USB Video Grabber Test
Many names
The device is sold under many names. EZ grabber, EZgrabber, EasyGrabber, STK1160, STK1150, LogiLink
Videograbber USB2, ezcap168,
The USB Video Grabber
Many buyers have trouble grabbing videos with the EasyGrabber. So did I. Since help was not available at the internet I wrote this test report. In Part 2 you get a tutorial for grabbing with the EasyGrabber.
Strengths and weaknesses
+ unbeatable cheap in price good image quality for VHS videos, absolutely sufficient no noise visible adequate sound quality for normal VHS tapes and VHS copies |
– captures only with 25 fps or 30 fps and throws together two frames without deinterlacing. (Subsequent deinterlacing necessary, otherwise strong comb artifacts visible in fast moving picture parts) no video visible during fast forward of VHS tape (Preview) below 60 Hz noise ratio reduces to 60 dB |
Measured values:
Video Image resolution Frame rate optical resolution via composite input optical resolution over S-Video input |
720×576 25 … 30 fps approximately 512×384 (see example below) |
Audio Sampling rate Resolution Frequency response (-3dB) Ripple (100 Hz ... 18 kHz) Noise ratio (40 Hz ... 20 kHz) Noise ratio (60 Hz ... 20 kHz) Stereo Separation (1 kHz, 200 Hz high pass) |
44.1 kHz / 48 kHz 16 bit 40 Hz … 20 kHz < 0.6 dB 60 dB > 75 dB > 56 dB |
For grabbing videos from VCR (VHS) these technical parameters are absolutely sufficient.
What can this grabber do in detail?
Image
The video grabber captures a full PAL image without border losses with 720×576 pixels.
Input signals can be composite (from VHS recorder) or S-Video. VHS video has
reduced sharpness because it uses only composite signals.
The grabber adds no visible noise to the picture.
Live preview
Here are some screenshots. You can compare the capture results of the composite and S-video input with a digital DVD test image. This is not a simulation, I actually recorded these pictures with the EasyGrabber. To show you the quality of a real-world VHS tape I recorded the signal also from DVD to VHS tape and grabbed it from the recorded VHS tape.
As the screenshot of the grabbed picture were compressed to 720 pixels, I have stretched them to 768 pixels.
Live preview of the image quality. Move the mouse over the image descriptions to see how image quality changes! Digital Original
Digital original S-Video from DVD
S-Video, small loss of details of the fabric Composite from DVD
Composite, blurred Composite from recorded VHS
Composite from VHS, blurred, strong loss of details in skin and fabric, below are switching distortions of the video heads
Scaling
The Grabber captures with a resolution of 720 × 576, thus the image scales about 9% to small.
To get back correct proportions, you need to enlarge to 768 pixels
in width, or you have to tell your video encoder that it must
use 720 pixels that are not square (pixel aspect ratio =??
see wikipedia:Pixel aspect ratio ).
Effective visual resolution
With VHS videos already 512 × 384 pixels satisfy the required visual resolution.
All details of a VCR image will be proper reproduced (I tested this via
composite input that was feed from a DVD). This limit is due to the
resolution of VHS, not the grabber.
Among other things, this results from the fact that the video bandwidth of 3 MHz of VHS is
significantly lower than that used for analogue TV signals. An
additional blurr in television arises because the lines partially overlap on screen.
Therefore the real resolution never reaches the number of lines that are transferred(550 lines). This
is important to know when scaling down the grabbed video.
Conclusion
The resolution of the composite input is sufficient to grab VHS videos without loss of sharpness.
A picture from DVD is sharper when recorded via the S-video input compared to the composite input.
But also S-video does not give the full sharpness of a DVD image that is
transferred to TV via separate RGB signals (SCART). This is not an
effect of the grabber, it results from the specifications
of composite and S-Video.
VHS players do not have RGB outputs, since a composite signal is sufficient to provide the full resolution of the recorded VHS signals. SCART connectors you find at VHS players do not carry separate RGB signals.
Limits
The resolution of current analog TV broadcasts is not achieved by the USB video grabber.
Using the USB video grabber makes no sense if you want to grab high-resolution analog material.
Audio
The USB Video Grabber digitizes audio signals in stereo at 44.1 kHz and 16 bit resolution.
Frequency response of the Grabber
The audio frequency response of the grabber is 40 Hz ...> 20 kHz (-3dB).
Ripple is less than <0.6 dB (100 Hz to 18 kHz).
Only with headphones you can hear a faint hum when no audio is present.
This hum can be removed with a 60 Hz high pass that is applied to the signal.
Absolutely sufficient for VHS.
Frequency Measured amplitude of the recorded signal
30 Hz -4,18
40 Hz -2,90 (this lower limit frequency at -3 dB)
50 Hz -2,00
60 Hz -1,49
80 Hz -0,89
90 Hz -0,71
100 Hz -0,58
125 Hz -0,58
160 Hz -0,19
200 Hz -0,10
400 Hz 0,02
1000 Hz 0,00
1500 Hz -0,09
3150 Hz -0,25
5000 Hz -0,26
10000 Hz -0,23
11000 Hz -0,14
12000 Hz -0,05
13000 Hz -0,09
14000 Hz -0,26
15000 Hz -0,45
16000 Hz -0,54
17000 Hz -0,51
18000 Hz -0,52
19000 Hz -0,70
20000 Hz -1,78 (this upper frequency limit of -2 dB)
Measured audio values of the EasyGrabber
Measured noise ratio
Below 60 Hz we have interference signals, which results in an over all noise ratio of 40 dB.
This is already sufficient. An audio tape recorder also achieves only the same SNR (with inactive noise reduction).
HiFi VHS recorders reach about 70 dB Signal to Noise Ratio. Here audio is record with FM modulated HF.
The USB video grabber cannot achieve this high audio quality. If you want to get
hi-fi audio from your tape your need to record audio with a
seperate high-quality sound card. However this will produce delay
between audio and video which you have to remove later.
You can use a simple trick to enhance the SNR of the recorded audio. As the
noise rises below 60 Hz you can filter the audio track by
means of any audio editor with a 60 Hz high pass filter. If
your video contains mainly speech, you will hardly notice any lost
of audio bandwith. By this means the noise ratio improves to very
good 75 dB (Figure below).
If your video contains music, you better use a 40 Hz high pass.
I have never filtered the audio recordings made by the Logilink video grabber, as the minimum noise is hardly noticeable.
After filtering with a 100 Hz high pass, about 75 dB SNR
The channel separation of this grabber is as good as the noise ratio.
I have measured a channel separation of 56 dB (1 kHz test signal, recording subsequently high-pass filtered with 200 Hz ).
Behavior of the grabber when fed with distorted signals
It is interesting what happens when the grabber is fed with a distorted signal. This is more common due to disturbances in VCR tape flow and crease on VHS cassettes.
In some grabbed videos I noticed that the picture sometimes freezes shortly in motion, and then jumped to keep running. There is a plugin for Avisynth that creates a list for all frames of a film, how much they differ from the direct predecessor frame. Actually this plugin has been developed to be able to analyze films from digital copies (DVD, Blueray), whether they have been subjected to an exotic interlacing, since for the conversion between different frame rates (NTSC and PAL) often some frames are doubled. This discrupts the smooth flow of movements.
Normally you cannot apply this tool on VCR recordings. Even with still images (eg a text panel) it will distinguish much difference between two successive, seemingly identical frames because they are superimposed by noise from VHS tape. The noise is so strong that it will rise the difference index between the two identical frames above 1,000. When I attached MultiDecimate to the recorded video, I discovered a lot of absolutely identical frame copies in the image stream. These came not frome the tape. The grabber and the software has inserted them at the points, when the grapper could not catch a correct picture because of disturbances in tape flow. There are artificially generated duplicate frames.
Duplicate fill frames are automatically generated to replace disturbed frames
Differences of successive frames in the list that has been generated by MultiDecimate
Large segments of the video were completely without these duplicate frames. In other sections whole series of up to 7 consecutive frames ware filled with identical digital duplicates.
In the graphic below I have summed up the total of duplicate fill frames over the duration of the video. It is apparent that the number of faults is greater in some sections.
Total amount of filling frames in a grabbed VCR recording
Duplicate frames are not an error or disadvantage of the USB video grabber. Something must be done if the input signal is disturbed. A repeat of the last undisturbed image is acceptable.
It would be a question of enhanced post-processing, to identify those duplicates in the film and replace them later by intermediate images that are automatically calculated from the last valid and next valid frame. This task is not trivial, since up to 7 frames must be replaced by a image sequence. For suggestions, I am grateful.
How to grab video from VCR
At Amazon, there are many disappointed buyers of this unbeatable cheap video digitizer.
My first results also were disappoiting. My videos had dropped frames and strong interlacing artifacts.
The trick is just take care of some things and use the right software. Then you can achieve very good results with this device. The following page helps you successfully digitizing your videos.