Ezcap HD Video Capture device any good? HDMI/Component

[DB83]

well I am somewhat lost on what exactly is being touted now.

The thread started by mentioning a HD device which is not listed on the official vendor's site.
Now we seem to be talking about a SD device which MIGHT be an upgrade from the official 116 except, if memory serves, that is also NOT listed on the official vendor's site.

[squallmx]

It´s about the EzCap HD (280), a China HDMI portable capture card that looks like a clone of the "Hauppauge HD PVR Rocket Portable Stand Alone" minus the PC Mode, the user DB83 brought this device and a older SD model, he is currently comparing the SD model vs the HD-PVR Rocket using SD sources, unfortunately I don´t have Depositfiles account so I can not watch the resulting files, but looks like the Rocket can capture from SD sources but the quality is not as good as with an old fashion analog capture card with Component input + Lossless compression.

There is another Chinese device from a nameless company with similar specs to the "Avermedia Lve Gamer Portable" sans PC Mode, that claims to have a internal HDCP stripper, so in theory you can capture directly from a PS3.

The lack of proper samples for both chinese items is the issue for me.

[ezcapper]

Got it squallmx, when the ezcap 280 comes in, I will upload to youtube so you can see it. For anyone who is interested, the avermedia clone is called the kingeda usb02. I must admit I am partial to ezcap, so don't know if I'll get this one, but here's the order site:
http://china.tradeford.com/cn337245/hdmi-hdcp-1080p-60hz-video-game-capture-hd-box-for...-_p406045.html
kingeda also makes a magewell hdmi dongle clone called the kingeda usb100hdmi. Like the magewell, it will is compatible with usb 2.0 by lowering the resolution, or will be full resolution with usb 3.0. Last but not least is the very unique device in the next link. I can honestly say I've never seen anything like this. Basically ezcap hdmi version, it's called HD Video Grabber (Kingeda usb101?). If they had just branded it as an ezcap, I would have definitely bought it:
http://www.aliexpress.com/store/product/2014-New-Arrival-HD-Video-Capture-Recorder-Box...031336514.html
side note: Kingeda makes an lgp lite clone as well called the kingeda usb01 that is available on ebay for $90:
http://www.ebay.com/sch/i.html?_from=R40&_sacat=0&_nkw=hdcp%201080p%20box&rt=nc&_pppn=...lo=60&_udhi=90

[DB83]

So, in a nutshell, the OP wants to tout anything and everything his company sells - official or not.

And just bring on a friendly 'brother' who happened to register a little later than he did to make the whole thread look genuine.

Me sceptic ? Nah !!!!

[usually_quiet]

I'm not sure I understand the file size argument. When I use MJPEG, the video bitrate is about 18mbps. With the h264 box, the max bitrate is 18mbps. The file sizes are very similar. Recording below that with either does not give good enough quality for me. If you reencode either, the result is not very good, so with either, it is best to keep the original file. H264 especially is not considered an intermediate codec, so I would imagine there would be more quality loss on the second encode. If I have time, I will do a second reincode test. I will use the program SUPER. There is a big difference between the EZcap 148 type devices than the dc60 type. Alot of the color problems mentioned earlier have been fixed

First you aren't capturing SD video at 18 mbps using the Rocket. If the Colossus is any example it won't let you do that.

Second, the Rocket isn't designed to do lossless H.264 encoding, which would be equivalent to lossless MJPEG. It was made for a different purpose. People buy something like the Rocket that does H.264 hardware compression precisely because they want to use H.264 for smaller file sizes, and don't want to re-encode. They almost certainly want to use High Profile, not Baseline Profile, and they would probably not try to use 18 mbps average VBR or CBR unless they were recording HD video.

Only a moron, or you apparently, would buy the Rocket intending to capture SD video using a H.264 baseline profile at 18 mbps average bitrate or constant bitrate. You are only doing a quality test against the Rocket because your intent was to create a skewed result for the purpose of deceiving the public.

A valid quality test would be a comparison between your EzCrap 148 and another device that could employ the same kind of MJPEG compression. A valid quality test would be to use your EzCrap 148 and H.264 software compression in real time with similar settings to those normally being used by the Rocket. ...but you aren't interested in performing a valid quality test, only a test that makes the product you are trying to sell look better.

Okay i posted the comparison files. A couple things you mentioned should be corrected a little bit. The Hauppage can do 18mbps on the max setting, but it only captures VBR so it varies a little based on the source file, but is always close to 18 (16-19mbps) on the max setting, which is labeled 18mbps. It cannot record higher than the h264 baseline profile.

I went ahead and did your "valid quality test" and set the ezcap 148 settings as similar to the rocket as possible. But I have a 5 year old laptop with a processor 1/5th the power of modern day intel core i5 etc processors, so I could not encode the audio in realtime. I will concede that the audio might be a little better on the hauppage, I can't tell much of a difference. In my opinion, this isn't as good of a test as the MJPEG one, because encoding with h264 in realtime, I had to use all the lowest quality settings (ultrafast baseline 3.0) just to get it not lag.

I did 18mbps. I captured a 5 minute clip from a feature film called Valkamma. It is Creative Commons license Share Alike. I tried to include a portion that had motion (the car scene) and faces at a close, medium, and far distance. The ezcap file is very slightly bigger because mostly because the audio is not compressed. Both the files look poor to me, but I slightly prefer the ezcap to the rocket. For example, when it cuts to the young man on the bench, the rocket file has his face very pixelated, but the ezcap file looks more like a human face. I think most people would say that the ezcap file looks a little better.

The ezcap can produce much better quality files at this bitrate by using a different codec. MJPEG is much lighter on the cpu and produces a much better picture quality at these bitrates. The rocket is set to it's highest quality settings and the ezcap is set to it's lowest. For my money, the ezcap still has better picture quality.

I also included an optional 4th file that shows the composite input quality of the generic 148, from a football game on tv, encoded realitime with your h264 18mbps valid quality test settings. I did this to show that they have improved composite input quality quite a bit since the d60. Remember the 148 is a $17 device.

Rocket file (720.56mb):
https://depositfiles.com/files/a9no729ey

Generic version of EZCAP 148 file (748.77mb):
https://depositfiles.com/files/jdwbquzy1

Original video clip:
https://depositfiles.com/files/8zir00o24

COMPOSITE input generic version of ezcap 148 (1min from football game 138.54 mb):
https://depositfiles.com/files/7vr3pikby

obviously, there is no postprocessing or editing on any of the files.

Both the AVI file and the TS file are darker and less detailed than the original VOB. The AVI file is the darkest.The VOB's framerate is 29.97 fps. The AVI file's framerate 29.97 fps. Strangely the TS file from the HD-PVR Rocket has a framerate of 58.941 fps.

I don't know how you are playing the video for capture by the HD-PVR rocket but either you are using a different method for playback, and a higher frame rate which could account for the pixelation, or the Rocket is set up to double the number of frames.

[ezcapper]

Thanks for the input, that really helped. I realize I had to go to the advanced settings in Hauppage Capture and change the frame rate from "source" to 30fps. I was using a dvd player to play the dvd and captured all the files from that same source. Thanks again for taking the time to watch the clips, it really helped me out.

[usually_quiet]

A second cup of coffee works wonders some days. I figured out what is going on.

The VOB is encoded progressive, but even so, when it uses 29.97 frames per second, the video output by the DVD player is interlaced. Each frame the DVD player outputs consists of 2 interlaced fields. That should give 59.94 frames per second when de-interlaced.

The AVI file is progressive, but shows interlacing artifacts/combing when I play it back. Each frame is being recorded without de-interlacing, giving 29.97 frames per second.

The Rocket is de-interlacing as it captures, to produce progressive video, although the frame rate is off from what I would expect. I think most of the "pixelation" is actually the result of bob deinterlacing, but each frame is also encoded using fewer bits than would be the case at 29.97 frames per second, because there are twice as many frames. So it is actually recording progressive video correctly from an interlaced source, but the other device is not.

Changing the Rocket's frame rate to 30 fps is only going to make the its captures look worse because it will frame decimate as well as perform bob deinterlacing. ...but that is the entire point of this exercise, to make the Rocket look worse. The Rocket's captures would look better if it could capture interlaced video as interlaced video, but I think it is not designed to do that.

[jagabo]

The Rocket is de-interlacing as it captures, to produce progressive video, although the frame rate is off from what I would expect.

This is a known problem with the Rocket. I've seen it mentioned elsewhere.

The AVI file is progressive, but shows interlacing artifacts/combing when I play it back. Each frame is being recorded without de-interlacing, giving 29.97 frames per second.

The OP can fix that by setting x264 to encode interlaced (--tff). The out of phase fields could be rectified with a field delay.

The Rocket's captures would look better if it could capture interlaced video as interlaced video, but I think it is not designed to do that.

That matches what I've read. And it makes the Rocket entirely useless for anything but progressive content. Nobody should be buying the Rocket to record standard definition TV.

[usually_quiet]

And it makes the Rocket entirely useless for anything but progressive content. Nobody should be buying the Rocket to record standard definition TV.

Very few would be buying the Rocket to record SD TV or SD video games for that matter. There are plenty of other devices able to record SD video that cost less. Some are even made by reputable companies that stand behind their products.

Most people who buy the Rocket want the device because it records without a PC and makes it easier to upload 480p. 720p, or 1080p video game captures to YouTube or similar.

[jagabo]

Isn't that what I said?

[usually_quiet]

Isn't that what I said?

Yes, and I'm agreeing with you.

[Edit] I looked at the specs for the Rocket knock-off again. The specs don't list 480i or 576i as supported input resolutions. 480p60 and 576p50 are the lowest supported input resolutions.

[squallmx]

After 2 months my unit finally arrived. Here are my initial thoughts:

1.- The thing is weightless, and looks kinda cheap.
2.- Remote control does not include battery, and there is no info about the battery type needed to function... so... is basically useless...
3.- AC Adapter included is the box is for Europe.
4.- Max Bitrate for 720p is 16 Mbps, for 1080p is actually 18 Mbps (frame-rate agnostic), Baseline@L4.0 AVC, min quantizer is 12, max quantizer is 51.
5.- It can encode HDCP protected signals
6.- Video quality is good at 24/30p 720p and 24 fps 1080p, decent for 30p 1080p, and mediocre for 60p 720p. Videogame content seems to be harder to compress than movies, posterization is common even in low quant frames.
7.- Audio quality is great, true AAC LR Stereo at 192 Kbps (approx), frequency cut is a little bit to aggressive IMHO, does not support Dolby Digital.
8.- No lag when capturing.
9.- Capture Card assumes Complete RGB Levels.
10.- "User Guide" is crappy and not completely accurate.

24p 1080p Sample (From Blu-ray Hard Candy):
https://mega.co.nz/#!M9Ii2KqL!XNK23Bpef8LUlUWPqMp_RhtTgGjtMBRfhJyKAWdFis8

30p (encoded as 60p) 720p Sample (From Final Fantasy X for PS3):
https://mega.co.nz/#!84hzlACQ!RGY1FvyOyJCT9v98eP-CkmfO4FIeFzA7bOt1SuKzUhw

30p 1080p Sample (From Final Fantasy X for PS3), crushed blacks are a bug from the game, not the capture box:
https://mega.co.nz/#!R4ACCZTB!6_QMrw1PvPTAilRE3hka_Zvv69eJ-tz6BKduffmqW0I

I still don't know is this capture card can be connected to a PC for higher bit-rate files...

[jagabo]

Thanks for that report and the samples. It captures as MKV file? Or did you remux or reencode the caps?

The 1080p24 sample had blown out brights, clamped at Y=235. From the other caps it's obvious the device can capture the full Y=0 to Y=255 range. So maybe the Blu-ray player was clamping the video? Or it might be the difference between capturing RGB vs YUV (HDMI can carry both). Are there any proc amp controls?

What happens with 1080i input?

Otherwise they look pretty good.

[turk690]

After 2 months my unit finally arrived...

Sometime after two weeks, I would have tried to trace exactly where the wretched unit is. You either have the patience of a saint or lots of stuff to kill in the meantime or both.

[squallmx]

Thanks for that report and the samples. It captures as MKV file? Or did you remux or reencode the caps?

The 1080p24 sample had blown out brights, clamped at Y=235. From the other caps it's obvious the device can capture the full Y=0 to Y=255 range. So maybe the Blu-ray player was clamping the video? Or it might be the difference between capturing RGB vs YUV (HDMI can carry both). Are there any proc amp controls?

What happens with 1080i input?

Otherwise they look pretty good.

The output files are always MP4, I had to trim the files a little bit, so I re-mux the streams as MKV, no Re-compression was made.

The 24p sample was from a Blu-ray which always is 16-235, I did compare the captured video vs the RAW m2ts, the images are identical sans additional compression defects from the encode, the results were the same in "YUV" and "RGB Complete" modes on my PS3.

24p Sample RGB:
https://mega.co.nz/#!Q85QCawT!dRu8_RghzVf91UX00PXyjZ2jj_Pb486mCptOielvYFY

24p Sample RGB Limited:
https://mega.co.nz/#!pgoFmKaZ!7bzylDPsWnw_RphWpvMdzUD3hzxT1k4D6P2cc505RgU

Source File:
https://mega.co.nz/#!41AU2b5J!h9jehtaZXCLHTad_HwYa8IvgUYGSkk3EJXpX3IFLdbc

1080i content is converted to 1080p 60p/50p, and then the HD Box captures only one of every two frames. Since I dont have a lot of 1080i the amount of testing was rather limited, but deinterlaced quality seems to be good.

Sometime after two weeks, I would have tried to trace exactly where the wretched unit is. You either have the patience of a saint or lots of stuff to kill in the meantime or both.

China Mail is painful slow this time of the year, I like to think that they use swimming donkeys and because of that the shipping time is so slow, the item finally arrived the last day covered by PayPal buyer protection, I was ready to escalate the complaint in order to get my money back.