Overclock.net › Forums › Industry News › Hardware News › [DH] Avivo HD Vs Purevideo HD
New Posts  All Forums:Forum Nav:

[DH] Avivo HD Vs Purevideo HD

post #1 of 7
Thread Starter 
Quote:
Every time a new graphics card is released we hear great claims about how the new video enhancements will improve our viewing experience or that the hardware is as perfect as it can get in the High Definition HQV Benchmark. Unfortunately, for many reasons most end users have to pretty much accept these claims as it is not really possible for the average consumer to set up two identical systems, side by side and see what the differences are. Additionally most reviews are just the opinion of a single person with no clear examples and can be hugely subjective, even in HQV tests.

But what if you could see the image quality of both ATI and Nvidia’s High Definition content, in full HD, on your screen, today? What would it tell you about how these two companies perform in real world use rather than a synthetic benchmark?

Well today you can do that. Here at Driver Heaven we have spent a considerable amount of time working through the various hurdles that stop most people from capturing images from HD-DVD and Blu-Ray and we are now ready to make the results available for everyone.
Source

Bit of a long (but interesting) read with no outright winner. ATi has better detail & Nvidia offers more vibrant colours. Seems it's all down to personnal taste & tweaking as usual!
arrrrgh
(13 items)
 
  
CPUMotherboardGraphicsRAM
Q6600 G0 @3.4'ish ASUS P5E XFX HD6950 s/w to HD6970 4GB OCZ DDR2 1066 @850 4-4-4-12 
Hard DriveOSMonitorKeyboard
3x1TB & 3x320GB Barracuda's Win 7 Ultimate 64 & XP PRO X64 ACER P243W 1920x1200 :D Logitech DiNovo 
PowerCaseMouseMouse Pad
OCZ ModXstream 700W Lian-li PC-A71B G5 Everglide Ricochet 
  hide details  
Reply
arrrrgh
(13 items)
 
  
CPUMotherboardGraphicsRAM
Q6600 G0 @3.4'ish ASUS P5E XFX HD6950 s/w to HD6970 4GB OCZ DDR2 1066 @850 4-4-4-12 
Hard DriveOSMonitorKeyboard
3x1TB & 3x320GB Barracuda's Win 7 Ultimate 64 & XP PRO X64 ACER P243W 1920x1200 :D Logitech DiNovo 
PowerCaseMouseMouse Pad
OCZ ModXstream 700W Lian-li PC-A71B G5 Everglide Ricochet 
  hide details  
Reply
post #2 of 7
This is a very, very biased review.

For one, some settings are enabled in the ATI panel whose equivalents aren't enabled in the NVIDIA panel (such as denoising, deinterlacing, color correction).

Deinterlacing doesn't seem to matter in the frame they used for comparison, but color correction does. The ATI panel also uses a lower value for Gamma.

Additionally, there seems to be a difference in the Levels value they outputted.

I'll explain Levels values in a very simple way: TV monitors have a Levels range from 16 to 235. PC monitors have a Levels range from 0 to 255.

When you view a video (or even an image) meant for a TV, it doesn't have those additional values, so it's padded. This is perceived as extra brightness, or a "white haze" all over the screen.
(Compare ATI (padded Levels, 16-235 to 0-255) to NVIDIA (proper Levels, 16-235 to 16-235))

The thing is that you can't really make the video "stay" at 16-235. It absolutely has to be padded to be displayed on a PC monitor, so these values have to be clipped later.

This often results in dark parts of the image becoming too dark, resulting in "loss" of detail (it's still there, it's just not visible), and bright parts of the image becoming too bright, also resulting in the same kind of loss.

The big problem with this review is that the NVIDIA features that are supposed to make up for this (e.g. color correction) are disabled. Woohoo!
(NVIDIA - dynamic contrast enhancement)

The ATI panel doesn't seem to even have a "smart" feature like this. The difference is likely just because in one it is completely disabled (NVIDIA), and in another (ATI) it's partially enabled.

I'll come back soon and apply the conversion to one of the ATI BMPs, just so you guys can see there is no (big) difference.

Also, haha. Professionals? More like wannabe professionals. Had they wanted to be more exact and objective in their comparison, they would've used a tool like Avisynth to subtract the differences between the two frames.

Since when do you use Photoshop for video? Many of the differences are only visible when you can separate each layer of video (e.g. luma, chroma).

After you save it to a screenshot, it's much harder to make an objective comparison since it was converted to RGB and you can no longer use software to show you the exact differences.

The point of my rant:

Watch material meant for the TV on your TV, through a hardware player. It looks much, much better. "The Way It's Meant To Be Viewed"
Akiyama Mio
(13 items)
 
  
CPUMotherboardGraphicsRAM
E6420 @ stock, 0.98v Asus P5N-E SLI Gainward GTX 460 1GB @ 800/1600/1900 2x2GB Kingston @ 800MHz 5-5-5-15 2T 
Hard DriveOptical DriveOSMonitor
WD 250GB, 320GB SATA/3, 16MB Cache, Seagate 1TB LG GSA-H62N 18x SATA Ubuntu 9.10 x86 & Win7 x86 Asus VW222U 
KeyboardPowerCase
Logitech Classic Corsair 650HX NZXT Apollo Black 
  hide details  
Reply
Akiyama Mio
(13 items)
 
  
CPUMotherboardGraphicsRAM
E6420 @ stock, 0.98v Asus P5N-E SLI Gainward GTX 460 1GB @ 800/1600/1900 2x2GB Kingston @ 800MHz 5-5-5-15 2T 
Hard DriveOptical DriveOSMonitor
WD 250GB, 320GB SATA/3, 16MB Cache, Seagate 1TB LG GSA-H62N 18x SATA Ubuntu 9.10 x86 & Win7 x86 Asus VW222U 
KeyboardPowerCase
Logitech Classic Corsair 650HX NZXT Apollo Black 
  hide details  
Reply
post #3 of 7
I have to admit that some bells went off when I saw they were doing photoshop comparisons. I'm not an expert, but I do know my way around AviSynth and what I found funny was when they said that they weren't comparing the exact same frames from one graphics card to another. AviSynth is frame-accurate, so that would have been the tool of choice. I mean, make a script and slap it into virtualdub, and you can extract exactly which frames you want, in PNG format no less.

Another thing that's interesting is that ATI was consistently lighter than Nvidia, which I think is likely a bi-product of the the colorspace choice (YUY2 was it?)

I often watch AVC on my PC, and unless I use RGB32 as the output in FFDSHOW, I get a washed out image. I read it had something to do with the established standard that video should always be output at TV levels, regardless of the actual display used... but I'd have to check back at the thread I started at doom9...

So yeah... I know that many enthusiast use ATI for HTPC builds (a quick glance at AVSForums will confirm my statement), so I'd probably go with that regardless.
My System
(13 items)
 
  
CPUMotherboardGraphicsRAM
i5 - 3570 Gigabyte GA-965P-S3 HIS 4670 IceQ 3x1GB G.Skill HK series 
Hard DriveOSMonitorKeyboard
4x1TB, 3 for storage 1 for OS Windows Vista Samsung Syncmaster 940BW -- 
PowerCaseMouse
Antec NeoPower 500W Antec P182 -- 
  hide details  
Reply
My System
(13 items)
 
  
CPUMotherboardGraphicsRAM
i5 - 3570 Gigabyte GA-965P-S3 HIS 4670 IceQ 3x1GB G.Skill HK series 
Hard DriveOSMonitorKeyboard
4x1TB, 3 for storage 1 for OS Windows Vista Samsung Syncmaster 940BW -- 
PowerCaseMouse
Antec NeoPower 500W Antec P182 -- 
  hide details  
Reply
post #4 of 7
Did they use HQV and HD-HQV at all?
Once again...
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 920 [4.28GHz, HT] Asus P6T + Broadcom NetXtreme II VisionTek HD5850 [900/1200] + Galaxy GT240 2x4GB G.Skill Ripjaw X [1632 MHz] 
Hard DriveOSMonitorKeyboard
Intel X25-M 160GB + 3xRAID0 500GB 7200.12 Window 7 Pro 64 Acer H243H + Samsung 226BW XARMOR-U9BL  
PowerCaseMouseMouse Pad
Antec Truepower New 750W Li Lian PC-V2100 [10x120mm fans] Logitech G9 X-Trac Pro 
  hide details  
Reply
Once again...
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 920 [4.28GHz, HT] Asus P6T + Broadcom NetXtreme II VisionTek HD5850 [900/1200] + Galaxy GT240 2x4GB G.Skill Ripjaw X [1632 MHz] 
Hard DriveOSMonitorKeyboard
Intel X25-M 160GB + 3xRAID0 500GB 7200.12 Window 7 Pro 64 Acer H243H + Samsung 226BW XARMOR-U9BL  
PowerCaseMouseMouse Pad
Antec Truepower New 750W Li Lian PC-V2100 [10x120mm fans] Logitech G9 X-Trac Pro 
  hide details  
Reply
post #5 of 7
Quote:
Originally Posted by Jimmy2Shoe View Post
I have to admit that some bells went off when I saw they were doing photoshop comparisons. I'm not an expert, but I do know my way around AviSynth and what I found funny was when they said that they weren't comparing the exact same frames from one graphics card to another. AviSynth is frame-accurate, so that would have been the tool of choice. I mean, make a script and slap it into virtualdub, and you can extract exactly which frames you want, in PNG format no less.

Another thing that's interesting is that ATI was consistently lighter than Nvidia, which I think is likely a bi-product of the the colorspace choice (YUY2 was it?)

I often watch AVC on my PC, and unless I use RGB32 as the output in FFDSHOW, I get a washed out image. I read it had something to do with the established standard that video should always be output at TV levels, regardless of the actual display used... but I'd have to check back at the thread I started at doom9...

So yeah... I know that many enthusiast use ATI for HTPC builds (a quick glance at AVSForums will confirm my statement), so I'd probably go with that regardless.
Most video is in the YV12 colorspace nowadays. x264, for example, only accepts YV12 input.

Also, all DVD movies are YV12. I imagine it's the same for all HD media (MPEG-2 HD and H.264 HD is in YV12, but I'm not sure about VC-1).

Doing a colorspace conversion to anything but RGB causes quality loss. Converting YV12 to RGB is kinda like converting 192kbps MP3 to 320kbps MP3.
There's no quality loss (or gain), it's just bigger.

If, as you say, they didn't use the same frames, then that puts this review at an even worse position.
Akiyama Mio
(13 items)
 
  
CPUMotherboardGraphicsRAM
E6420 @ stock, 0.98v Asus P5N-E SLI Gainward GTX 460 1GB @ 800/1600/1900 2x2GB Kingston @ 800MHz 5-5-5-15 2T 
Hard DriveOptical DriveOSMonitor
WD 250GB, 320GB SATA/3, 16MB Cache, Seagate 1TB LG GSA-H62N 18x SATA Ubuntu 9.10 x86 & Win7 x86 Asus VW222U 
KeyboardPowerCase
Logitech Classic Corsair 650HX NZXT Apollo Black 
  hide details  
Reply
Akiyama Mio
(13 items)
 
  
CPUMotherboardGraphicsRAM
E6420 @ stock, 0.98v Asus P5N-E SLI Gainward GTX 460 1GB @ 800/1600/1900 2x2GB Kingston @ 800MHz 5-5-5-15 2T 
Hard DriveOptical DriveOSMonitor
WD 250GB, 320GB SATA/3, 16MB Cache, Seagate 1TB LG GSA-H62N 18x SATA Ubuntu 9.10 x86 & Win7 x86 Asus VW222U 
KeyboardPowerCase
Logitech Classic Corsair 650HX NZXT Apollo Black 
  hide details  
Reply
post #6 of 7
Flog that dead horse Driver Heaven!
PWNzershreck
(15 items)
 
  
CPUMotherboardGraphicsRAM
4930K @ 4.6 GHz ASUS Rampage IV Black Edition MSI GTX 1080 FE Heatkiller Acetal 16 GB Corsair Vengeance 1600C9 
Hard DriveOptical DriveCoolingOS
2x Samsung 840 Pro  ASUS DVD-RW SATA Koolance 380i & 2x HW Labs 480GTX Arch Linux x86_64, Windows 7 x64 
MonitorKeyboardPowerCase
LG UC88-B Ultrawide, ASUS VS278Q Ducky Corsair AX1200i Caselabs STH10 
MouseMouse PadAudio
Logitech G500 Func 1030 ASUS Xonar Essence STX 
  hide details  
Reply
PWNzershreck
(15 items)
 
  
CPUMotherboardGraphicsRAM
4930K @ 4.6 GHz ASUS Rampage IV Black Edition MSI GTX 1080 FE Heatkiller Acetal 16 GB Corsair Vengeance 1600C9 
Hard DriveOptical DriveCoolingOS
2x Samsung 840 Pro  ASUS DVD-RW SATA Koolance 380i & 2x HW Labs 480GTX Arch Linux x86_64, Windows 7 x64 
MonitorKeyboardPowerCase
LG UC88-B Ultrawide, ASUS VS278Q Ducky Corsair AX1200i Caselabs STH10 
MouseMouse PadAudio
Logitech G500 Func 1030 ASUS Xonar Essence STX 
  hide details  
Reply
post #7 of 7
So I took the first BMP screenshot. I saved them as PNG (which is a lossless format, for the uninformed). God knows why they decided BMP is superior.

In any case:
The original ATI image PNG (1.75MB), BMP (7.91MB)
ATI image after Levels PNG
(you can clearly see the 0-15 and 236-255 ranges were clipped in the reduction in filesize)
Code:
The script used for conversion:
imagesource("ati1.bmp")
Levels(16, 1, 235, 0, 255, coring=false)
NVIDIA image PNG (1.5MB), BMP (7.91MB)

What the NVIDIA image could have looked like with color correction enabled:
ATI image after Levels, tweaked (to prevent detail loss): PNG

Compare the ATI image after levels to the NVIDIA image and you'll see that there's no difference aside from the NVIDIA image being redder.
(I know Craig's collar partially "merges" with the shirt, but this is most likely because I had to convert to YV12 to tweak it. If I had the original video, this would've probably not happened.)

Also, I really don't think the entire Levels conversion thing was done by the NVIDIA card. It's more likely to have been done by the software decoder (even if the decoding was hardware assisted).
It's usually a postprocessing filter.
Akiyama Mio
(13 items)
 
  
CPUMotherboardGraphicsRAM
E6420 @ stock, 0.98v Asus P5N-E SLI Gainward GTX 460 1GB @ 800/1600/1900 2x2GB Kingston @ 800MHz 5-5-5-15 2T 
Hard DriveOptical DriveOSMonitor
WD 250GB, 320GB SATA/3, 16MB Cache, Seagate 1TB LG GSA-H62N 18x SATA Ubuntu 9.10 x86 & Win7 x86 Asus VW222U 
KeyboardPowerCase
Logitech Classic Corsair 650HX NZXT Apollo Black 
  hide details  
Reply
Akiyama Mio
(13 items)
 
  
CPUMotherboardGraphicsRAM
E6420 @ stock, 0.98v Asus P5N-E SLI Gainward GTX 460 1GB @ 800/1600/1900 2x2GB Kingston @ 800MHz 5-5-5-15 2T 
Hard DriveOptical DriveOSMonitor
WD 250GB, 320GB SATA/3, 16MB Cache, Seagate 1TB LG GSA-H62N 18x SATA Ubuntu 9.10 x86 & Win7 x86 Asus VW222U 
KeyboardPowerCase
Logitech Classic Corsair 650HX NZXT Apollo Black 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Hardware News
Overclock.net › Forums › Industry News › Hardware News › [DH] Avivo HD Vs Purevideo HD