XOR42's Defective XA-10 Image Quality
Problem Examples
Updated with boring details for fellow nerds: 2012-01-05 @ 0130 Hours
These examples are arranged in order of worst to best.
The first two are terrible in normal/casual/lower room light, measured as EV 4-6 (soft,
not dark).
The last two show phenomenal color rendition in bright overcast daylight, EV
10-13.
But, even in the best two, if you ignore superb color/contrast,
you'll see the XA-10 still sands out detail!
Subjective Testing Equipment and Protocols
Camera | Device | Canon XA-10. Native Lens. Mid-Zoom for most clips. |
Purpose | Camera under test | |
Calibration | Factory Reset performed to defaults before test sessions. | |
Resolution | 1920x1080 for Hi-Q test. 1440x1080 Anamorphic
for mid-Q tests. Those settings are not selectable. They default to quality mode. |
|
Connection | N/A. Image rays input via fixed mount zoom lens. | |
Rendering | N/A. Godess of Nature provides virgin images. | |
Display 1 | Device | Pioneer 1140HD Plasma 50" Monitor = Excellent! |
Purpose | Subjective visual testing and review | |
Calibration | Color matched to test targets via advanced menus. | |
Resolution | 1365x768, HD = 720i Native = Oustanding | |
Connection | Canon HMDI output | |
Rendering | Canon HDMI engine | |
Display 2 | Device | Dell 24" 2405FPW LCD Graphics Monitor. |
Purpose | Subjective visual testing, review, and screen capturing | |
Calibration | Monthly calibration via Datacolor Spyder 3 | |
Resolution | 1900x1200 @ 75 Hz | |
Connection | Digital Video Interface (DVI) | |
Rendering | PC (See Computer) | |
Capture SW | SnagIt Screen Capture Utility | |
Display 3 | Device | Samsung SyncMaster 24" 2343BWX LCD |
Purpose | Subjective viewing on pedestrian monitor. | |
Calibration | Monthly calibration via Datacolor Spyder 3 | |
Resolution | 2048x1152 @ 60 Hz | |
Connection | Digital Video Interface (DVI) | |
Rendering | PC (See Computer) | |
Computer | Device | Generic Intel I7 Quad Core, 3.2 Ghz, 8 threads. |
Memory | RAM=8 GB, DASD=6 TB (eSata, Firewire, USB) | |
Media Player | WMP 12.0.7601.17514 | |
Capture SW | SnagIt v9.1.3 Screen capture utility | |
Method | Video Source | Recorded on XA-10 |
Record Parms | See each sample cell below for lighting and recording details. | |
Delivery | Clips copied to PC (eSata drive) | |
Playback | WMP Renders clips to Dell LCD | |
Selection | Exemplary frame chosen that matches dynamically viewed content Key frames preferred to ensure minimal interpolation errors. |
|
Capture | SnagIt used to capture selected frames at 1:1 ratio. | |
Processing | NONE. Image only cropped. No tweaks. File Q=Max. | |
Comparison | Plasma and LCD images compared. Only those equal are used. Selections compared to Sony XR550V and CX700V cameras. Also compared to online but YouTube is too poor for this level of test. |
Sample 1: Example of very heavily smoothed and blocked
up surfaces like hands, paper, rug, etc. Like old VHS tape. Lighting: Approximately EV 5-6 (soft, even lighting, 300W ceiling lamp + 6 x 60W spots, dimmed no kore than 75%) Camera Mode: Manual. Factory reset performed. Exposure Mode: Auto. Factor Defaults. Periodic inspections ~ 1/30 @ F2.0 Gain 9 db. Record Mode: Default FP+ 12 mbps bitrate. Framerate: 60i. Capture/Crop: 1:1. No resize. Only cropped to problem area so as to fit this table. No quality lost. ![]() |
Sample 2: Example of very heavily smoothed and blocked
up surfaces like hands, paper, rug, etc. Like old VHS tape. Lighting: Approximately EV 5-6 (soft, even lighting, 300W ceiling lamp + 6 x 60W spots, dimmed no kore than 75%) Camera Mode: Manual. Factory reset performed. Exposure Mode: Auto. Factor Defaults. Periodic inspections ~ 1/30 @ F2.0 Gain 9 db. Record Mode: Default FP+ 12 mbps bitrate. Framerate: 60i. Capture/Crop: 1:1. No resize. Only cropped to problem area so as to fit this table. No quality lost. ![]() |
Sample 3: Moderate Kitchen Light: Very Poor
Detail / Smoothing of Skin/Hair Detail Lighting: Approximately EV 5 (moderate even lighting, 8 x 40W overhead pot lights, not dimmed). Camera Mode: Manual. Factory reset performed. Exposure Mode: Auto. Factor Defaults. Periodic inspections ~ 1/30 @ F2.0 Gain 12 db. Record Mode: Maximum MXP 24mbps. Framerate: 60i. Capture/Crop: 1:1. No resize. Only cropped to problem area so as to fit this table. No quality lost. ![]() |
Sample
4: Moderate to Full Kitchen Light: Poor
Detail / Smoothing of Skin/Hair Detail Lighting: Approximately EV 6 (moderate even lighting, 8 x 40W overhead pot lights, not dimmed). Camera Mode: AUTO. Factory reset performed. Exposure Mode: Auto. Factor Defaults. Periodic inspections ~ 1/60 @ F2.0 Gain 6 db. Record Mode: Maximum MXP 24mbps. Framerate: 60i. Capture/Crop: 1:1. No resize. Only cropped to problem area so as to fit this table. No quality lost. ![]() |
Sample 5: Very Good (Not quite Superb) in Bright
Indoor Window Light Example of this unit's best quality as seen indoors but in bright lighting form picture window. It looks really great at first but a closer look reveals unwanted smoothing evan at daylight levels. Lighting: Indoors with bright overcast window. Approximately EV 8. Camera Mode: AUTO. Factory reset performed. Exposure Mode: Auto. Factor Defaults. Periodic inspections ~ 1/60 @ F4.0 Gain 0 db. Record Mode: Maximum MXP 24mbps. Framerate: 60i. Capture/Crop: 1:1. No resize. Only cropped to problem area so as to fit this table. No quality lost. Result: Image appears smeared by smoothing. ![]() |
Sample 5: Excellent (Not quite Superb) in Bright
Outdoor Light Example of this unit's best quality as seen outdoors in bright overcaste day. It looks really great at first but a closer look reveals unwanted smoothing even at this bright level. Lighting: Outdoors, bright overcast day. Approximately EV 12. Camera Mode: Manual Factory reset performed. Exposure Mode: AUTO. Auto. Factor Defaults. Periodic inspections ~ 1/125 @ F5.6 Gain 0 db. Record Mode: Maximum MXP 24mbps. Framerate: 60i. Capture/Crop: 1:1. No resize. Only cropped to problem area so as to fit this table. No quality lost. Result: Even at this maximum light level, skin still being smoothed. ![]() |