Repository files navigation

BitcoinUtxoVisualizer (shortbuv) can generated videos of the evolution of the evolution of Bitcoin'sUTXO (Unspent Transaction Outputs).

Watch the video on Youtube:

WARNING: Generating such video is a time & resource intensive task, as Bitcoin's database is continuously growing.

This currently only works in Linux. Prerequisites are a C++ compilerg++ (>= v9) (or, my prefered choice,clang++), CMake (>= 3.13), and OpenCV (libopencv-dev).

1. fetch

   git clone --recurse-submodules https://github.com/martinus/BitcoinUtxoVisualizer.git

2. compile

   mkdir BitcoinUtxoVisualizer/buildcd BitcoinUtxoVisualizer/buildcmake -DCMAKE_BUILD_TYPE=Release ..make -j12

3. Run all tests, should printSUCCESS!

   ./buv

This is a 3 step process:

1. Have a fully syncedBitcoin Core node running locally.
2. Make sure to enable transaction index by addingtxindex=1 tobitcoin.conf.
3. buv makes heavy use of Bitcoin Core's JSON RPC, so you need to enable this as well. Also, make sure theRPCs have enough threads for processing. To sum this up, I have these settings in mybitcoin.conf file:

   server=1rest=1rpcport=8332rpcthreads=12rpcworkqueue=24txindex=1dbcache=2000# generate username & password with 'bitcoin/share/rpcauth.py <username> -'rpcauth=martinus:xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Once Bitcoin Core is fully synced and RPC is enabled, you can preprocess the UTXO database. This fetches all blocks with full transaction data from bitcoin core, extracts UTXO data, and writes a compact data file. First, configure by editingbuv.json. For this step, you only need to updatebitcoinRpcUrl and the outputblkFile:

"bitcoinRpcUrl": "http://127.0.0.1:8332","blkFile": "/run/media/martinus/big/bitcoin/BitcoinUtxoVisualizer/changes.blk1",

The outputblkFile will be ~7.5GB large (as of Block 660,000). It contains block information & all satoshi amounts that were added or removed for each block. The format is tuned to be very compact and very fast to parse.

./buv -ns -tc=utxo_to_change -cfg=../buv.json

On my computer this takes about 1 1/2 hours, saturates 12 cores, and takes ~6.5GB of RAM. I have spent a long time to speed this up, initially this took 4 days and >30GB of RAM.

After generating theblkFile, this file can be converted into an image stream that is directly piped intoffmpeg to generates a video. Preview is possible withffplay.

The configuration filebuv.json has several options to configure the output.

To watch a preview, I usually updatebuv.json to start at a reasonably late block:

"startShowAtBlockHeight": 200000,

1. In one window, startffplay:

   ffplay -f rawvideo -pixel_format rgb24 -video_size 3840x2160 -framerate 60 -i "tcp://127.0.0.1:12987?listen"

2. In another window, startbuv to connect toffplay and pipe its output into it.

   ./buv -ns -tc=visualizer -cfg=../buv.json

   Oncebuv has processed up to block 200000 ffplay will pop up and show a life preview.

If you are happy with what you see, instead offfplay useffmpeg and startbuv again:

ffmpeg -f rawvideo -pixel_format rgb24 -video_size 3840x2160 -framerate 60 -i "tcp://127.0.0.1:12987?listen" -c:v libx264 -profile:v high -bf 2 -g 30 -preset slower -crf 24 -pix_fmt yuv420p -movflags faststart out.mp4

For 660000 this will create a ~3 hour 4K x 60Hz video, where each frame represents a single block. The video is about 21GB large.

Here is the final image of that video. Click for high resolution 4k image: