Where is ALL the NEEDED/USEFUL information found on the web?

Hey all - Beau here!

QUICK BACKSTORY: I’m a graphic designer and pro photographer (I’m actually The WatchTographer & PIX-EDC) so I have seen my share of watch faces. :grinning: and even designed a few. I recently purchased a GW4 Classic and started transferring some of my face designs from photoshop to WFS. Sweet.

So, here are my questions.

  • Where is the most complete list of TAGS and CODES that we can use in WFS - and where in the program, are those editable fields? I see people referencing and advising what codes to use in certain conditions, but where did they, find that information in the first place?

  • I understand that the best size for ‘full-face’ images in WFS, is 450x450. But, when designing those images, is it best to create them at that size, or make them much larger, and then re-size down to 450?
    • Does this apply to BITMAP fonts as well?

  • How do you include multiple ‘faces’ in a single design, that are not just a different ‘style’ (AKA a separate image) selected in the STYLE > ADD STYLE section - but rather a separate face (group) altogether?
    EXAMPLE: I have a face that is a BG with a semi-transparent PNG above it (that uses the GYRO function) to create the illusion of light moving on the face, placed in a GROUP. How do I tell the SW that I have more groups (set up in this manner) that I want to use as a selectable STYLE within the CUSTOMIZE function of the watch? As you can with COLORS or HANDS.

Maybe @r.liechty_SDR can point this n00b in the proper direction?

FYI, be on the lookout for more questions to come!

Thanks in advance!

Hello,

Here is a Code lab for WFS tag expression. You can also see the https://developer.android.com/training/wearables/wfs/tag-expression site. You can also visit the Tag Expression | Samsung Developers site which is for GWS.

You can check this thread response Battery Optimization in Watch Faces - #2 by sinjae

Welcome to Watch Face Designing Warning this can become too much fun.

This currently is not possible there is a hint of it being possible in a future release but I may just be wishing for it. :slight_smile:

Be sure to read the training documentation on the Android wearable site HERE but ask questions some of the training needs explaining.

Ron
Samsung Developer Relations

1 Like

You can check this thread response Battery Optimization in Watch Faces - #2 by sinjae

Thank you for the replay, but you basically directed me to an answer that is a non-answer. I know that 450x450 is ideal, but I have seen that either designing images at 450 or converting them to 450, GREATLY reduces the clarity of the result. This is especially true in regard to text or numbers.

So, maybe this is the better question: How LARGE, can we get away with and NOT lose image quality, while maintaining a decent file size? Also, I want to keep processing use to a minimum as well. It is maybe 900px? Just doubling the destination size?

@r.liechty_SDR, any thoughts?

Thanks.

This may be a question to ask on a photoshop discussion.
I know on GWS the theory for simple images was to make it small and then increase the size in GWS to save the size of the binary. But if you have a detailed image I just have no idea and with multiple watches with various displays no one answer may be best.

I can ask a designer that works for Samsung but not until after Oct 12th as he is busy working on SDC22 at the moment.

Ron
Samsung Developer Relations

1 Like

Maybe you should make some tests on the watch. In theory I would expect for static images like backgrounds it should be least computing, if you insert finished image that has same pixel count as the display. If you put in anything else, you force the watch to compute the resized version for its display and also rely on its internal algorithm to make it satisfyingly well (sharp enough etc.).

Theoretically, the best way for both quality and performance (high performance = minimum processing = low power consumption) is to use images of exactly the same size as displayed on the display. However, this is impossible because the display sizes of consumer devices are different. WFS provides a 450x450 canvas according to the display size of the most representative Galaxy Watch4/5, so the next best way is to produce an image to import according to the size of 450x450.

More tips for parts that require sophisticated quality,

  1. Use 1:1 size image as mentioned above
  2. Do not use lines with 1 pixel in text or line
  3. Do not use decimal coordinates if possible

Among the resolutions of the currently released WearOS devices, 450x450 is quite large, so you don’t have to worry about quality loss due to upscale when producing the original image based on 450x450. Conversely, in a down sampling situation occurring in a device with a small resolution, it is generally difficult to expect a serious quality regression, but it is necessary to consider the 2nd tip above.

If extreme sophistication is not required, it is also okay to use images of a smaller size. The results of GWS made for old Galaxy Watches (Tizen OS) with 360x360 resolution are converted and used a lot on Galaxy Watch with 450x450 display, but no serious quality problems have been reported.

4 Likes

That would be EXCELLENT! Thank you!

As a designer, I ALWAYS design large and downsample - it just provides better results.

Also, it would be handy to know just HOW larger/smaller images are displayed on the screens. Are they just ‘blown up’ or 'reduced? Or are they actually resampled by the watch once installed, or during the build process?

  • Beau

@PIX-EDC

I think you have valid questions that others are curious about. Graphic manipulations is just done with automagic to me so I can’t answer.

Sinjae is a Samsung Engineer and he responded above did he answer what you asked or does it need clarification. Do you have further questions.

Ron
Samsung Developer Relations

2 Likes

I (as not a designer, but a perfection-enthusiastOR), always design everything in vector then export it whatever resolution is needed. :slight_smile:

1 Like

Agreed. I vector all but the textures, then downsize.