V2 Sparse: Zstd versus Uncompressed

[bwinter]

When using FPP Connect to upload to FPP, the FSEQ Type options are

• V1
  V2
  V2 Sparse/zstd
  V2 Sparse/Uncompressed

What's the difference between zstd and uncompressed, and what's the deciding factor on which to use?

[dkulp]

In general, use zstd.   The files will be about 1/10th the size so transfer a lot faster. 

The uncompressed might be needed for various low power devices that don't have the CPU speed to decompress the zstd OR don't even have a zstd library at all.  For next year, we're kind of working with other device folks to provide a way for FPP connect to upload sequences to other types of devices, but not all of them are as powerful as a Pi/BBB and thus may need the uncompressed.   

The PiZero and very old Pi's also might consider the uncompressed.   I'm still on the fence on this.   For my testing of using a hat to drive some pixels, the zstd is fine.   But I have some reports where there have been some lagging and I'm not sure if it's a config issue or not enough CPU for zstd or crappy SD card (where uncompressed would make it worse) or what.   

[bwinter]

Thanks for the clarification.

I use a lot of Zeros (and 3s) with Hats, so I'll keep that in mind if I start seeing any lag with the PiZero.

[obejohnknobe]

Beginner with  XLights. 
1.  No idea how to post here just grabbed this one and guessed
2.  It appears I have everything setup on the pi and XLights but no idea where to find the problem with 2 - P5 panels 32 vercal and 64 wide 1/16 scan.  When doing the test lights from the FPP, I get all the colors needed with some glitches.  The glitches have 8 vertical strings and 1 panel wide will blank but the next cycle it will appear.  All random when it goes out.  When applying a butterfly sequence, the first 8 top vertical strings across both panels is doing what is programmed but the next 8 vertical strings below on both panels wide look like they are going the opposite way.  This does this for the 4 sets of 8 vertical strings.

My guess is there is something I did not check or wrong info.  I have looked in the manual  tried YouTube videos.  No idea where to begin other than changing items in both FPP and XLights to see what it will do.

[plaberge]

The PiZero and very old Pi's also might consider the uncompressed.   I'm still on the fence on this.   For my testing of using a hat to drive some pixels, the zstd is fine.   But I have some reports where there have been some lagging and I'm not sure if it's a config issue or not enough CPU for zstd or crappy SD card (where uncompressed would make it worse) or what.   

I use a few PiZeros in my more remote self-contained props and have not seen any noticeable lag with zstd. I did upload uncompressed while debugging - I couldn't see any differences between compressed and uncompressed. Given these PiZeros drive only a few channels each, uploading zstd files is lightning fast. I'd tolerate a few milliseconds delay just to take advantage of that

[jnealand]

Given these PiZeros drive only a few channels each, uploading zstd files is lightning fast. I'd tolerate a few milliseconds delay just to take advantage of that

Me too.

[bwinter]

To clarify?does the ?decompressing? only occur once, after the file is newly-uploaded?  Or does it actively decompress as the sequence is running (each time)?

[allknowing2012]

The file sits on the sd card as is -- so compressed. I assume it just reads it back in the zipped format. There are code libraries that allow you to read a zipped file without unzipping.