---
title: "ProTrackR vs ProTrackR2"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{ProTrackR vs ProTrackR2}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

## ProTracker2 and its predecessor

The predecessor to this package ([ProTrackR](https://pepijn-devries.github.io/ProTrackR/)) was
entirely programmed in R. Although technically possible, it was challenging and slow with recursive algorithms.
The new version is a complete overhaul in C/C++, based on Olav Sørensen's
[ProTracker clone](https://16-bits.org/pt2.php). With it, came some design changes which are worth mentioning
here.

## Overview of differences

The table below summarises the differences between ProTrackR2 and its predecessor.
<!-- TODO explain each of these differences in more detail -->

```{r tab-dif, echo=FALSE, message=FALSE, warning=FALSE}
library(kableExtra)
data.frame(
  Feature    = c("Effect commands", "Infrastructure", "File readers",
                 "Audio output", "OpenMPT test cases"),
  ProTrackR  = c("Limited set implemented", "R script", "Optimized for format preservation",
                 "tuneR S4 Wave class", "Passes 6 out of 12 selected tests"),
  ProTrackR2 = c("All PT2.3d effects implemented", "Compiled C/C++", "Optimized for PT2.3d compatibility",
                 "audio S3 audioSample class", "Not tested yet")
) |>
  kbl()
```

## Benchmark

A benchmark test where the same module (the one provided with this package) is rendered with
both `ProTrackR` and `ProTrackR2`. The settings for both tests were similar and performed
on the same system and repeated 10 times. On average `ProTrackR2` renders 8.8 times faster
than `ProTrackR`.