[X] My Assistant

I have one of those, assuming that's got an em286x chip in it. Mine's Roxio branded but it's the same thing. Also like the Ezcap DC60+.

Apparently there are four entirely different variants of this device, [linuxtv.org] https://linuxtv.org/wiki/index.php/Easycap. The last version of windows that supported my card was Windows 7, but I remember finding some other generic driver that made it work in Win8. It's always worked "out of the box" for me in Linux distros.

Works in linux, and it's actually pretty good for capturing 60FPS 240p video when I run it through a couple ffmpeg filters to separate the fields into frames and correct the aspect ratio. It is annoying to configure for the first use, though.

I don't usually use it for audio capturing, just for video, but I suppose I could use it for audio on my laptop, actually. Thanks. I forgot about mine.

And the use case is mixing sound with my PC's audio for my USB headset so I can play Mega Drive games while appearing to be in class on zoom.

let me know how that goes if you mess with it in linux, currently mine does not like to do audio capture anymore (nothing but static on the audio input that shows up) and I cant really figure out if i can blame the drivers or this terrible hardware.
luckily the one usecase I had of watching output of tv box on pc is now covered with a new equally shoddy hdmi capture stick, which does have working audio.

I often forget how competent people on electric matter we have here. One thing has been bugging me for months is what is an "euro module coaxial socket"?

Pic:


I ripped one of those from my antenna when moving into my current flat and substituted by normal SAT coaxial cables (RG6 - some people call it F type). And my reception has not been particularly good since then - given that I live on the line of sight from the antenna transmitter (on a huge hill) I find it silly that cannot get a great reception.

For all purposes this "euro module" looks just like a coaxial cable extension. But that extra thing behind it, is there some form of filter? i.e. did I rip off the filter that reduces extraneous frequencies into the TV?

I often forget how competent people on electric matter we have here. One thing has been bugging me for months is what is an "euro module coaxial socket"?

Pic:


I ripped one of those from my antenna when moving into my current flat and substituted by normal SAT coaxial cables (RG6 - some people call it F type). And my reception has not been particularly good since then - given that I live on the line of sight from the antenna transmitter (on a huge hill) I find it silly that cannot get a great reception.

For all purposes this "euro module" looks just like a coaxial cable extension. But that extra thing behind it, is there some form of filter? i.e. did I rip off the filter that reduces extraneous frequencies into the TV?

it's usually not all that exciting whats actually in these: (IMG:[1.bp.blogspot.com] http://1.bp.blogspot.com/_8yDjtDOx9ws/TFt0DqJJP8I/AAAAAAAAFCw/z6nSWUh7aX4/s1600/3-way.JPG)

whats inside these boxes is usually just a couple resistors, sometimes a slightly more advanced filter. this is only needed because we these days hook tons of stuff up to the same antenna connection. this poses a problem as not all of these will have the same impedance(read as: resistance of device's input). without these splitters/connectors to 'match' impedance one device would ruin the signal for anything else down the line. so most likely all thats in this wall port is just a couple resistors. maybe some caps for further filtering.

so if you removed these and are now directly connecting your antenna in to devices, might want to splice some of these in again. I grossly glossed over a ton of details(I dont know enough about the black radio magic of antennas) but thats the jist of why these splitters have magical metal boxes behind them and why you usually dont want to get rid of them. its also perfectly possible it had some simple form of signal amplification in it, but usually that'd be labled.

it's usually not all that exciting whats actually in these: (IMG:[1.bp.blogspot.com] http://1.bp.blogspot.com/_8yDjtDOx9ws/TFt0DqJJP8I/AAAAAAAAFCw/z6nSWUh7aX4/s1600/3-way.JPG)

whats inside these boxes is usually just a couple resistors, sometimes a slightly more advanced filter. this is only needed because we these days hook tons of stuff up to the same antenna connection. this poses a problem as not all of these will have the same impedance(read as: resistance of device's input). without these splitters/connectors to 'match' impedance one device would ruin the signal for anything else down the line. so most likely all thats in this wall port is just a couple resistors. maybe some caps for further filtering.

so if you removed these and are now directly connecting your antenna in to devices, might want to splice some of these in again. I grossly glossed over a ton of details(I dont know enough about the black radio magic of antennas) but thats the jist of why these splitters have magical metal boxes behind them and why you usually dont want to get rid of them. its also perfectly possible it had some simple form of signal amplification in it, but usually that'd be labled.

Those look like ferrite beads to me, not resistors.

let me know how that goes if you mess with it in linux, currently mine does not like to do audio capture anymore (nothing but static on the audio input that shows up) and I cant really figure out if i can blame the drivers or this terrible hardware.
luckily the one usecase I had of watching output of tv box on pc is now covered with a new equally shoddy hdmi capture stick, which does have working audio.

Is yours em28xx-based? Because like I said there are actually four varieties that all use different drivers.

If you're just getting static, is it possible it's just a little endian/big endian mixup somewhere?

For me:

CODE

ffmpeg -acodec pcm_s16le -f alsa -i hw:1,0 -f flac outfile.flac

might be a good way to test that. Replace 'pcm_s16le' with 'pcm_s16be' and see what happens/changes. Obviously hw:1,0 is just one possible device name. If you use pulse, all bets are off.

This post has been edited by dragontamer8740: Aug 29 2020, 18:46

Those look like ferrite beads to me, not resistors.

im afraid they actually are, which makes everything I mentioned about those things being for stopping low impedance devices utterly worthless, if thats also in that euro plug it might just be the increased emi now that you removed the ferrite beads from the picture that now causes worse signal.

Is yours em28xx-based? Because like I said there are actually four varieties that all use different drivers.

If you're just getting static, is it possible it's just a little endian/big endian mixup somewhere?

For me:

CODE

ffmpeg -acodec pcm_s16le -f alsa -i hw:1,0 -f flac outfile.flac

might be a good way to test that. Replace 'pcm_s16le' with 'pcm_s16be' and see what happens/changes. Obviously hw:1,0 is just one possible device name. If you use pulse, all bets are off.

naturally I'm using pulse, and its very likely that something like that was the actual cause. I'll try to remember that the next time I think about opening that box. but with how good this new hdmi one is working I might consider a terrible analog->hdmi->hdmicapture card solution first if a situation comes up

im afraid they actually are, which makes everything I mentioned about those things being for stopping low impedance devices utterly worthless, if thats also in that euro plug it might just be the increased emi now that you removed the ferrite beads from the picture that now causes worse signal.

naturally I'm using pulse, and its very likely that something like that was the actual cause. I'll try to remember that the next time I think about opening that box. but with how good this new hdmi one is working I might consider a terrible analog->hdmi->hdmicapture card solution first if a situation comes up

When I last used this device, I specifically remember having to kill pulseaudio to get any sound out of it. Pulse likes to respawn, too, so be sure it's good and dead.

Pulse likes to respawn, too, so be sure it's good and dead.

whenever something tries to bother pulse while it isnt running the pulseaudio.socket and systemd magic setup will cause it to automatically start again. which my volume indicator happens to do very frequenly. so I cant really not run it.
im pretty sure I could trick ffplay into playing the devices audio with the correct codec though.

This post has been edited by cate_chan: Aug 29 2020, 20:59

whenever something tries to bother pulse while it isnt running the pulseaudio.socket and systemd magic setup will cause it to automatically start again. which my volume indicator happens to do very frequenly. so I cant really not run it.
im pretty sure I could trick ffplay into playing the devices audio with the correct codec though.

Isn't is just:

/etc/pulse/client.conf

CODE

autospawn = yes

Most distros have that these days.

BTW, thanks for the help guys, Now I have some more testing fun ahead of me - there are even ferrite clips these days.

Isn't is just:

/etc/pulse/client.conf

CODE

autospawn = yes

Most distros have that these days.

a fair guess but arch decided it'd be more 'simple' to set autospawn = no and having:
/usr/lib/systemd/user/pulseaudio.socket

CODE

[Unit]
Description=Sound System
ConditionUser=!root

[Socket]
Priority=6
Backlog=5
ListenStream=%t/pulse/native

[Install]
WantedBy=sockets.target

This post has been edited by cate_chan: Aug 29 2020, 21:47

whenever something tries to bother pulse while it isnt running the pulseaudio.socket and systemd magic setup will cause it to automatically start again. which my volume indicator happens to do very frequenly. so I cant really not run it.

Yes you can; I have done so. You don't even need to remove systemd-as-init (although I did that too later on).
in /etc/pulse/client.conf (on my machine):

CODE

autospawn = no
daemon-binary = /bin/true

Just `autospawn` isn't enough though. You also need `daemon-binary` set.

im pretty sure I could trick ffplay into playing the devices audio with the correct codec though.

Probably.
ffplay -f alsa -acodec pcm_s16le -ar 48000 hw:{whatever} (or dsnoop:{whatever} maybe)

Looks like AC '97 audio is supposed to be MSB-first (little-endian). So if your chip uses AC '97 internally (many do and have a separate chip for the AC '97 codec), it's likely little-endian.

Alternately try pcm_s16be, for instance, for 16-bit big endian. Also experiment with -ar {number}, like -ar 48000 or -ar 44100.

You can get a large list of PCM-like formats ffmpeg can decode with with:

CODE

ffmpeg -codecs|grep -i '^ D.*pcm_'

if you can tell me exactly what chipset your capture device is using (or what kernel module(s)), I can try to find a datasheet for the chip and tell you what formats to try first.

Another thing you may want to do is try setting the 'card' parameter while loading the em28xx module, if your card is actually em28xx-based. My card is recognized as type 9.

---------

UPDATE:
Just plugged in my card and tried it; what I suggested worked for me, but I first had to do:

CODE

v4l2-ctl -d /dev/video2 -c mute=0

…Where /dev/video2 is of course the USB capture card. Looks like it's muted by default (according to `v4l2-ctl -d /dev/vide02 -L`). After doing that, I was able to run:

CODE

ffplay -f alsa -acodec pcm_s16le -ar 48000 hw:3,0

.
Using one of the automagically-created `dsnoop` devices also worked, e.g. 'dsnoop:CARD=Device,DEV=0', where that string was obtained from `arecord -L`.

Pictures don't indicate working audio, but I swear it works too.


My full command-line, since I also separate the 480i fields back into 240p "frames" and dump output to a file while also playing back a preview:

CODE

ffmpeg -f v4l2 -timestamps 2 -i /dev/video2 -f alsa -acodec pcm_s16le -ar 48000 -async 1 -i plughw:CARD=Device,DEV=0 -vf "setfield=tff,separatefields,scale=640x480:flags=neighbor,setdar=4/3" -vcodec libx264 -acodec libvorbis -ab 128k -f matroska filename.mkv -acodec pcm_s16le -vcodec rawvideo -f nut - -y | ffplay -

mplayer can be substituted for ffplay above, of course.
This was after I had changed the frame width in v4l2-ctl as well, to make it a 4:3 PAR.

CODE

#ntsc
v4l2-ctl -d /dev/video2 --set-standard ntsc
# input 1 for s-video, 0 for CVBS
v4l2-ctl -d /dev/video2 --set-input 0
v4l2-ctl -d /dev/video2 --set-fmt-video pixelformat=YUYV
# width is 720 by default, changed for 4:3 PAR
v4l2-ctl -d /dev/video2 --set-fmt-video width=640
# bump up "sharpness" a wee bit for CVBS. range 0-15.
v4l2-ctl -d /dev/video2 --set-ctrl sharpness=7
# unmute
v4l2-ctl -d /dev/video2 --set-ctrl mute=0

This post has been edited by dragontamer8740: Aug 30 2020, 09:39

CODE

daemon-binary = /bin/true

I laughed, love it.

pcm_s16le

yes this did the trick for me as well with ffplay, mine was also using the em28xx driver as it was em2861 based. though I remembered having a perfectly working line in on this pc, and the hdmi capture card works fine as well. its still good to know it works.

mplayer can be substituted for ffplay above, of course.

found that ffplay has an ever so slightly smaller delay on the video than mplayer/mpv, even with all the low latency and performance options available. which really shows quickly when watching content featuring people talking.

cheers for the detailed process of what you did, not many people seem to use those capture devices anymore(or at all under linux)

I laughed, love it.

If it works, it works.

yes this did the trick for me as well with ffplay, mine was also using the em28xx driver as it was em2861 based. though I remembered having a perfectly working line in on this pc, and the hdmi capture card works fine as well. its still good to know it works.

Good, glad it helped.

found that ffplay has an ever so slightly smaller delay on the video than mplayer/mpv, even with all the low latency and performance options available. which really shows quickly when watching content featuring people talking.

cheers for the detailed process of what you did, not many people seem to use those capture devices anymore(or at all under linux)

That could well be. I have never been able to make mine have very much under a second's latency.
I can't even find a driver for this card above Windows 8, actually. I haven't tried recently but I suspect I might be able to just add my card's vendor and product IDs to an .INF from another seller of em28xx devices.

As it stands, my newest windows machines run XP and 98 (not SE).

Fun fact I discovered while reading the datasheet for the chip: em2861 doesn't actually exist; it's just an em2860 with an audio source connected (an AC '97 codec chip, like emp202). Without the codec chip it can't capture audio and will simply report itself as an em2860. 2861 is a purely software distinction.

This post has been edited by dragontamer8740: Aug 31 2020, 00:57

As it stands, my newest windows machines run XP and 98 (not SE).

They eventually fixed the audio latency issues for Windows 7 onwards.

I switched from Windows XP after melting some ram after using patches from a Chinese dev to get it to work with 16 gigs of ram with PAE.

I need to learn how to disassemble this thing before the dust ruins it.

work

kimochi kimochi

work

kimochi kimochi

oh god the general chat plague is spreading!

They eventually fixed the audio latency issues for Windows 7 onwards.

I switched from Windows XP after melting some ram after using patches from a Chinese dev to get it to work with 16 gigs of ram with PAE.

we're not talking about the same thing.

I'm talking about some inherent lag in my capture card between when it receives a picture/audio sample and when it is relayed to my screen. That happened in windows 7 as well, it's a hardware issue.

On the subject of windows 7 audio though, windows vista and 7 ruined MIDI by removing the built in mapper.

I need to learn how to disassemble this thing before the dust ruins it.

You do that by saying "fuck it" and doing so. Assuming "this thing" means a laptop.

Some laptops have publicly available service manuals that can make it easy and remove a lot of guesswork, though.

This post has been edited by dragontamer8740: Aug 31 2020, 20:48

I need to learn how to disassemble this thing before the dust ruins it.

depending on your device it might be a better idea to blow it out in one piece, I've slightly ruined the look of more things than I want to remember taking them appart and breaking little 'one time snap-in' plastic clips holding cases together.

kimochi kimochi

I wish the general experience of dealing with technology was like this