(2014-05-20, 01:28)nooryani84 Wrote: [ -> ] (2014-05-19, 09:25)kbaggen Wrote: [ -> ]Hi there,
just got this "ALTERNATIVE NUC....SOUND PROOF" 4 days ago:
* MS-Tech CI-70 (including, PicoPSU 120W)
* ASRock Q1900B (Quad-Celeron 2ghs-2,5ghz/MiniITX) (Q1900B-ITX)
Can we please stop calling everything that's miniature a NUC alternative? The NUC is A LOT smaller than the mini-itx motherboard used here http://www.asrock.com/mb/Intel/Q1900B-ITX/
Comparing the two form factors is like comparing ATX to mATX.
NUC (UCFF): 10.16 × 10.16 cm(4 × 4 inches)
Mini-ITX: 17 × 17 cm (6.7 × 6.7 in)
Not to mention the NUC doesn't need anything but memory and an HDD/SSD, it's a much more elegant and clean solution. It's very misleading for people who are actually considering a NUC, which are very highly recommended for people who aren't interested in building a traditional HTPC box.
.......and why dó you Care?
This is a Bay trail tread on motherboard, hence, for builders. And we are in my case discussing a case where you Mount motherboard, Then ram, and ssd. It is so easyjet and close to the NUCs and defacto then an alternative. So what Else to Call it? Secondly, by calling what it is, eg. NUCs alternative, Then People's hopefull nget the idea it is an alternative and have the coice of getting a bang for there money, Eg, better value.
(2014-05-20, 07:13)kbaggen Wrote: [ -> ] (2014-05-20, 01:28)nooryani84 Wrote: [ -> ] (2014-05-19, 09:25)kbaggen Wrote: [ -> ]Hi there,
just got this "ALTERNATIVE NUC....SOUND PROOF" 4 days ago:
* MS-Tech CI-70 (including, PicoPSU 120W)
* ASRock Q1900B (Quad-Celeron 2ghs-2,5ghz/MiniITX) (Q1900B-ITX)
Can we please stop calling everything that's miniature a NUC alternative? The NUC is A LOT smaller than the mini-itx motherboard used here http://www.asrock.com/mb/Intel/Q1900B-ITX/
Comparing the two form factors is like comparing ATX to mATX.
NUC (UCFF): 10.16 × 10.16 cm(4 × 4 inches)
Mini-ITX: 17 × 17 cm (6.7 × 6.7 in)
Not to mention the NUC doesn't need anything but memory and an HDD/SSD, it's a much more elegant and clean solution. It's very misleading for people who are actually considering a NUC, which are very highly recommended for people who aren't interested in building a traditional HTPC box.
.......and why dó you Care?
This is a Bay trail tread on motherboard, hence, for builders. And we are in my case discussing a case where you Mount motherboard, Then ram, and ssd. It is so easyjet and close to the NUCs and defacto then an alternative. So what Else to Call it? Secondly, by calling what it is, eg. NUCs alternative, Then People's hopefull nget the idea it is an alternative and have the coice of getting a bang for there money, Eg, better value.
Oh don't get me wrong, I'm interested in BayTrail and building - though this doesn't serve as informative information for new users that aren't as knowledgeable in this area. You have to remember that we have all kinds of users that frequent these forums, both novice and intermediate. It's our job to make sure that our information is as accurate as possible so that they don't mistakenly make a purchase that isn't right for them. This isn't a even close to a NUC and the value isn't necessarily better as it depends on the users' criteria for an HTPC. We might as well call all ATX pc's mini-HTPCs since the size ratio between NUC / mini-itx and microATX are comparable. It's like saying the Wii is a Xbox360 alternative
Value is in the eye of the beholder.
(2014-05-20, 01:28)nooryani84 Wrote: [ -> ]Oh don't get me wrong, I'm interested in BayTrail and building - though this doesn't serve as informative information for new users that aren't as knowledgeable in this area. You have to remember that we have all kinds of users that frequent these forums, both novice and intermediate. It's our job to make sure that our information is as accurate as possible so that they don't mistakenly make a purchase that isn't right for them. This isn't a even close to a NUC and the value isn't necessarily better as it depends on the users' criteria for an HTPC. We might as well call all ATX pc's mini-HTPCs since the size ratio between NUC / mini-itx and microATX are comparable. It's like saying the Wii is a Xbox360 alternative
Value is in the eye of the beholder.
Ok, we dont get even close to understand each other. In my eyes it is a NUC alternativ, and Wii is an alternative to xbox360! Yes, there is differency in power of the system, just as there differency between the various Intel NUCs! The name NUC is just some cleaver markingting people trying to sell.......they are all mini-pc´s, mini-HTCPs.........so please go to all the NUC treads and ask them to call the NUC what it is = a mini-pc!
But whatever, let us keep the NUC people in the dark ages, let them think there is no alternative, let all money go to Intel, let us sit back and let the none-competition ruin the market
NUC is a model from a brand [Intel] It's not clever marketing.
A NUC alternative would be one that is similar in size [these aren't] running similar specs and cost less.
If you called them a mini HTPC you'd be spot on. But there are many mini HTPCs. Fully customisable or prebuilt. A NUC is a simple and effective mini PC that can be an HTPC too. You'll find most might start with a NUC. Then want to leave the "walled garden" and build their own. NUC users aren't in the dark ages. NUCs aren't exactly cheap once fully built. But they're more than capable. It's down to the end user. If a NUC is too expensive. People will look elsewhere. To say NUC users are stuck in the dark ages is narrow minded. I've had the mk1 847, played with a mk2 i3 and currently have the 2820. For their abilities and size. They're spot on. The new fanless one is my next view!
A wii isn't an alternative to an Xbox. Can you play GTA on wii? Just as my van isn't an alternative to a Ferrari!
(2014-05-19, 15:24)fritsch Wrote: [ -> ]Hi,
I need a tester, does not matter, which Baytrail System. We are currently thinking of giving VPP, the vaapi postprocessing lib a second chance. sraue of OpenELEC has put together a build, which you can use for testing. The advantage of VPP is only used when you watch LiveTV which is interlaced.
VPP implements three different methods for Deinterlacing:
- Simple Bob (IVB, HSW)
- MADI - Motion Adaptive Deinterlacing (IVB, HSW)
- MACI - Motion Compensation Deinterlacing (HSW)
As the Baytrail is much too slow to copy the decoded surfaces back to system memory and then do some advanced deinterlacing, this VPP is highly useful.
You can download the image here:
http://snapshots.openelec.tv/test/OpenEL...063daa.tar
Testing Samples:
Transformers (576i): https://dl.dropboxusercontent.com/u/5572...formers.ts
Servus TV HD (1080i50): https://dl.dropboxusercontent.com/u/5572...g-test.mp4
How to test:
Download sample Transformers and Servus TV HD, play it, navigate to the film reel icon and select: Deinterlace Auto -> Deinterlacing Method: VAAPI (BOB), Scaling: Bilinear.
Press "o" and see if you get 50fps, if there are skips and drops (count only the numbers that appear from that point). Watch sample till the end
Now watch the sample again and select Deinterlacing Method: VAAPI (MADI)
Press "o" and see if you get 50fps, if there are skips and drops (count only the numbers that appear from that point). Watch sample till the end
This was the test for the 576i Sample.
Now repeat the above test with the 1080i50 sample.
Thanks very much
Peter
On Haswell: All methodes despite BOB are broken - which was the reason we abondonnned it long ago
Did some testing on the Q1900-ITX.
I couldn't find the VAAPI (MADI) option. There were only VAAPI BOB, VAAPI AUTO, BOB and BOB (Inverted) deinterlancing methods.
The VAAPI BOB and VAAPI AUTO had same results.
VAAPI BOB:
576i sample: Frame rate remains smooth @ 50 FPS, no noticeable drops or skips. Average CPU usage throughout 4 cores is about 10%. Most of the load goes to Core #1 and Core #2, Core 3 & 4's barely get any load (2-5% avg)
1080i sample: Frame rate jumps all over the place, from 5 to 60 and so on. Constant dropping and skipping. Video plays for 1 second then freezes for about 2-3 seconds and so on. Average CPU usage throughout 4 cores is about 15% with the #1 core doing most of the work (~20%ish)
Couldn't test with the VAAPI MADI since I did not find the deinterlancing method.
(2014-05-20, 10:40)kbaggen Wrote: [ -> ] (2014-05-20, 01:28)nooryani84 Wrote: [ -> ]Oh don't get me wrong, I'm interested in BayTrail and building - though this doesn't serve as informative information for new users that aren't as knowledgeable in this area. You have to remember that we have all kinds of users that frequent these forums, both novice and intermediate. It's our job to make sure that our information is as accurate as possible so that they don't mistakenly make a purchase that isn't right for them. This isn't a even close to a NUC and the value isn't necessarily better as it depends on the users' criteria for an HTPC. We might as well call all ATX pc's mini-HTPCs since the size ratio between NUC / mini-itx and microATX are comparable. It's like saying the Wii is a Xbox360 alternative
Value is in the eye of the beholder.
Ok, we dont get even close to understand each other. In my eyes it is a NUC alternativ, and Wii is an alternative to xbox360! Yes, there is differency in power of the system, just as there differency between the various Intel NUCs! The name NUC is just some cleaver markingting people trying to sell.......they are all mini-pc´s, mini-HTCPs.........so please go to all the NUC treads and ask them to call the NUC what it is = a mini-pc!
But whatever, let us keep the NUC people in the dark ages, let them think there is no alternative, let all money go to Intel, let us sit back and let the none-competition ruin the market
The thing is... the NUC is ENTIRELY different to what you posted. The NUC is a form-factor and not clever marketing. The fact of the matter is that nobody has packaged this kind of horsepower in such a small package before, not at this price point (for what it is, it is cheap considering the size).
(2014-05-20, 10:50)jammyb Wrote: [ -> ]NUC is a model from a brand [Intel] It's not clever marketing.
A NUC alternative would be one that is similar in size [these aren't] running similar specs and cost less.
If you called them a mini HTPC you'd be spot on. But there are many mini HTPCs. Fully customisable or prebuilt. A NUC is a simple and effective mini PC that can be an HTPC too. You'll find most might start with a NUC. Then want to leave the "walled garden" and build their own. NUC users aren't in the dark ages. NUCs aren't exactly cheap once fully built. But they're more than capable. It's down to the end user. If a NUC is too expensive. People will look elsewhere. To say NUC users are stuck in the dark ages is narrow minded. I've had the mk1 847, played with a mk2 i3 and currently have the 2820. For their abilities and size. They're spot on. The new fanless one is my next view!
A wii isn't an alternative to an Xbox. Can you play GTA on wii? Just as my van isn't an alternative to a Ferrari!
Broadly speaking your examples are alternatives, even the van and Ferrari, and the Wii is a specific (gaming) alternative to the Xbox, and so is a PC, you are just restricted by content, but you can still play and enjoy gaming.
So broadly speaking and for all intents and purposes as we are on xbmc.org, any system capable of running xbmc is an alternative. Bay-Trail can be a specific alternative to a NUC, if you have a very small case then it is reasonable to call it an alternative, with the caveat that it is indeed larger than a NUC. A bay trail in a full-size case would not be an alternative.
Price has little to do with being an alternative, unless that is one of your main criteria. For example, there could be a more/less expensive alternatives to the NUC with a similar size.
(2014-05-20, 12:05)nooryani84 Wrote: [ -> ]The thing is... the NUC is ENTIRELY different to what you posted. The NUC is a form-factor and not clever marketing. The fact of the matter is that nobody has packaged this kind of horsepower in such a small package before, not at this price point (for what it is, it is cheap considering the size).
*cough* Zotac *cough*
Fine, whatever, I just hope others can use the info I try to bring if they as me wish to have a complet soundless system, in a nice HiFi look/n/like box. I am pretty sure the people ending up here on this tread can disecrate the wording "NUC alternative" as what it is!
(2014-05-20, 12:05)lewzer Wrote: [ -> ] (2014-05-19, 15:24)fritsch Wrote: [ -> ]Hi,
I need a tester, does not matter, which Baytrail System. We are currently thinking of giving VPP, the vaapi postprocessing lib a second chance. sraue of OpenELEC has put together a build, which you can use for testing. The advantage of VPP is only used when you watch LiveTV which is interlaced.
VPP implements three different methods for Deinterlacing:
- Simple Bob (IVB, HSW)
- MADI - Motion Adaptive Deinterlacing (IVB, HSW)
- MACI - Motion Compensation Deinterlacing (HSW)
As the Baytrail is much too slow to copy the decoded surfaces back to system memory and then do some advanced deinterlacing, this VPP is highly useful.
You can download the image here:
http://snapshots.openelec.tv/test/OpenEL...063daa.tar
Testing Samples:
Transformers (576i): https://dl.dropboxusercontent.com/u/5572...formers.ts
Servus TV HD (1080i50): https://dl.dropboxusercontent.com/u/5572...g-test.mp4
How to test:
Download sample Transformers and Servus TV HD, play it, navigate to the film reel icon and select: Deinterlace Auto -> Deinterlacing Method: VAAPI (BOB), Scaling: Bilinear.
Press "o" and see if you get 50fps, if there are skips and drops (count only the numbers that appear from that point). Watch sample till the end
Now watch the sample again and select Deinterlacing Method: VAAPI (MADI)
Press "o" and see if you get 50fps, if there are skips and drops (count only the numbers that appear from that point). Watch sample till the end
This was the test for the 576i Sample.
Now repeat the above test with the 1080i50 sample.
Thanks very much
Peter
On Haswell: All methodes despite BOB are broken - which was the reason we abondonnned it long ago
Did some testing on the Q1900-ITX.
I couldn't find the VAAPI (MADI) option. There were only VAAPI BOB, VAAPI AUTO, BOB and BOB (Inverted) deinterlancing methods.
The VAAPI BOB and VAAPI AUTO had same results.
VAAPI BOB:
576i sample: Frame rate remains smooth @ 50 FPS, no noticeable drops or skips. Average CPU usage throughout 4 cores is about 10%. Most of the load goes to Core #1 and Core #2, Core 3 & 4's barely get any load (2-5% avg)
1080i sample: Frame rate jumps all over the place, from 5 to 60 and so on. Constant dropping and skipping. Video plays for 1 second then freezes for about 2-3 seconds and so on. Average CPU usage throughout 4 cores is about 15% with the #1 core doing most of the work (~20%ish)
Couldn't test with the VAAPI MADI since I did not find the deinterlancing method.
Those are really, really bad results :-(
That means that the GPU is too slow to do Deinterlacing and Decoding - ouh man - I would never have expected that.
Can you post an xbmc.log please from bootup and with those samples? Are you sure that scaling was set to Bilinear?
Background: We are currently using swfilter in OpenELEC. This copies the surfaces back from the GPU to system memory and there we deinterlace on the CPU. With that method that Baytrail Chip can do 1080i50 with BOB quite fine.
That's a desaster :-(
(2014-05-20, 12:33)ertman Wrote: [ -> ] (2014-05-20, 10:50)jammyb Wrote: [ -> ]NUC is a model from a brand [Intel] It's not clever marketing.
A NUC alternative would be one that is similar in size [these aren't] running similar specs and cost less.
If you called them a mini HTPC you'd be spot on. But there are many mini HTPCs. Fully customisable or prebuilt. A NUC is a simple and effective mini PC that can be an HTPC too. You'll find most might start with a NUC. Then want to leave the "walled garden" and build their own. NUC users aren't in the dark ages. NUCs aren't exactly cheap once fully built. But they're more than capable. It's down to the end user. If a NUC is too expensive. People will look elsewhere. To say NUC users are stuck in the dark ages is narrow minded. I've had the mk1 847, played with a mk2 i3 and currently have the 2820. For their abilities and size. They're spot on. The new fanless one is my next view!
A wii isn't an alternative to an Xbox. Can you play GTA on wii? Just as my van isn't an alternative to a Ferrari!
Broadly speaking your examples are alternatives, even the van and Ferrari, and the Wii is a specific (gaming) alternative to the Xbox, and so is a PC, you are just restricted by content, but you can still play and enjoy gaming.
So broadly speaking and for all intents and purposes as we are on xbmc.org, any system capable of running xbmc is an alternative. Bay-Trail can be a specific alternative to a NUC, if you have a very small case then it is reasonable to call it an alternative, with the caveat that it is indeed larger than a NUC. A bay trail in a full-size case would not be an alternative.
Price has little to do with being an alternative, unless that is one of your main criteria. For example, there could be a more/less expensive alternatives to the NUC with a similar size.
Broadly speaking, sure. But I'm not broadly speaking. You are.
If I was specifically looking for a NUC alternative for XBMC. Then as you said. A Zotac would be near it. :
Same goes if I specifically wanted to play GTA, CoD or even Batman. I'd choose Xbox or Playstation. The Wii isn't an alternative.
But broadly speaking. They can all play games. But I'm not broadly speaking.
My van isn't red, nor has 100's of horses or leather interior. If I was looking specifically to buy a supercar. I wouldn't be looking in van trader. But broadly speaking. [which I'm not, remember] I'd have more room in the back and better MPG!
We all understand where the OP comes from. If anything it should be "NUC inspired" not "NUC alternative".
(2014-05-20, 13:21)jammyb Wrote: [ -> ] (2014-05-20, 12:33)ertman Wrote: [ -> ] (2014-05-20, 10:50)jammyb Wrote: [ -> ]NUC is a model from a brand [Intel] It's not clever marketing.
A NUC alternative would be one that is similar in size [these aren't] running similar specs and cost less.
If you called them a mini HTPC you'd be spot on. But there are many mini HTPCs. Fully customisable or prebuilt. A NUC is a simple and effective mini PC that can be an HTPC too. You'll find most might start with a NUC. Then want to leave the "walled garden" and build their own. NUC users aren't in the dark ages. NUCs aren't exactly cheap once fully built. But they're more than capable. It's down to the end user. If a NUC is too expensive. People will look elsewhere. To say NUC users are stuck in the dark ages is narrow minded. I've had the mk1 847, played with a mk2 i3 and currently have the 2820. For their abilities and size. They're spot on. The new fanless one is my next view!
A wii isn't an alternative to an Xbox. Can you play GTA on wii? Just as my van isn't an alternative to a Ferrari!
Broadly speaking your examples are alternatives, even the van and Ferrari, and the Wii is a specific (gaming) alternative to the Xbox, and so is a PC, you are just restricted by content, but you can still play and enjoy gaming.
So broadly speaking and for all intents and purposes as we are on xbmc.org, any system capable of running xbmc is an alternative. Bay-Trail can be a specific alternative to a NUC, if you have a very small case then it is reasonable to call it an alternative, with the caveat that it is indeed larger than a NUC. A bay trail in a full-size case would not be an alternative.
Price has little to do with being an alternative, unless that is one of your main criteria. For example, there could be a more/less expensive alternatives to the NUC with a similar size.
Broadly speaking, sure. But I'm not broadly speaking. You are.
If I was specifically looking for a NUC alternative for XBMC. Then as you said. A Zotac would be near it. :
Same goes if I specifically wanted to play GTA, CoD or even Batman. I'd choose Xbox or Playstation. The Wii isn't an alternative.
But broadly speaking. They can all play games. But I'm not broadly speaking.
My van isn't red, nor has 100's of horses or leather interior. If I was looking specifically to buy a supercar. I wouldn't be looking in van trader. But broadly speaking. [which I'm not, remember] I'd have more room in the back and better MPG!
We all understand where the OP comes from. If anything it should be "NUC inspired" not "NUC alternative".
That's a more sensible answer than I gave and couldn't agree more.
(2014-05-20, 12:40)kbaggen Wrote: [ -> ]Fine, whatever, I just hope others can use the info I try to bring if they as me wish to have a complet soundless system, in a nice HiFi look/n/like box. I am pretty sure the people ending up here on this tread can disecrate the wording "NUC alternative" as what it is!
Wasn't meant as criticism and I know you were just trying to be helpful
Thing is, these products are quite different in the area of HTPCs.
(2014-05-20, 13:06)fritsch Wrote: [ -> ] (2014-05-20, 12:05)lewzer Wrote: [ -> ] (2014-05-19, 15:24)fritsch Wrote: [ -> ]Hi,
I need a tester, does not matter, which Baytrail System. We are currently thinking of giving VPP, the vaapi postprocessing lib a second chance. sraue of OpenELEC has put together a build, which you can use for testing. The advantage of VPP is only used when you watch LiveTV which is interlaced.
VPP implements three different methods for Deinterlacing:
- Simple Bob (IVB, HSW)
- MADI - Motion Adaptive Deinterlacing (IVB, HSW)
- MACI - Motion Compensation Deinterlacing (HSW)
As the Baytrail is much too slow to copy the decoded surfaces back to system memory and then do some advanced deinterlacing, this VPP is highly useful.
You can download the image here:
http://snapshots.openelec.tv/test/OpenEL...063daa.tar
Testing Samples:
Transformers (576i): https://dl.dropboxusercontent.com/u/5572...formers.ts
Servus TV HD (1080i50): https://dl.dropboxusercontent.com/u/5572...g-test.mp4
How to test:
Download sample Transformers and Servus TV HD, play it, navigate to the film reel icon and select: Deinterlace Auto -> Deinterlacing Method: VAAPI (BOB), Scaling: Bilinear.
Press "o" and see if you get 50fps, if there are skips and drops (count only the numbers that appear from that point). Watch sample till the end
Now watch the sample again and select Deinterlacing Method: VAAPI (MADI)
Press "o" and see if you get 50fps, if there are skips and drops (count only the numbers that appear from that point). Watch sample till the end
This was the test for the 576i Sample.
Now repeat the above test with the 1080i50 sample.
Thanks very much
Peter
On Haswell: All methodes despite BOB are broken - which was the reason we abondonnned it long ago
Did some testing on the Q1900-ITX.
I couldn't find the VAAPI (MADI) option. There were only VAAPI BOB, VAAPI AUTO, BOB and BOB (Inverted) deinterlancing methods.
The VAAPI BOB and VAAPI AUTO had same results.
VAAPI BOB:
576i sample: Frame rate remains smooth @ 50 FPS, no noticeable drops or skips. Average CPU usage throughout 4 cores is about 10%. Most of the load goes to Core #1 and Core #2, Core 3 & 4's barely get any load (2-5% avg)
1080i sample: Frame rate jumps all over the place, from 5 to 60 and so on. Constant dropping and skipping. Video plays for 1 second then freezes for about 2-3 seconds and so on. Average CPU usage throughout 4 cores is about 15% with the #1 core doing most of the work (~20%ish)
Couldn't test with the VAAPI MADI since I did not find the deinterlancing method.
Those are really, really bad results :-(
That means that the GPU is too slow to do Deinterlacing and Decoding - ouh man - I would never have expected that.
Can you post an xbmc.log please from bootup and with those samples? Are you sure that scaling was set to Bilinear?
Background: We are currently using swfilter in OpenELEC. This copies the surfaces back from the GPU to system memory and there we deinterlace on the CPU. With that method that Baytrail Chip can do 1080i50 with BOB quite fine.
That's a desaster :-(
fritsch,
I ran those 2 samples on my "NUC
altenative inspired" system yesterday J1900, eg. win7, hence I assume you cannot use it for much, but with the following specs I got 50 fps, 1 or 2 skips if I recall!.
*Deinterlace (OSD video setting), DVXA BEST or BOB (same fps)
*Upscalering (OSD video setting): DVXA
*Under System--> Video -->
1. Hardwareaccelation ON --> Pixel-shardes
2. 10%, upscalere .
3. Hardwareaccelation kan use DVXA2
(Also adjusted the framerat to movies = ON, at start/stop)
I also used LAcross3 (only posible for the i576 sample), and it gave also 50fps........If I understood your former topic that should not be poosible?
In all it looked very smooth!
Jep, that should not be possible. Cause another tester made windows tests for me and ended with just 30 fps. Do you have a logfile? Btw. Can you check the CPU load while doing that. IIRC DXVA copies the surface back to memory anyways, that one should be seen in the cpu load.
I also still want the other OE tests logfiles to verify that there were no other issues.