Login at Kodi Home

wesk05 · (This post was last modified: 2016-01-11, 19:33 by wesk05.)

(2016-01-09, 02:23)hdmkv Wrote:
(2016-01-09, 01:55)wesk05 Wrote: Once I get back from CES, I will upload a modified 3D MVC ISO monoscope test pattern that you can use to check the TV Lines of resolution. This would be display dependent, also varies between active/passive 3D, but can be used to compare devices connected to the same display.

Excellent, thanks.

I have run into some problems with the monoscope pattern. I cannot seem to get the optimal 3D conversion for the pattern.

On my way back, I got to research on 3D resolution and chanced upon two excellent articles by Dr. Raymond Soneira (DisplayMate) and Dr. Martin Banks & Joohwan Kim (UC. Berkley). Back in 2012, when LG first introduced passive 3D displays, Samsung was debating and advertising how passive 3D only has half-resolution, produces inferior 3D effect and such. LG had complained to the National Advertising Division of BBB about this and the NAD actually ruled in LG's favor. NAD had drawn their conclusion based on these studies and another one from JKP.

Dr. Banks' study used stereoscopic visual acuity charts. Whereas, Dr. Soneira's study used the Blu-ray of IMAX: Space Station 3D. The study checked the sharpness and resolution of small texts that are present throughout the documentary. I think we can use the same methodology to differentiate between FHD3D and TAB/SBS 3D output. Although Dr. Soneira's study found that passive 3D displays produced sharper text, I didn't get that same result on my passive 3D Sony TV. I found the texts to be sharper and more easier to read on my active 3D TVs/projector.

Here are the links to the studies:

1) Joohwan Kim & Martin Banks - Effective spatial resolution of temporally and spatially interlaced stereo 3D televisions

2) Raymond Soneira - 3D TV Display Technology Shoot-out

3) Joe Kane - 3D Display Resolution

I have uploaded a clip from IMAX: Space Station 3D that has several of the texts that Dr. Soneira had used in his study- IMAX: Space Station 3D Test Sample Clip (1.1GB).

The clip is from 00:29:00 to 00:32:00. Here are the location of the texts to look out for in the clip.

I have checked the clip on RPi2, Zidoo X6 and Minix U1 (VidOn 14.2.6 & XBMC for Minix). Surprisingly, VidOn is the only player that cannot play this clip. I am not sure whether this has something to do with VidOn or Minix firmware or the clip itself.

noggin · (This post was last modified: 2016-01-11, 11:42 by noggin.)

(2016-01-11, 10:43)wesk05 Wrote: Dr. Banks study used stereoscopic visual acuity charts. Whereas, Dr. Soneira's study used the blu-ray of IMAX: Space Station 3D. The study checked the sharpness and resolution of small texts that are present throughout the documentary. I think we can use the same methodology to differentiate between FHD3D and TAB/SBS 3D output. Although Dr. Soneira's study found that passive 3D displays produced sharper text, I didn't get that same result on my passive 3D Sony TV. I found the texts to be sharper and more easier to read on my active 3D TVs/projector.

Do you have a 1080p or a 2160p Sony TV?

I assume the 1080p to 540p downconversion for each eye - which will be done in the TV (if the source presented is HSBS or Frame Packed 1080p or 720p, but probably not if HTAB 1080p) - can impact on this, as downconversion of TV signals is actually quite tricky to get right (it's more difficult than up conversion in some cases) This is because you want to retain as much vertical information as possible BUT avoid aliasing. This will mean different approaches will deliver differing levels of vertical resolution on a TV-by-TV basis.

**hdmkv** · 2016-01-11, 18:33

Thanks wesk05. Now I'm on a business trip this week so will check next week. I have that IMAX 3D disc. I'm pretty convinced already that VidOn's output via S95 is top/bottom as my PJ clearly indicates that. With Pi2 and HiMedia Q5, PJ shows "frame packing".

wesk05 · (This post was last modified: 2016-01-11, 19:43 by wesk05.)

(2016-01-11, 11:41)noggin Wrote: Do you have a 1080p or a 2160p Sony TV?
I assume the 1080p to 540p downconversion for each eye - which will be done in the TV (if the source presented is HSBS or Frame Packed 1080p or 720p, but probably not if HTAB 1080p) - can impact on this, as downconversion of TV signals is actually quite tricky to get right (it's more difficult than up conversion in some cases) This is because you want to retain as much vertical information as possible BUT avoid aliasing. This will mean different approaches will deliver differing levels of vertical resolution on a TV-by-TV basis.

The two Sony TVs that I have are both 1080p (one is passive and the other one is active 3D). Yes, indeed the algorithm used for sub-sampling differs between brands of TV. Stephen Dawson had done some extensive study on LG TVs.

Passive 3D - How does LG work now?

He has actually written quite a lot on 3D - http://hifi-writer.com/wpblog/?s=3D

wesk05 · (This post was last modified: 2016-01-11, 19:52 by wesk05.)

(2016-01-11, 18:33)hdmkv Wrote: Thanks wesk05. Now I'm on a business trip this week so will check next week. I have that IMAX 3D disc. I'm pretty convinced already that VidOn's output via S95 is top/bottom as my PJ clearly indicates that. With Pi2 and HiMedia Q5, PJ shows "frame packing".

Apparently, it is not that easy to judge 3D resolution just by visual inspection. You can get definite answers only from an HDMI protocol analyzer/waveform monitor/device log or a device that can show information about the 3D format like a projector/video processor.

Yes, we can say without any doubt that VidOn 3D output on Amlogic is not frame packing FHD3D.

movie78 · 2016-01-11, 20:32

(2016-01-11, 10:43)wesk05 Wrote: I have uploaded a clip from IMAX: Space Station 3D that has several of the texts that Dr. Soneira had used in his study- IMAX: Space Station 3D Test Sample Clip (1.1GB).

The clip is from 00:29:00 to 00:32:00. Here are the location of the texts to look out for in the clip.

I have checked the clip on RPi2, Zidoo X6 and Minix U1 (VidOn 14.2.6 & XBMC for Minix). Surprisingly, VidOn is the only player that cannot play this clip. I am not sure whether this has something to do with VidOn or Minix firmware or the clip itself.

That is some nice details testing, i will take look at the IMAX Space Station and do some comparison chat.

**hdmkv** · (This post was last modified: 2016-01-19, 01:23 by hdmkv.)

This post may be my last testing & report-out for Beelink Mini MX (at least on the Android side). There's minor progress on this box, for both its Android OS & custom OpenELEC OS, but overall, it has many significant issues. Not sure I want to invest any more precious time or energy on this thing. I've highlighted the key positives in green, and the key negatives in red...

With Stock Android OS: Latest OTA firmware LMY47V.20160105 test-keys installed

Kodi 16 Beta 3 is preinstalled
Kodi UI is 720p; ability to alter resolution is disabled (greyed-out)
Now handles up to 120Mbps bitrate clips (vs. previous 90Mbps) smoothly
Videos w/TrueHD = no audio!
Videos w/DTS-MA = downmixes to standard DTS5.1, or is passing just the DTS core
Videos w/LPCM 5.1/7.1 = still 2.0 output
3D MVC MKV's & 3D ISO's still play in 2D mode
Kodi froze & crashed often, especially with my TrueHD & LPCM test clips

With Custom OpenELEC: From Amlinuxmedia forum (tested with latest build as of January 16, but, now appears gone!?)

Kodi 16 Beta 5 build
Kodi UI is full 1080p (unlike above)
UI/performance is noticeably faster than on Android (not Chromebox-fast, but quite good)
Standard DD/DTS5.1 pass fine as DD/DTS5.1
TrueHD 5.1 & 7.1 passthrough (bitstream), which is great
ATMOS also passes as TrueHD 7.1 well (can't verify true ATMOS output as my AVR doesn't support it)
DD+ 5.1 & 7.1 = no audio!
DTS-MA 5.1 & 7.1 = crash, no picture; every time! (Could be due to last point below)
Tried a DTS:X movie ('Ex Machina (2015)') = picture blanks out (nothing, must reboot box)
3D MVC MKV plays in top/bottom mode. No audio! What's weird is the OSD (playback controls) are side-by-side. Changing 3D modes doesn't help, but using my projector to force TAB output does render a proper 3D pciture
3D ISO brings up "Play with 3D mode" option (as in Zidoo X6 & HiMedia H8), but same TAB output as above w/OSD SBS, and no audio!
Found that anything 23.976 (that's like 90% of videos, especially movies & TV shows, out there!!!) causes picture to blank out, with "No signal" reported by my projector, forcing reboot of box! I even tried a re-encoded 10GB movie, and same blank out issue

If I had any hope for the Mini MX, it would be under the custom OpenELEC work being done by whichever developer is behind it.

noggin · 2016-01-19, 00:31

(2016-01-19, 00:12)hdmkv Wrote: [*]Videos w/DTS-MA = downmixes to standard DTS5.1

Are you sure it is downmixing ? Isn't it more likely to be just passing through the DTS core and ignoring the DTS-HD MA additional stream ?

**hdmkv** · 2016-01-19, 00:35

Hmm, not sure. But, what you're saying is correct I believe. Updated post above.

noggin · 2016-01-19, 00:41

(2016-01-19, 00:35)hdmkv Wrote: Hmm, not sure. But, what you're saying is correct I believe. Updated post above.

Don't think Kodi re-encodes/transcodes to DTS - so not sure how it can be anything else.

**hdmkv** · (This post was last modified: 2016-01-19, 01:08 by hdmkv.)

Right, I think the core is always present with DTS-MA, while with TrueHD, core is a +track... which you can lose in MKV as TrueHD & AC3 must be muxed in as separate tracks if I recall correctly.

Re: Amlinuxmedia forum, they now require registration to view/access...

Quote:Sorry folks.. but this section is for members only.. please register for free to view.

http://amlinuxmedia.com/forum/index.php?topic=13.0

DrJogalog · 2016-01-19, 01:24

With regard to the OpenElec install. Dolby Digital Plus plays fine without any issues on my box. The DTS HD MA tracks do not crash the box, but no audio signal is present either. The video plays without issue, just no sound. (If you switch off passthrough, it will output a stereo PCM signal for the DTS HD MA tracks).

I also have no issue playing 23.976fps videos either.

It has replaced my Windows Kodi box for now as this thing is silent and plays HVEC without issue. I have very few DTS HD MA videos so not an issue there for me.

**hdmkv** · 2016-01-19, 01:30

(2016-01-19, 01:24)DrJogalog Wrote: I also have no issue playing 23.976fps videos either.

Hmm, I may try re-flashing & re-trying. Maybe the crash was due to DTS-MA or DTS present in 23.976 videos.

dukester · 2016-01-19, 04:33

Looks like a good start to me, I'd try forcing 24hz and turning auto refresh rate changing off for now, I think the kernel will need patching but it's a new 64 bit kernel so maybe it's not the same as the s8xx devices, can u test the hevc hdr uhd videos from the last pages of the wetek thread and see how they play?

mylle · 2016-01-19, 07:47

(2016-01-19, 00:46)hdmkv Wrote: Right, I think the core is always present with DTS-MA, while with TrueHD, core is a +track... which you can lose in MKV as TrueHD & AC3 must be muxed in as separate tracks if I recall correctly.

Re: Amlinuxmedia forum, they now require registration to view/access...

Quote:Sorry folks.. but this section is for members only.. please register for free to view.
http://amlinuxmedia.com/forum/index.php?topic=13.0

If you look a little closer you will find that the topic is actually not there anymore. I dont know what happened Sad

http://amlinux.ru/temp/tmp/

This is where Alex posts his builds. Hopefully this is not the last of it.

Regards
Jacob