147 replies to this topic

[buyukbang]

Hi guys,

I need especially developer responses and my aim is to find the root cause of the bad picture quality with Enigma2.

As you know there are some boxes are praised for their picture quality like Katherin, Humax, Techisat , Azbox...

I used only Azbox with its official firmware and now using VU+ . I can easily say that SD picture quality of Azbox with its official is a way much better than VU+ and Dreambox that I tested.

Another interesting point here, Enigma2 can be installed on both Azbox and Katherin, but E2 on this boxes sucks as VU+ and Dreambox when compared to their official firmware. So I'm sure that this problem should not be related with hardware?

I don't want to use other boxes with their silly and inflexible firmware just for picture quality, instead I would really like to have a better picture quality on my lovely UI, E2.

So what is the wrong with E2 in terms of picture quality. Can't we do anything to improve it with just coding ? Is it related with drivers or some scaler part of E2?

Any ideas would be appreciated....

Edited by buyukbang, 6 March 2012 - 12:52.

[Hump]

You are right. I still own a Kathrein Receiver with original software in flash and e2 bootable (multiboot) on HD. the picture quality with the original software is much better as with e2. e2 doesn't even get close to the perfect picture quality of original software.

e2 on my mipsel box et9x00 is pretty good but there is still room for improvement until it reaches the picture quality of the Kathrein original software.

[Erik Slagter]

The "picture quality" of SD channels is only determined by the scaler, that is a hardware feature, although it may be tweaked a bit.

So step one is to adjust the scaler parameters to your likings using the "video settings" plugin.

If that still doesn't solve the issue, try to compare the SoC of both boxes. It may be that the Azbox has a SoC that simply has a better (?) scaler.

Enigma itself has no influence on the picture quality whatsoever, it is never in the middle of the video stream.

[MiLo]

Enigma2 doesn't do any video rendering whatsoever, so it's the chipset and the drivers that do all the work. If you are using a HDMI connection, there should be no difference at all, regardless of what you're using. If there is a difference in the digital stream from one receiver to another, that would indicate a plain bug, not a "quality" issue. Of course, assuming that the receivers aren't doing any postprocessing (e.g. scaling, deinterlacing) but you should let the TV do that anyway.

[Kers]

Enigma2 doesn't do any video rendering whatsoever, so it's the chipset and the drivers that do all the work. If you are using a HDMI connection, there should be no difference at all, regardless of what you're using. If there is a difference in the digital stream from one receiver to another, that would indicate a plain bug, not a "quality" issue. Of course, assuming that the receivers aren't doing any postprocessing (e.g. scaling, deinterlacing) but you should let the TV do that anyway.

As the question was about SD picture quality, I dont think its an option to let the TV do the scaling and deinterlacing. Wouldnt that mean having to use SD resolution for the OSD?

I also have a Vu+ Duo and an Azbox Premium. Hands down the best SD picture quality is on the Azbox with original firmware, unfortunately that firmware is in all other regards hopeless. Especially deinterlacing performance is much better in Azbox, even with tweaking "video settings" plugin.

[Erik Slagter]

In that case it's simply the (perceived!!!) difference in the scalers of both SoC's, you won't get around that...

[malakudi]

As far as STI boxes are concerned, the difference of original software vs enigma2 software lies in the drivers used. Drivers of original software use STAPI, an API for communicating with hardware from STi. enigma2 is not STAPI compatible, so other drivers are used that conform to DVBAPI. DVBAPI drivers for STi boxes are based to an unsupported project called "Havana" from STi. The source code of this project was leaked and most DVBAPI drivers for STi boxes are based on these sources. However, the quality of these drivers is much worse than the STAPi drivers, and thus the lower image quality in E2.
The problem from Azbox is similar, different drivers for E2 releases, not from Sigma but from other developers.

Broadcom chipset picture quality is indeed lower than STi or Sigma solutions with original drivers.

[Erik Slagter]

Broadcom chipset picture quality is indeed lower than STi or Sigma solutions with original drivers.

You can't say that, only that the perceived quality of the scaler is different. If you watch HD over HDMI, there can be no difference anyway.

[Kers]

Yes, we are obviously talking about perceived quality and at least for me there is no issue with HD.
Thanks for explaining the driver situation malakudi, now it makes alot more sense.

The main issue that is visible for me in the SD quality is that deinterlacing is "slow", with the Azbox original firmware there is hardly any jagged interlacing lines even with fast moving picture. With the Vu+ Duo even with slow movement the intenlacing shows up.

On a related note, is there a difference in terms of different generations of Broadcom SoCs? I have understood that Vu+ Uno and Ultimo have more recent and powerful chip. Are you aware of any improvements or side by side comparisons bethween older/newer chips?

[Erik Slagter]

The main issue that is visible for me in the SD quality is that deinterlacing is "slow", with the Azbox original firmware there is hardly any jagged interlacing lines even with fast moving picture. With the Vu+ Duo even with slow movement the intenlacing shows up.

There you're contradicting yourself

If deinterlacing is performed unconditionally and thoroughly, the picture maybe perceived as "slow" and "smeared out" during quick movements. There will be absolutely no jagged edges (comb effect) in that case whatsoever. If deinterlacing is performed dynamically, aka motion adaptive interlacing, interlacing is performed much more subtile, which prevents the smear-effect, but also can give a bit of a comb effect now and then. Please note that deinterlacing throws away a considerate amount of motion data, so it should be used as sparingly as possible.

On the same note, you shouldn't let your stb do any deinterlacing at all. If the stb deinterlaces, all of the additional motion information is simply discarded. The TV should do any interlacing because it can make 50/60fps pseudo progressive from 50fps interlaced material and thus does not need to throw away any motion information.

If you configured your STB correctly, it would not deinterlace any material, but would apply "bob+weave" before scaling and "interleave" after scaling. You will find these terms on web pages describing deinterlacing (eg. 100fps.com), but technically none of them are. Never in this process any motion information is discarded.

What it does is: make full frames from every field, so the material becomes "pseudo" 50p. In this process some blending is performed to prevent (excessive) combing. The same considerations apply as with actual deinterlacing - tradeoff between fluid motion and possible combing - although no actual deinterlacing is performed! This pseudo 50p material is then scaled and afterwards interleaved, taking odd lines from odd frames and even lines from even frames, creating 50 fields per second from 50 frames per second, this is again 50i material and completely fluiently watchable on your television set.

As mentioned before, the "scaler sharpness" value on your enigma stb determines the tradeoff between sharpness and possible combing. I accept a little combing because it comes with a much sharper image. If you don't want to see any combing at all, you should decrease this setting at the cost of less sharpness.

On a related note, is there a difference in terms of different generations of Broadcom SoCs? I have understood that Vu+ Uno and Ultimo have more recent and powerful chip. Are you aware of any improvements or side by side comparisons bethween older/newer chips?

Afaik all current Broadcom SoC's have the same scaler hardware, only the dm800 has a SoC with a clearly lower quality scaler (according to some).

[malakudi]

Broadcom chipset picture quality is indeed lower than STi or Sigma solutions with original drivers.

You can't say that, only that the perceived quality of the scaler is different. If you watch HD over HDMI, there can be no difference anyway.

I also have an IPBOX 9000HD with STi chipset. Picture, even on HD channels is much better. Colors are better. I didn't do double blind test, but that is what it seems to me. I also experimented with autoresolution plugin and native resolution for SD channels to avoid scaler, but I still believe STi has better picture quality.

[Erik Slagter]

If you see any difference between STB'es IF the material's resolution is equal to the panel's native resolution AND you're using digital transmission (HDMI/DVI) THEN something MUST be wrong, simple.

No STB should alter the (digital) video stream as received, especially not without the user's explicit consent.

[buyukbang]

Thanks for this great clarification. Thigs are much clear for me now.

So why STB producers cannot modify the source codes to get a better picture quality ? Simply unskilled developers are the reason ?

As far as STI boxes are concerned, the difference of original software vs enigma2 software lies in the drivers used. Drivers of original software use STAPI, an API for communicating with hardware from STi. enigma2 is not STAPI compatible, so other drivers are used that conform to DVBAPI. DVBAPI drivers for STi boxes are based to an unsupported project called "Havana" from STi. The source code of this project was leaked and most DVBAPI drivers for STi boxes are based on these sources. However, the quality of these drivers is much worse than the STAPi drivers, and thus the lower image quality in E2.
The problem from Azbox is similar, different drivers for E2 releases, not from Sigma but from other developers.

Broadcom chipset picture quality is indeed lower than STi or Sigma solutions with original drivers.

I can also confirm the same about the autoresolution, I tried it but it sucks for me. Even my VU+ duo does it job better. So I would definitely prefer Azbox / Katherin for the SD/HD picture quality. Bur I cant give up E2 , neither

I also have an IPBOX 9000HD with STi chipset. Picture, even on HD channels is much better. Colors are better. I didn't do double blind test, but that is what it seems to me. I also experimented with autoresolution plugin and native resolution for SD channels to avoid scaler, but I still believe STi has better picture quality.

[Kers]

Maybe my terminology was not quite correct and I realize its a trade off bethween sharpness, motion blur and combing. That being said, I still stand by my evaluation that I see less combing, sharper picture and less motion blur with the Azbox than with the Vu+ Duo.

I'm also not sure that you can state categorically that you should only do deinterlacing and scaling at the display. Wouldnt that depend on the comparative quality and capabilities of the DSP in the TV, STB and possible AV ampllifier?

I will play around with the autoresolution setting together with my TV settings and see what I can achieve.

One question though,If the STB upscales the picture to 1080i without deinterlacing it (for high resolution OSD), would the TV even understand how to deinterlace the picture correcly as there would by two layers of interlacing in the material it receives (576i of the source material and the 1080i STB->TV link)?

[malakudi]

If you see any difference between STB'es IF the material's resolution is equal to the panel's native resolution AND you're using digital transmission (HDMI/DVI) THEN something MUST be wrong, simple.

No STB should alter the (digital) video stream as received, especially not without the user's explicit consent.

It's the video decoder code and algorithm in the broadcom chips that is worse in my opinion. Input is digital, output is digital, but you are not forwarding the digital content to the tv set for decoding. STB chip decodes MPEG2 or H.264 data and converts it to HDMI signal, this convertion is worse in broadcom chipsets in my opinion. I see saturated reds for example in all broadcom chipsets, from old DM800 to latest BCM7413 found in VU+ Ultimo and ET9200. Colors on STi seems much more "correct". When I do tests for picture quality I usually use prerecorded material, 1080i, so scaling and deinterlacing is not in the equation.

Edited by malakudi, 6 March 2012 - 16:04.

[Erik Slagter]

Kers, I'd really like you to read again all what I wrote earlier on this subject, because you really didn't understand some of it.

For example, why you NEVER would want to do deinterlacing (according to the strict definition) on a STB. You're discarding information, the picture becomes jittery, like a true 25p recording! 50i really has more motion information than 25p can ever hold.

There is no need all to use autoresolution, did I ever mention that? If you simply leave all settings to their defaults, it will bob+weave, scale, interleave as I described above.

The frames sent to the television are NOT "doubly interlaced", never. See my explanation. The scaler makes 50p from the 50i material, at that point there is no more interlacing (although this is NO deinterlacing) and then scales. You cannot scale interlaced material without horrible artifacts.

[Erik Slagter]

So why STB producers cannot modify the source codes to get a better picture quality ? Simply unskilled developers are the reason ?

Scaler = hardware. There is no software involved.

Also what you call better, may not be perceived the same way by others.

[Erik Slagter]

It's the video decoder code and algorithm in the broadcom chips that is worse in my opinion. Input is digital, output is digital, but you are not forwarding the digital content to the tv set for decoding. STB chip decodes MPEG2 or H.264 data and converts it to HDMI signal, this convertion is worse in broadcom chipsets in my opinion. I see saturated reds for example in all broadcom chipsets, from old DM800 to latest BCM7413 found in VU+ Ultimo and ET9200. Colors on STi seems much more "correct". When I do tests for picture quality I usually use prerecorded material, 1080i, so scaling and deinterlacing is not in the equation.

Video DECODING is a lossless operation. If two decoders yield different bitstreams, then at least one of them is faulty. This clearly as opposed to video ENCODING. There is only one lossy step in the whole chain and that is at the encoder at the provider, after that, everything (including beaming up and reception) is lossless.

[Kers]

I realize that 50i has more motion information than 25p, I would imagine that being the reason why interlacing is used. I mentioned autoresolution as it seems to give more granular control over what is done to the picture in different modes. I will try with the default settings without any deinterlacing in the STB, but with 1080/50i output.

My point is that as far as I understand the Vu+ duo does not support 1080/50p output, only 1080/50i. Wouldnt this mean that the actual interlacing in the source material (576/50i) differs from the interlacing in the medium between STB and TV? In other words, when the STB does the process you described (576/50i->576/50p->1080/60p->1080/50i), the combing lines do not match the interleaved lines, thus making it more difficult for the TV to properly do the deinterlacing.

Does that make any more sense?

[buyukbang]

OK, but this still cannot explain my below quote. I followed Ripper while he is porting the E2 to Azbox, and I remember that he told he used Sigma's API's for scaling/deinterlacing, if I'm not wrong. But the results are very different from the original firmware. Katherin users report exactly the same. So I think this problem is hardware independent and malakudi's explanation seems more logical to me. I think all developers preparing drivers to E2 use some common code (may be just for cheap and easy e2 compability ) which make the picture quality terrible.

So why STB producers cannot modify the source codes to get a better picture quality ? Simply unskilled developers are the reason ?

Scaler = hardware. There is no software involved.

Also what you call better, may not be perceived the same way by others.

Another interesting point here, Enigma2 can be installed on both Azbox and Katherin, but E2 on this boxes sucks as VU+ and Dreambox when compared to their official firmware. So I'm sure that this problem should not be related with hardware?

Edited by buyukbang, 6 March 2012 - 17:22.