A Practical Look at Lee Smoke And Pickles

I've been running this process in my workflow for a while now, mostly because it solves a specific problem with audio file handling that a lot of other tools fumble. The basic idea is straightforward: you feed it a batch of files, it cleans them up using a proprietary algorithm, and spits out something more usable for editing or distribution. That said, the setup has some quirks that aren't obvious until you've hit them head-on. The installation itself is simple enough. You download the package from their site, run the installer, and after a few minutes you're greeted with a dashboard that's not particularly pretty but does everything it needs to. The license registration is the part where people tend to get stuck. You need an account on their forum to generate a key, and the activation server occasionally times out. If that happens, just wait ten minutes and try again. It usually goes through.

Lee Smoke And Pickles Setup And Workflow

Once you're past activation, you'll want to configure your input and output directories before importing anything. I learned this the hard way on my first session. I pointed it at a folder full of WAV stems and tried to process them all at once. The application hung, then crashed, and I lost the progress on about forty files. The workaround is simple: batch them in groups of ten or fewer. It also helps to make sure your source files are all in the same sample rate. Mixing 44.1 and 48kHz files in the same job will cause silent drops and click artifacts that are a pain to track down later. The processing engine uses a hybrid approach. It applies spectral denoising first, then time-stretches or resamples depending on your settings, and finally normalizes to your target LUFS. The default settings are conservative, which means they sound safe but often leave too much noise floor for broadcast or film work. I usually bump the denoise threshold up to around -42dB and set the normalization target to -14 LUFS integrated. That gets you to streaming standards without sounding crushed. The output files come out as WAV by default, and you can change that to MP3 or AAC if you're just doing quick edits.

Where It Falls Short

It's not a magic bullet. The tool struggles with highly transient material like acoustic drums or cymbals. The denoising tends to smear transients if you push it too far, and there's no way around that. You can run a light pass on the noise floor first, then go back in for a second lighter pass, but even that only goes so far. I've found that dedicating a separate track or bus for noise reduction on drum sources works better than trying to do everything in one pass. Another limitation is that the software doesn't handle metadata very well. ID3 tags, embedart, and ISRC codes all get stripped during processing. If you're working with music that needs to stay tagged for distribution, you'll need to re-apply metadata after export. I keep a CSV reference sheet for the projects I run through it regularly. It saves time compared to rebuilding the metadata from scratch.

Get the Full Details

Smoke and Pickles by Edward Lee | Hachette Book Group
Smoke and Pickles by Edward Lee | Hachette Book Group

Alternatives To Consider

If your work involves a lot of dialogue or speech, you might be better off looking at iZotope RX or Acon Digital's DeNoise. Those are purpose-built for voice and tend to preserve clarity better. Lee Smoke And Pickles is more suited to musicians and producers who need batch processing with a single tool rather than piecework with multiple applications. For that use case, it holds its own, especially at the price point. The main thing is to go in knowing what it does well and where it breaks down, then design your workflow around those edges instead of fighting them.