## What does RMS error mean?

Share your tips and tricks here or get help with any Autopano Pro / Giga problem!
No bug reports (of any kind) in this forum!

macneubie
New member

Topic author
Posts: 2
Likes: 0 post
Liked in: 0 post
Joined: Tue Mar 11, 2014 4:45 pm
Info

### What does RMS error mean?

Newbie reporting here, hence be kind to me

As above, what does it mean? Higher numbers means a less than perfect stitch, lower is better? What are the typical values that determine if the stitch is good or bad? Thanks!

DrSlony
Moderator

Posts: 1892
Likes: 0 post
Liked in: 0 post
Joined: Sat Nov 03, 2007 6:30 pm
Location: Sweden
Info

### Re: What does RMS error mean?

Hey

Yes, lower=better, higher=worse.
Assuming Autopano's RMS refers to the mathematical root mean square, Wikipedia explains it: http://en.wikipedia.org/wiki/Root_mean_square

AlexandreJ/devs, could you confirm that this is the same RMS?

AlexandreJ
Kolor Team

Posts: 5988
Likes: 7 posts
Liked in: 10 posts
Joined: Mon Nov 14, 2005 4:56 pm
Location: Francin, France
Info

### Re: What does RMS error mean?

A tutorial has been written on that : http://www.autopano.net/wiki-en/action/ ... _RMS_value

DrSlony
Moderator

Posts: 1892
Likes: 0 post
Liked in: 0 post
Joined: Sat Nov 03, 2007 6:30 pm
Location: Sweden
Info

### Re: What does RMS error mean?

That's a great explanation, thank you!

klausesser
Member

Posts: 8830
Likes: 5 posts
Liked in: 63 posts
Joined: Mon May 22, 2006 12:18 am
Location: Duesseldorf, Germany
Info

### Re: What does RMS error mean?

macneubie wrote:Newbie reporting here, hence be kind to me

As above, what does it mean? Higher numbers means a less than perfect stitch, lower is better? What are the typical values that determine if the stitch is good or bad? Thanks!

I can´t tell about the maths and the theory. But i can tell a lot about practise:

When you shoot hires indoors - which means about 700MPx to 6GPx for a sphere - you will realize a *relevant* difference between RMS of 2 and RMS of 1 for exemple.

Usually invisble stitch-errors suddenly become visible when you zoom in @100% . . . which is the goal of shooting hires, is it.

The difference of 2 RMS and 1 RMS and lower can mean to edit the result for hours in Photoshop for searching and retouching them . . or achieving an almost perfect stitch where you NOT need to search for and edit stitch-errors.

Working commercially this is a relevant point i mean . .

best, Klaus

Jeremiahscott
Member

Posts: 93
Likes: 7 posts
Liked in: 1 post
Joined: Fri Apr 05, 2013 4:07 am
Location: Washington
Info

### Re: What does RMS error mean?

Klaus, I'm curious. I've seen you mention this RMS of 1 before. I've played with this a bit, and while I can get down very close to an RMS of one, (and while I completely agree with you about the many hours of photoshop time being wasted with a high RMS number), I also end up losing many of the control points to get down to an RMS of 1. if I can get 2000-3000 control points and an RMS of around 2 I'm quite happy. while I can spend another 30 minutes getting the RMS down to 1, I end up losing at least half of those control points, and then seeing errors in other places . there has to be a trade off between the "perfect RMS" number and a good spread of control points..... right?

mediavets
Moderator

Posts: 16414
Likes: 2 posts
Liked in: 130 posts
Joined: Wed Nov 14, 2007 2:12 pm
Location: Isleham, Cambridgeshire, UK.
Info

### Re: What does RMS error mean?

Jeremiahscott wrote:Klaus, I'm curious. I've seen you mention this RMS of 1 before. I've played with this a bit, and while I can get down very close to an RMS of one, (and while I completely agree with you about the many hours of photoshop time being wasted with a high RMS number), I also end up losing many of the control points to get down to an RMS of 1. if I can get 2000-3000 control points and an RMS of around 2 I'm quite happy. while I can spend another 30 minutes getting the RMS down to 1, I end up losing at least half of those control points, and then seeing errors in other places . there has to be a trade off between the "perfect RMS" number and a good spread of control points..... right?

Yes, see:

viewtopic.php?f=2&t=21772&p=140884#p139330

klausesser
Member

Posts: 8830
Likes: 5 posts
Liked in: 63 posts
Joined: Mon May 22, 2006 12:18 am
Location: Duesseldorf, Germany
Info

### Re: What does RMS error mean?

mediavets wrote:
Jeremiahscott wrote:Klaus, I'm curious. I've seen you mention this RMS of 1 before. I've played with this a bit, and while I can get down very close to an RMS of one, (and while I completely agree with you about the many hours of photoshop time being wasted with a high RMS number), I also end up losing many of the control points to get down to an RMS of 1. if I can get 2000-3000 control points and an RMS of around 2 I'm quite happy. while I can spend another 30 minutes getting the RMS down to 1, I end up losing at least half of those control points, and then seeing errors in other places . there has to be a trade off between the "perfect RMS" number and a good spread of control points..... right?

Yes, see:

viewtopic.php?f=2&t=21772&p=140884#p139330

What i said is a - my - "rule of thumb". I - i - realized that having max. around RMS 1 leads to the best results.

Usually the results are not related direcly to the number of CPs. You do not need to have thousands of CPs between the images. You only need the right ones.
Optimizing in APG can be be a kind of art sometimes . .

I could not reveal how to manage to reach a number on RMS1 in APG 3.5 . . . . the best i achieve is around 2. And that´s not perfect with 700MPx when you zoom in.
So at the moment i use 3.0.8 as long as i´m a bit more familiar with 3.5 and find how to achieve around a RMS of 1.
3.5 maybe providing better automatic quality - i don´t care - but going deeper in it working manually it starts just confusing me. Documentation traditionally is . . . well.

best, Klaus

### Who is online

Users browsing this forum: No registered users and 3 guests