Author Topic: Y-axis scaling issue when displaying grade profile (Read 8341 times)

guney · « **on:** March 25, 2008, 01:26:26 PM »

I have just downloaded and installed the demo version of TopoFusion Pro (v3.31). By the way, I've been looking for a tool just like this for weeks! Thank you for the excellent job you've done on this piece of software. I'm sure I'll be purchasing a full license soon.

There is, however, one glitch that I've noticed: In the Profile dialog, when I set the Y axis to display the grade, the scale of the Y axis is not adjusted to fit in the highest and lowest points in the dataset. Consequently, some parts of the graph are cut off.

Here's an example:

In that screenshot, the cursor is at a point on the graph where the grade is shown as 16.7%, but the Y axis does not seem to go above 10%.

It also works the same way in the opposite extreme:

In this second image, the cursor is at a point where the grade is -22.6%, but the Y axis bottoms out around -10%.

I'm not sure if this is happening because the Y axis is simply hard-coded to display a 10% to -10% range at all times, or whether an auto-scaling computation is failing for some reason. If it's the former (with some kind of reasoning like "if the extremes are fully included, the average parts of the plot would be too flattened to show enough variation"), I hope I'm not alone in thinking that having something like a scaling control for the Y axis would be welcome flexibility, allowing a more inclusive picture when desired.

I've also seen the same problem when plotting "pace" in the same dialog. It seems to work fine for speed and elevation, though I don't know if that's only due to the fact that my data happens to fall within any hard-coded boundaries in either of those cases. (I haven't tried the other variables; my dataset doesn't have those.)

Is this a bug? A known limitation? Or am I missing something?

Krein · « **Reply #1 on:** March 25, 2008, 01:42:21 PM »

Hi,

Thanks for the feedback, and glad you're digging the software.

You are right that the grade profiles are cut off at 10% and -10%. It's not really a bug, but an arbitrary limit. It should in minimum allow a larger range (likewise with pace, and power).

Speed you can control the high point (albeit not in an easy-to-access way) using the "minimum speed allowed" setting in Options->Preferences.

What I envision, for the next version of TF, is user selectable high and low points for all types of data. These would apply both to the profile limits and the track shading on the map. It's a sorely needed addition and will be made a priority.

guney · « **Reply #2 on:** March 25, 2008, 08:43:12 PM »

I don't want to sound presumptuous, but am I wrong in thinking that it might be more intuitive and straightforward if the Y-axis value range is simply calculated on the fly to span the min-to-max range of the dataset at hand? It seems like this is what's being done (at least for the "max") in case of plotting elevation profiles.

Krein · « **Reply #3 on:** March 25, 2008, 08:46:45 PM »

Generally you're right, but not so much in the case of grade. It's very easy to get ridiculous grade numbers such that it's quite useful to have a cap.

But the current limit is too small, and if you've got good data, you want to see it all.

guney · « **Reply #4 on:** March 25, 2008, 09:00:46 PM »

Makes sense. Thanks for the quick responses. I'll look forward to the next version of TopoFusion to be able to see my grade plots more completely...

'>

guney · « **Reply #5 on:** March 26, 2008, 07:34:39 PM »

You know what? I just came across another example that goes against what we just discussed. Take a look at this grade plot:

In this screenshot, the max value on the Y axis is, again, 10%. (It seems to be cut off due to window size, but no amount of vertical resizing expands the Y axis range in that direction. Only the labeling changes.) However, the min value on the Y axis is somewhere around 19%. In fact, this graph successfully accommodates the lower extreme of its dataset. Only the upper extreme gets cut off.

This example makes it look to me like an on-the-fly value range computation is being attempted but (at least for the max end in this example, and for both ends in the previous one) is failing. At the very least, it's not a hard-coded 10% to -10% range.

I'm not trying to prove a point.

'> Just wanted to prevent a potential misconclusion.

Krein · « **Reply #6 on:** March 31, 2008, 09:10:11 PM »

Gotcha. Just got back from a trip, but will look at this ASAP.

The way the profiler computes grade is a little funny. There's some smoothing in there, so that may be part of what's throwing it off.

Thanks for the post.

Krein · « **Reply #7 on:** April 10, 2008, 09:22:51 AM »

Hi,

The latest version clears up this issue. Thanks much for pointing it out.

The min/max was being set by the min/max of all loaded files, but it was their average uphill and downhill grade. Thus the limited but moveable range.

That was definitely a bug.

Now you'll get the full range for the currently profiled track. There's a tiny bit of smoothing right now. In the next version we'll have buttons to increase/decrease smoothing which should help quite a bit for grade plots.

Krein · « **Reply #8 on:** April 10, 2008, 09:23:38 AM »

Just wanted to add that this bug with the grade range was introduced in the last few versions -- older versions were doing it correctly, but when we upgraded the profiler to do HR/Cad/Power it got fouled up.

guney · « **Reply #9 on:** April 11, 2008, 04:22:29 PM »

Thanks! And I'm glad to be of help.