Archiver, 7z Alternative and FreeArc Replacement

[Registered55]

Hi everyone,

I've been experimenting with archiver.exe and testing out various multi-codec chains (combining precomp sub-codecs, srep, and zstd), but I'm finding it a bit tricky to dial in the optimal syntax without official example scripts.

Since documentation on command-line configurations is sparse, would any of the experts or the developer be willing to share a few working example command lines?

Specifically, I'm looking for two or three profiles:

Normal / Balanced Profile: A clean, reasonably fast command for everyday folder archiving.

High-Compression / Max Profile: A heavy archival command (using precomp/rep/zstd or alternative internal codecs) pushed to its limits for maximum size reduction.

Extraction Command: The matching clean extraction syntax for those specific profiles.

Having a small cheat sheet of tested examples on the forum would be immensely helpful.

for context, i'm currently testing on PS3 ISO's Decrypted.

[panker1992]

Generally i am trying to understand what is it that you are trying to do ??

if you want to max compression why would you use ZSTD?

precompression is there to help you pack better, so if you precompress kraken and leviathan streams you need equivalent or better to go for more

therefore ZSTD for max compression is a no go,

you can use rep or addex in my archiver both are good deduplications engines self made.

rep is a sliding window, symmetric deduplication engine which uses the same ram to compress/decompress and its very very fast and efficient.

Addex will be better to gather reps that are scattered i used logic that came from Srep to make this happen.

of you want to compress max

xz:l9:d256mb:t10 will give you the best lz77 family can get you

if you want to go oodle, then you can do

oodle:l9:a13:v1:t10:d256mb,
this will make leviathan scoop the input because v1 uses previous block as a dictionary for the next. using :v0 will half the ram

zstd:l22:w31:f1:s9 will give you the best ratio zstd can offer, remember zstd is using Entropy + tANS so it will take 2 hours to compress and will give you worse ratio than xz

if you are interested in details give me a PM and i will help you more

[Joe Forster/STA]

Originally Posted by panker1992

if you are interested in details give me a PM and i will help you more

I think everyone is interested in the details; that's what this forum is for.

[panker1992]

Details vary per request,

What details do you require??

Do you want to maximize ratio? or get the best possible speed??

Do you have compress once decompress many model ?? like repackers where compressing once takes multiple hours but the unpack will happen to millions of PCs?

Do you want RAM oriented performance ??

how am i supposed to provide detailed guidance in any case.


There is something else as well.

Archiver has Build-in Compressors you are highly familiar with for example lzma2,xz,zstd and more. i didnt change their params to something else so if you know how to use zstd in FA then you can use it on my tool as well in the same way.


Details for my own algorithms is a welcoming question


I have forked xtool 7.9 and i have added native code for packmp3,packjpg,wavpack, and my own algo for OGG, called oggpre.

the same goes for addex, and rep.

So the request for details must at least include a topic.

[Registered55]

Thanks for the guidance! To give you the exact context of what I'm aiming for:

My use case is "compress once, store forever" (archiving game ISOs like PlayStation 3 dumps and heavy media libraries onto a home NAS). Compression time and CPU/RAM load don't matter at all—I have a 32GB system and I want to squeeze out the absolute maximum space savings.

Once I lock down the ideal settings, my ultimate goal is to wrap this into a PowerShell batch script so the whole pipeline runs completely hands-off. I've experimented with MiniCompressor and DiskSpan 2 in the past, but getting them to cooperate reliably in unattended batch modes proved difficult. Because your archiver.exe is built from the ground up as a standalone entity, it opens up fantastic automation options.

Before I write the wrapper script, I'd love to clarify a couple of quick technical details on how your native algorithms behave:

addex vs. rep: Are both addex and rep fully baked natively inside archiver.exe, or do they rely on external binaries? Furthermore, is addex essentially the advanced evolution of rep for scattered matches (using SREP-inspired logic, as you mentioned)?

Parameter Syntax / .ini behavior: When I tried passing -mrep:m3 directly via the command line, the -m3 parameter seemed to be ignored, and I had to manually edit the .ini file to force it from level 1 to level 3. Is there a specific syntax I should be using in the command string to pass parameters directly to rep or addex without touching the .ini file?

More than anything, seeing a couple of working example command lines (e.g., a standard balanced profile and a max-compression profile using XZ or Oodle combined with your native dedup) would be immensely helpful. It's less about copy-pasting the strings directly, and more about seeing the exact syntax grammar, parameter formatting, and delimiter rules for how your tool's parser expects everything to be structured.

If you can share a couple of reference examples along those lines, I'd love to test them out and start building the PowerShell automation script!

so far i have been building something like this

archiver.exe a "R:\PS3\TEST\bejeweled_3_archive.arc" "R:\PS3\TEST\bejeweled 3 [BLUS30865]" -mprecomp:creflate:czlib:cpng:clz4:czstd:cflac:cbrunsli:cpackjpg:cpackmp3:ckraken:b1024:t50p:d2:dd+srep+xz:l9:d512:t8 -b4096m --tempdir="C:\Archiver\temp"

but to be clear, i'm learning as i go and no idea if i'm on thr right track, but so far this is the latest string i have cobbled together lol.

another issue i have been having, i'm finding it difficult to get archiver to go above 35% of cpu usage, when it starts it jump to around 50-60% great, but as it gets to stage 2 and 3, the cpu goes down to 35% leaving a lot of CPU performance on the floor.

i have tried playing around with threads, but knowing where to put the -t options where and what number has so far alluded me.

one more thing is, let's say i wanted to run lolz (i have found this in your external archive) where would this be placed in the command string?