, which I have not tested. The ‘free‘ version works well, and was able to quickly find a number of duplicates in my collection. As that is the primary function I am after, and it does it well enough, I saw no reason to purchase. Perhaps that will change at some point in the future.

      Of the claimed benefits to ‘premium’, only three seem to have any value.

You would gain access to a more precise comparison method. The ‘free’ version applies this to a small fraction of the songs it scans, and it does appears to more reliably find duplicates. But if you’re on a budget, ‘good enough’ overrules ‘better’.

You would be able to save the results of the scanning. I do other things with my computer that mitigate the benefit here. But for the average person, I could see this being a useful feature to have.

Supports future development. Which in turn, would lead to improvements to the ‘free’ version as well. When you boil down to it, this seems to be the primary reason to buy software, any software.

      There are other features gained, but they seem to be focused on particular niches. Useful for some, but not everyone. I listen to music, looking at a ‘sonogram‘ is of no use to me.



Other Features

      Similarity will check images too, and looks to do a decent enough job of this. There are already plenty of other tools that will check images, and this program doesn’t particularly stand out in how this task is handled. Not done poorly, I just think there are better ways to handle the problem of duplicate images. Having said that, there is some value in using one tool to solve two problems.

      In addition to duplicates, it will ‘Analyze’ audio files. This rates various aspects of the audio quality. Bitrate, clicks, channels, size, spectrum, etc. Then aggregates to provide an overall rating. On it’s own, not of much value, but it could be a starting point towards finding better versions of audio you already have.

      And that seems to be it. This program is focused on a job, and does that job well. No need to add additional ‘fluff’.





      In case you missed any of the links above: Similarity can be found at similarityapp.com

Leave a Reply