Streaming from an arcade cabinet is easy!  In this page, we will go through the components needed to obtain the best quality, and to build a modular, multipurpose capture setup that can also be used for capturing arcade boards or consoles.

Warning: This is not what you would call cheap, but it is the best quality you can get in terms of direct capture and is super flexible.  You definitely get value for money.

The secret to great capture – RGB

Cast your mind back to a time when you connected your consoles to a CRT TV.  If you lived in America, chances are that you probably used RF or composite, or if you were really fancy, S-video.

But did you know that Europe and Japan had something even better?  RGB; and it was delivered over something called a SCART cable.  So what makes RGB so good?  In short, it’s the separation of signals and increased bandwidth.

The way S-video works is that it employs a clever technique that converts the image into separate greyscale and colour channels (a bit like layers in photoshop) and sends each part down its own conductor.  The greyscale channel is sent at full resolution with a sync signal, but the colour channel is sent at half resolution due to bandwidth constraints.

RGB on the other hand, splits the image into 3 full resolution channels of red, green and blue, and like S-video, sends each part down its own conductor (in most cases there is a separate sync channel too).  What’s more, for devices that are capable of outputting RGB (eg Sega Genesis), it means you avoid the colour conversion that S-video does.  The result is a higher definition image than S-video and no colour conversion loss.

The other secret to great capture – Avoiding Interlacing

You may be familiar with the terms 480i, 720p and 1080i.  The number tells us how many lines make up the image and the letter at the end tells us what type of image it is. P means progressive and I means interlaced – simple.  But what is an interlaced or progressive image?  Why are they used?  Which is better?  Well, in short, progressive is; but we’ll get to that.

Progressive scan displays draw each frame from top to bottom.  It happens so fast that due to persistence of vision, you can’t even tell.  Unless you are reading this on a CRT TV, your display is almost certainly progressive, and that’s a problem for you and all your viewers if your capture source is interlaced.

Interlacing is a technique designed to increase the perceived frame rate of television while adhering to bandwidth, broadcast and technical limitations of the time.  In interlaced video, we talk about fields rather than frames.  So what is a field?  Quite simply, it’s half a frame made up of the odd or even lines.

Here’s the clever bit.  Because an interlaced display shows these fields one after another, it means you can pack 2 “frames” into 1!

On the left we have an interlaced frame showing two fields.  Click the right image to see how it would appear on a TV!

This means you can transmit 30 of these spliced frames, and because the TV displays the fields one after another, you effectively get 60 half resolution “frames” per second.  It might not look like the best thing ever here, but on an interlaced CRT it looked pretty convincing!

What about interlaced video today?  Viewing such video on a progressive display will give you frames that look like the one on the left.  The bad news doesn’t stop there either.  Interlaced material compresses very poorly, meaning your captures will require more bandwidth and be lower quality as a result.  Interlaced video also doesn’t resize well and will most likely suffer from frame blending as a result.

Fortunately, there is a way to deal with it, but it still doesn’t come close to a progressive capture.  Some capture and stream programs offer the option to deinterlace, and if you are lucky enough to own a video processor or upscaler, that might too.  Deinterlacing will try to reconstruct the image, and there are many different methods of doing this, all with varying success and CPU demands.  For more information on deinterlacing, see Wikipedia.

Both composite and S-video are typically interlaced (480i/576i), but SCART supports progressive (240p/288p) and interlaced (480i/576i).  Fortunately for us, CPS2 outputs a progressive RGB signal that is compatible with SCART so we are able to avoid interlacing entirely.

More about SCART

SCART is a multi-purpose connector that can carry composite, S-video, RGB, data and stereo audio signals.  Since composite video contains a sync signal, the composite pin is also used for RGB sync (sync on green is also supported).

SCART was also used in Japan where it is known as JP21 and although they share the same connector, they are wired differently.  The good news is that both standards are interoperable with a simple converter that corrects the pinouts.

We suggest avoiding flat cables as the shielding is usually not as good

There are a couple of things you should know to avoid any future headaches.

Not all SCART cables are fully wired!
The SCART specification allows for cables partially wired.  This means it is possible to buy cables that only have connections for composite and/or audio.  You can usually tell as these cables will have pins missing in the connector, or the cable will be very thin.  It is important to look for fully wired cables if you want RGB.

SCART is directional
If you ever wondered why there are so many pins in a SCART connector,  it’s because some are for input and some are for output.  That means that if you buy a SCART to RCA cable, it will take audio from a SCART device and output to RCA – you won’t be able to flip it round and use it to send RCA audio into a SCART device.  Fortunately, converters are available with input/output switches.

The capture chain

We’ve decided that the best way to capture is by using RGB, and that SCART will help us achieve that.  Now it’s time to build our capture chain.

We will be using a JAMMA/SCART breakout PCB that goes between the game board and JAMMA harness to provide our RGB SCART feed.  We then the SCART to an upscaler, which converts the analog 240p RGB into 720p or 1080p HDMI.  From there we can send it to the capture device or another display or projector.

Next, we’ll go through each part of the chain in detail.  At each stage of the chain, there will be a link to a recommended product which we have used and an alternative.  This will guide will give you video and audio from the cabinet via HDMI, just as if you were capturing from a modern console.  If you want to do more fancy stuff like adding commentary, player headsets or additional displays, check out the event stream section.

Arcade Cabinet/PCB

This guide is written for the streaming and capture of Super Street Fighter II Turbo using a standard JAMMA cabinet and has been tested to work with the Versus City and linked Astro City cabinets.  This capture method should work for most, if not all CPS2 and CPS3 games, or other 15kHz/240p JAMMA games with RCA audio output.

It is common for Super Street Fighter II Turbo to be played in a head to head configuration by linking two solo cabinets by way of a custom JAMMA harness.  There’s a chance you may experience reduced quality or even drop outs in your capture due to splitting the video signal passively.  You can mitigate this by using short, high quality cables and keeping the JAMMA harness length to a minimum.

High quality upscalers like the XRGB offer ways to adjust sync and signal sensitivity levels which in combination with quality cables, can eliminate drop outs entirely and give near emulator quality capture.

We have had the best results with Versus City cabinets, and mixed results with linked Astro City cabinets.  One setup required some sync adjustment and was solid, but another setup had an unstable video feed despite making adjustments in the XRGB.  We may have just been unlucky and it was an issue with the board’s video output, but we provide this warning and advice regardless.

You can see some sample videos below.

Versus City (good)  Linked Astro City (good)  Linked Astro City (poor)

Recommended – Smallcab JAMMA to SCART PCB
You will also need – A short SCART cableRCA cable, SCART audio block (optional)
Alternative – Retroelectronik JAMMA to SCART PCB

The JAMMA to SCART PCB connects between the game board and the JAMMA harness.  It effectively adds a SCART socket to your game board, allowing you to tap into the video signal without hacking up your JAMMA harness.  The SCART cable then connects to the next part in the chain, the upscaler.

Note that these PCBs do not provide audio via SCART.  Audio must be taken from the RCA jacks on the arcade board and then you can either combine the audio into the SCART cable with an adapter like this or connect the arcade board’s audio out to the audio input of the upscaler or capture card.

For other options or more advanced usage including mixers, see the event stream section

SCART to HDMI Upscaler
Recommended – Micomsoft XRGB Mini (aka Frame Meister) with Euro SCART adapter
You will also need – Travel adapter for Japanese plugs to your socket
Useful links – XRGB mini settings | XRGB mini firmware | CPS2 Stream Overlays
Alternative – Open Source Scan Converter

The next part of our chain is the upscaler, the XRGB mini.  We’ll be using it to take the 240p RGB SCART signal and give us 720p or 1080p HDMI.  The XRGB mini features front RCA audio ports where you can connect the arcade board audio from earlier (front port audio is selectable in the XRGB menu).  We then feed the HDMI into the capture card.

The XRGB mini is the most expensive component here, but it’s far more than a simple SCART to HDMI converter – it’s an analog to digital Swiss army knife!  It has inputs for composite, S-video, SCART, component (via D terminal adapter) and can also function as a HDMI switch.  It was designed to allow gamers to play their old consoles on their new TVs with minimal input lag and far higher quality than scalers built into TVs.  There is also an option to generate scanlines, but these should not be used for streaming as they present similar kinds of issues as interlacing.

You may remember that we mentioned JP21 earlier; well the XRGB mini was designed for use in Japan and uses JP21.  That means that the supplied adapter will not work with European SCART, so make sure you buy the European SCART adapter linked above.  This adapter has a built in sync cleaner that is powered by the USB port on the back of the XRGB mini.  Though the sync cleaner has no effect on arcade capture since the sync is already clean, it can improve the quality of other SCART devices where the sync signal is derived from the composite video signal.

You should connect the XRGB mini to a display rather than a capture card for initial setup.  The default resolution is 480p, so make sure you change it to match the resolution you will be streaming at for the best quality (we suggest 720p).  Some general XRGB settings for 240p input such as arcade boards and consoles can be found in the links above, but other settings will depend on your hardware and preferences.

Settings you will want to pay particular attention to are the AD level and sync level.  The AD level is sort of an analog sensitivity setting.  If your signal is weak and the colours are dull, raising this will help compensate.  On the other hand, if you have a good quality signal and this is set too high, it can cause colours to crush or look over saturated.  Different equipment will have different ideal settings, for example streaming from two linked solo cabinets will require a higher AD level than a single solo cabinet due to the signal being weaker.

If your signal is so weak that it can’t maintain a steady image (rolling or other glitching), adjusting the sync level might help.  We have run into this issue before when streaming from linked solo cabinets and adjusting the sync level produced a steady image.

It is also worth spending some time doing colour correction.  Just as the AD level can vary, so can colour.  We have provided a reference image of the Super Turbo colour bar test from an emulator for the correct RGB values.  There are transparent sections in the image so that you can load it in your stream program and see your capture through it.  Simply put your board into the colour test mode and as you make adjustments, the live feed image bars and reference should seem to blend together.

Left: Reference image overlayed on live capture.  Right: Comparing the board’s colour bars to the reference

For colour correction we will be concentrating on the bottom greyscale bar.  Due to what seems to be a capture card limitation, the primary colours can look off.  We know they are correct because red, green and blue mixed makes white and the greyscale/white bar matches the reference very closely.  If your greyscale/white bar is off colour, adjusting the primary colours should help.  The settings you are most likely to adjust are brightness, black level and gamma.  Adjusting the black level can also hide noise.

The good news is that the XRGB mini supports saving your settings to a microSD card, so you can create profiles for different games or capture setups.

Due to the unusual resolution of CPS2 games, the upscaled image will not fill the frame completely and will have black borders.  While the XRGB does have settings to resize and reposition the image, having borders doesn’t matter for streaming since you can crop them out in the stream program and use the space for other stream elements such as player cameras or stream chat.

In fact, not filling the frame will actually give us higher quality scaling (and a convenient place for player names and scores).  But why?  Let’s briefly talk about image scaling.

The most basic image scaling algorithm, known as nearest neighbor, simply copies pixels.  If you have an image 2 pixels wide and wanted to make it 4 pixels wide, each pixel becomes 2 pixels.  If you made it 6 pixels wide, each pixel would become 3 pixels.  When an image is scaled by a whole number we call this integer scaling.

But what if you were to scale the 2 pixel image to 5 pixels?  You can’t have part pixels, so it’s not possible for each pixel to become 2.5 pixels.  Instead, one pixel becomes 2 pixels, and the other 3 pixels.  This is what’s known as non-integer scaling, and we want to avoid it.

A 2 pixel wide image, scaled 2x, 3x and 2.5x using the nearest neighbour algorithm

We can check if something is an integer or non-integer upscale by dividing the target resolution by the original resolution.  If you wanted to fill a 720p frame with a 224p image, you’d need a scale factor of around 3.21.  That’s a non-integer scale, which results in uneven scaling.

Left: A 3.21x upscale to 720p.  Right: A 3x upscale to 672p.  Uneven scaling is most noticeable in pixel grids like the background and life bars.

But what if we go for the nearest integer – a 3x upscale?  That gives us 672 lines which fills most of a 720p frame and we get even scaling of all the pixels.

Ah, but there is a catch.  For 1080p, there isn’t an integer scale we can perform that will still fill most of the frame.  A 4x scale would be 896 lines, and 5x would be 1120 lines, cutting off some of the image and not leaving much space for names and scores.  We can go for the same image to frame ratio as we did with 720p, which happens to be a 4.5x scale giving us 1008 lines.  It’s not ideal, but at least it’s a uniform, alternating 4 and 5 pixel scale.

We also need to adjust the width of the image, because CPS2 games were designed for use on CRTs and appear too wide on modern displays.  We can calculate how wide the image needs to be from its aspect ratio.  We know that Super Turbo was designed to be played on “square” CRTs (as opposed to widescreen), and the aspect ratio of those is 4:3.  The aspect ratio defines the relationship between the width and height of something, in this case a display or image.

We can use the aspect ratio in a couple of ways.
To find the width:  Divide aspect ratio (4/3), and multiply by the height (672) = 896
To find the height:  Flip the aspect ratio then divide (3/4), and multiply by the width (896) = 672

As you can see from above, we have determined that if we have an image height of 672 pixels, that it needs to be 896 pixels wide for it to be the correct 4:3 aspect ratio.

However, we run into an unavoidable problem with scaling!  CPS2 games are 384 pixels wide, and we need the image to be 896 pixels wide which is a 2.33x scale.  We can’t even pick the nearest integer to scale to because that would cause the image to be the wrong aspect ratio.

We have found a workaround though.  We do the width correction in the stream program rather than the XRGB, as stream programs have better scaling algorithms for dealing with non-integer scales.

But before we do that, we need to resize the image to make sure the XRGB is performing an integer scale like we did for the height.  First we find what the nearest integer scale we can do is.  If you remember what we did earlier, this is the target resolution divided by the original resolution, so in this case 1280 divided by 384, which gives us a 3.33x scale.  Since 3 is the nearest integer, a 3x scale at 384 pixels gives us 1152.

So how do you go about getting perfect scaling in your XRGB?  Integer scaling.  We’ve found that we can do a 3x scale of the width and height to give us 1152×672, which fits nicely in a 1280×720 frame.  Normally it’s a case of trial and error; making adjustments, screenshotting and checking in an image editing program.  Fortunately, we’ve done this for you and provided some stream templates for getting the correct scaling in the XRGB and fixing the aspect ratio in your stream program, so all you have to do is adjust the image until it fits.

About the OSSC
Although we are focusing on the hardware that we have experience with, the OSSC deserves a special mention.  It it open source and significantly cheaper than the XRGB and seems to have even higher image quality.  Sounds awesome right?  So why isn’t it the recommended upscaler?  The XRGB and OSSC have different approaches to upscaling, and in the end it comes down to compatibility.  We recommend the XRGB because of the two, it’s the most likely to work for you.

Remember earlier how the XRGB upscaled the image but left a black border?  That’s called padding.  Because our upscaled image was 672 lines, extra lines are added to give a standard 720p signal.  The XRGB can do this because it has a frame buffer, but it’s the frame buffer that causes the processing delay, said to be about 20ms or 1.5 frames.

The OSSC on the other hand is built for speed and does not use a frame buffer.  This gives the OSSC the incredible latency of less than 2 scanlines (less than 0.15ms) but at a cost.  With no frame buffer, the OSSC is unable to pad the video which means it is possible to output non standard resolutions.  For example, performing a 3x line multiplication on CPS2 games results in a 672 line image and your capture card or monitor may not accept it.  To further complicate matters, CPS2 games run at 59.63fps which is not a standard frame rate.  The XRGB has an option to convert it to standard 59.94fps, but the OSSC is not able to do this.

That doesn’t mean the OSSC is bad, in fact we bought one to replace our XRGB.

We should also mention the generic SCART to HDMI upscalers you can buy from places like Amazon and ebay.  Though they offer little to no control over how the image is scaled, they have very good image quality given the price.  We found that they didn’t work with the Live Gamer Portable, so be aware that compatibility might be hit and miss with certain equipment.  Also make sure that the listing specifies RGB support.  We bought two and they had different PCBs inside.  One supported RGB and the other didn’t.

Capture Card
Recommended – Elgato HD60S (USB 3)
Alternative – Avermedia Extremecap U3 (USB 3) | Avermedia Live Gamer Portable (USB 2)

Capture cards are available as internal PCI cards for desktop computers, or external devices usually in USB form.  Internal cards are generally better and have much more bandwidth available to them than USB alternatives meaning they can support multiple HDMI inputs, higher resolutions, higher quality sampling and may offer lower capture latency.  If you are using a laptop you are limited to external cards, so we’ll focus on those.

When it comes to USB capture cards, they are effectively separated into two classes – USB 2 and USB 3.  Due to the limited bandwidth, USB 2 cards rely on internal video compression to be able to transfer the video to your computer.  Most only support 30fps at 1080p too.  Not only do you lose quality due to the lossy nature of the compression, but there is also a delay between what is happening in game and what is displayed in your stream program due to the compression process.  According to this test, the Live Gamer Portable has around 340ms of delay (around 20 frames).

This can cause other stream elements to be out of sync such as commentary, player cameras or a projector.  Fortunately, the main streaming programs such as OBS and Xsplit offer options to delay stream elements so you can work around the sync issue, but it’s something that can be avoided by using USB 3 or internal capture cards.

Thanks to the extra bandwidth, USB 3 capture cards can capture uncompressed video, and due to not having the internal compression step, have much lower delay – around 50-70ms (3 or 4 frames).

Something to be aware of if you have an older computer, is that some earlier implementations of USB 3 are not full speed.  Notably, first generation Core i3, i5 and i7 CPUs have a 2.5GT/s bus which is half the speed of the USB 3 standard.  In practical terms, it means you won’t have enough bandwidth to capture 1080p at 60fps, but 720p at 60fps should be fine.

If you don’t have USB 3 and own a desktop PC, you can always buy a USB 3 PCI card or internal capture card.  If you are a laptop owner without USB 3, you are stuck with USB 2.  However, if you have an expresscard slot, you’ll be happy to know that you can buy a USB 3 expresscard adapters which add a USB port or two.  We have had good results with the Startech ECUSB3S11 which uses a Renesas/NEC – µPD720202 chipset.  Although the expresscard is rated as 5GT/s, it will work on systems with a 2.5GT/s bus.

Capture cards are also available with different features and inputs.  It’s quite common for capture cards to feature a HDMI passthrough, which clones the input to the capture card so you have a lag free output for your monitor.  Since our players are playing on cabinets, we don’t need the pass through.  This might be something you want to consider if you are streaming other consoles, but a HDMI splitter will do the same thing anyway.

It was also common for capture cards to feature component input, usually through a dongle or breakout cable.  While component isn’t used much today, it might be a welcome feature for streamers who can’t stretch to an XRGB mini since SCART to component converters are more affordable, just be sure to check that low resolutions such as 240p are supported.

Most capture cards have analog audio input too which you can connect the arcade board’s audio or a mixer to.  Be aware that just because a capture card has analog audio inputs, it doesn’t necessarily mean that you can capture from HDMI and the analog audio inputs at the same time – in most cases you can, but it depends on the driver/software.  All the capture cards we have listed do support HDMI with analog audio in.  Don’t let no analog audio input be a dealbreaker though – you can always use the audio in on the XRGB or a microphone/line in on your computer.

Our first capture card was the Avermedia Live Gamer Portable.  Despite the drawbacks of USB 2, one great feature it has over other capture cards is the ability to capture to SD card without a computer!  This is great for people getting started, for capturing gameplay from other setups, or for use as a backup device.  In fact, the first event we captured was with the PC free mode, with scores and overlays added in later in video editing.  The Live Gamer Portable also supports 240p over component, so coupled with a SCART to component converter, can make a cheap offstream capture setup.  Not to be confused with the Lite model which does not have SD card capture or component, but is a very affordable entry into HDMI capture where latency isn’t a concern.

We also have an Avermedia Extremecap U3, which was one of the first USB 3 capture cards.  It was a big step up in quality from the Live Gamer Portable and had far lower latency.  One notable omission is a HDMI passthrough which means for HDMI consoles, you should use a HDMI splitter if latency is a concern.  At the time we were using it, we experienced some audio crackle in Xsplit that goes away after a minute or two.  Format detection did not seem to work, so we had to manually select the resolution in Xsplit.  Changing the framerate had no effect, so it would capture internally at 60.00fps which can cause occasional stutter in 59.94 fps sources.  We contacted Avermedia and they have been working on a fix much to their credit.

Our main capture card is the Elgato HD60S and has been a solid performer.  We should also mention that the software is very good.  If you go into the HD60S configuration within Xsplit, you will find a visual representation of the audio levels which is incredibly useful for setting levels for commentary and avoiding clipping.  The Elgato drivers also feature the ability to clone audio output to multiple devices, so rather than audio coming out of your speakers -or- HDMI output, you can set it to both.  That means you can send your stream to a projector or TV and still be able to monitor the audio on your computer.  Something to note, the minimum requirements state Windows 10, but we are using it with Windows 7.

Streaming for an event

If you are streaming for an event you might want to include commentary, player cameras, a sound system or have your stream displayed on a projector or large display.  Next we’ll go over the components that will help you build a full on event setup.

Splitters and Distribution Amplifiers

Once you get to this stage of your stream setup, it’s likely you are going to run out of outputs and will need to duplicate or split video and/or audio feeds.  There are two types of splitter – passive and active.  Active splitters are also known as distribution amplifiers, and is how we will refer to them in this guide.

Passive splitters take the available signal and share it between the connected devices.  For analog audio and video, it will result in progressively lower quality with each device added.  Video will be darker and noisy, and audio will be quieter and have audible hiss.  For digital signals such as HDMI, you don’t get a progressive loss in quality, however the signal does degrade until it reaches a point where it will start to show corruption or no image at all.  With digital video, you might be able to passively split and run two displays at the same quality as one display, but adding a third could cause image corruption or no image at all.  With analog  signals, you lose quality as soon as you add another device.

Distribution amplifiers take the available signal and amplify it so that each output is a copy of the input.  Because each output is independent, it does not matter how many devices are connected, the quality of each output remains the same.  You may find a little quality loss with analog audio and video compared to the input, but digital devices are effectively a perfect copy.  The only real down side with distribution amplifiers is that they are more expensive than passive splitters and they require power.

Adapters and Converters

Just like with splitters, there are different types of adapters and converters, and you are likely to come to a point where an output and input connector are different.  It is quite common for people to just call both types converters, so you have to be careful and know your stuff!  With the right combination of adapters and converters, you can convert pretty much anything.

Adapters simply change one type of connector to another and do not modify the signal.  You are probably familiar with the 3.5mm jack to RCA cable, or headphones to aux as you may know it.  High end headphones often use a 6.3mm headphone jack and come with a 6.3mm to 3.5mm adapter for use on portable devices.  Adapters are great for when input and output formats match but the connector is different.  Fortunately for us, adapters are generally all we need for analog audio within the scope of capture and streaming.

Converters are used to change one signal format to another.  This is more common in video than audio due to the number of different standards and formats.  The XRGB mini is effectively the converter in our capture chain, taking analog RGB and converting it into digital.  You could swap out the XRGB mini for a different converter if you wanted RGB to component or S-video for example.

Adding Commentary

It’s easy to add commentary to your existing capture setup.  There are a few different ways you can add commentary, but we’ll be covering the use of a mixer and broadcast headsets.

Recommended – Yamaha MG10 with 6.3mm Y splitter
You will 
Alternative – Yamaha MG06,

Mixers are probably the most intimidating looking part of the stream setup if you aren’t familiar with them, but they are actually very easy to use.  In normal operation, once you’ve got everything set up the way you like, you will probably only need to adjust one or two dials if commentators change and they are too loud or quiet.

Its function is simple, to combine many audio inputs into one output.  In our case, we’ll be connecting a couple of microphones and the audio from the arcade board.  The mixer then connects into your capture chain, usually where the arcade board was connected.  We’ll cover the exact hookup later on as there are a few different configurations depending on your preferences and requirements.

We can break our mixer down into 3 sections:  mono inputs (left), stereo inputs (middle) and stereo outputs (right).  The microphones we will be using are mono and will connect to the inputs in the left section, CPS2 output is stereo and connects to one of the pairs of inputs in the middle, and the right section provides the audio output going into your stream.

Mixers are laid out in columns with the input at the top and all the controls for that input in-line.  Let’s take a look at channel 1 on the Yamaha MG10XU.


The first thing you’ll notice is the input.  Usually you will find an XLR and/or 6.3mm jack at the top, but Yamaha use these fancy Neutrik combination sockets which take XLR and 6.3mm jacks.  Your microphone connects here.

Next you’ll find a couple of push buttons.
Pad is a fixed 26db attenuation.  You can use this if your signal level is too high.
HPF is a high pass filter that cuts out frequencies below 80Hz and can improve voice clarity.


Gain controls the input level into the channel.  Cranking it up will make the microphone louder, a bit like a sensitivity setting.  We suggest leaving gain in the shaded area.

Yamaha mixers feature 1 dial compression.  Turning it up compresses the dynamic range, so quiet sounds are louder and loud sounds are quieter.


The group of green dials labeled high, mid and low are tone controls.  High will boost or cut the treble, low the bass, and mid the mid range.  In a setting where you have a mixer connected to a PA speaker, cutting the bass can help improve voice clarity.


Yamaha mixers with the XU suffix feature digital effects such as reverb and pitch change, and this dial controls the amount of the effect.  We don’t need effects so keep it at 0.

Pan controls the position of the channel in a stereo mix.  Putting it all the way to the left will have that channel only be heard through the left channel or speaker.

Level controls the output level of the channel.  This is the dial you will adjust if a different commentator comes on and they are too loud or quiet.  The peak LED will let you know if the signal level is too high.



And that’s all there is to it.  Most of these controls are replicated on each input, meaning that once you understand how one channel works you pretty much understand all of them.  The only real difference is that the stereo channels usually have different input connectors and fewer controls (no gain, no compression).

The stereo channels typically use 6.3mm jacks for their input, though some mixers offer RCA inputs as well, just like our Yamaha MG10XU.  It’s just a matter of convenience really.  It doesn’t matter if a mixer has RCA jacks or not because you can use an RCA to 6.3mm jack adapter or cable.

The output section is very easy to understand.  We’ve got a few different options here, but the only thing we are interested in are the 6.3mm outputs labeled stereo out and phones.  Stereo out will be your mixed game audio and commentary which you then feed back into the capture chain via the XRGB, capture card or line in.

The phones output is a stereo jack and is where you will connect your commentator’s headphones.  Since there is only one phones jack, we will need a splitter.  Fortunately the output from the mixer is powerful enough to drive two pairs of headphones, so we can use a simple Y splitter.

Down in the bottom right hand corner of the mixer, we have some level controls.  The red dial labeled stereo level controls the output level of the stereo out jacks going into the stream.  With this, you can raise or lower the level of the entire mix without adjusting each channel individually.

The white dial labeled monitor/phones controls the level of output to the phones jack – in other words, what the commentators hear.  Adjusting this up or down has no effect on the audio going to the stream.


Setting your audio levels will likely require some trial and error and will depend on your preferences, the microphones you are using and the output level of the CPS2 board (for example this will be lower than normal if you are splitting passively as opposed to a direct connection).  You should refer to your mixer’s manual for the correct procedure.

If you are using the Elgato HD60S capture device, you can open the settings page in Xsplit and use the level meter to assist you.  Xsplit also has a built in level meter in the main window, but it is small and not very detailed.  Another option may be to capture a short video with some test audio and drop it into a video editing program like Adobe Premiere and check the levels from there.

You should set the microphone levels first as the level of the game audio will be relative to the commentary.  If you are using one of the Yamaha MG series mixers, start out by setting the red stereo level dial to the 3 o’ clock position marked by the arrow, then do the same for the white level dial for the channel you are adjusting.  You are now ready to set the gain.  We suggest setting it so that your output to the stream peaks around -6db during normal commentary.  This will give you a bit of headroom so that when your commentators get hype, the audio doesn’t clip or sound distorted.

We found that for our headsets, putting the gain in the shaded area at around 10 o’ clock gave us good results.  Of course, how much headroom you need will depend on how loud your commentators get.  Using compression might help to reduce the difference between talking and hype levels, though compression can cause the pick up of more background noise.  There’s no right or wrong way – you might want clean audio, or you might want to capture some of the ambiance.

Once you have found a level that you are happy with for commentary, you can then set the game audio.  If you check the side of a CPS2 motherboard, there is a test button and volume up and down buttons.  You’ll want to make sure the volume is at the maximum level before adjusting your mixer.  The trick is to find a level that’s audible but doesn’t make the commentary difficult to hear.  We’d suggest something around -18db and adjusting to your liking.

reference levels game sfx tone generator diff mixer or mic same level

Recommended – Micomsoft XRGB Mini (aka Frame Meister) with Euro SCART adapter
You will also need – Travel adapter for Japanese plugs to your socket
Useful links – XRGB mini settings | XRGB mini firmware | CPS2 Stream Overlays
Alternative – Open Source Scan Converter

You’ll need a microphone or two to go along with your mixer for commentary, but there are many different types – hand held, desktop, headset and wearable types that you might have seen TV presenters using.  There are also different technologies such as dynamic and condenser and different pick up patterns.  To top it all off, there are wireless versions too.

We use the Audio Technika BPHS1 headsets which feature a dynamic microphone with a cardioid pickup pattern with XLR output.  But what does all that mean, and how does it apply to you?

Form factor
The type of microphone you pick is an important consideration.  Hand held microphones can be inexpensive and a good way to get into commentary.  They can also offer great sound quality.  One disadvantage is that because the microphone is not a fixed distance from your commentator, you can experience changes in sound level and quality if commentators hold their microphones at different distances or move around.  If you use a hand help microphone, you might want to experiment with holding it close using low gain versus holding it far away with high gain.

Desktop microphones aren’t too different from their hand held counterparts, but can help the issue of commentators holding the microphone too close or not close enough.  However, due to them being farther away from the source of sound, they need to be more sensitive, and so will pick up more background noise.  Also bear in mind that a desktop microphone will transmit vibrations from the desk, so loud music, a noisy computer fan, people putting stuff on the desk and so on.

Headsets offer headphones and a microphone in one convenient package.  The microphone is boom mounted meaning it’s always at a set distance from the commentator, giving you consistent audio levels.  Because it is close to the mouth, it means you need a less sensitive mic (or use lower gain) resulting in less background noise.  Thanks to the microphone being headset mounted, it means you are free to operate the stream.  The only real disadvantages are that they are more expensive than a stand alone microphone.

Lapel, or lavalier microphones as they are sometimes called can be a good alternative.  They clip on to clothing and provide a fixed distance from the commentator for consistent sound.  They are often available in wireless form.  Sound quality may not be as good as other microphones of a similar price due to their small size.

Dynamic VS Condenser
These terms tell us what kind of technology the microphone uses.  Dynamic microphones are simple and work by way of a diaphragm connected to a wire coil which produces a current as the sound waves move the diaphragm and in turn, the coil over a magnet.  It’s quite similar to a speaker, just in reverse.

Condenser microphones on the other hand feature a metal plate and a conductive diaphragm close to eachother forming a capacitor.  As the sound waves hit the diaphragm, the distance between the plates changes causing the capacitance to change.  Condenser microphones require power.  This can be supplied by batteries or by your mixer if the input has phantom power.

Generally speaking, condenser microphones offer higher sensitivity and are preferred for high quality applications such as studio recordings and offer better frequency response.  The downside is that they require power and can distort with high volumes.

For commentary purposes, you may find dynamic microphones more suitable as the lower sensitivity will result in less background noise being picked up.  They are also cheaper than condenser microphones and more hard wearing – ideal if you travel.

*note about pc mics

Polar Patterns
Not all microphones respond to sound the same way.  Polar patterns describe the frequency response characteristics of a microphone in relation to the direction of sound.  The most common polar patterns are omni directional, bi directional, cardioid and super cardioid.

Audio Technica BPHS1 Headset Cardioid pattern

If the above pattern doesn’t mean much to you, imagine you are at 0° and you are talking directly into the microphone.  180° would be directly behind the microphone, and 90° and 270° would be to the left and right of it.

What the cardioid pattern shows us is that it has high pickup directly in front of it, some to the left and right and very little behind it.  Cardioid microphones are a great choice in settings where there is noise in front of you that you’d rather cut out.  Think of singers at concerts or streamers in front of a noisy crowd.

Super cardioid as the name suggests, is like a buffed version of cardioid with a narrower pickup pattern that rejects more noise from the left and right, however they do pick up some sound from behind.  You’ll need to check out individual patterns for microphones to see if it’s a worthwhile trade off.

Omni directional is easy to understand – the microphone picks up sound in all directions, mostly equally, giving us a circular polar pattern.  If you want to capture the ambiance of an event, omni directional might be worth checking out.

Bi-directional is also called figure of 8 due to its shape and its pattern looks a little like two cardioid microphones back to back.  It has high pick up at the front and rear, and very low pick up at the sides making them a good choice for interviews.

In cases where each commentator has a micrphone, we’d recommend sticking with cardioid unless you have a specific requirement/usage.

Player Cameras
Recommended – Micomsoft XRGB Mini (aka Frame Meister) with Euro SCART adapter
You will also need – Travel adapter for Japanese plugs to your socket
Useful links – XRGB mini settings | XRGB mini firmware | CPS2 Stream Overlays
Alternative – Open Source Scan Converter

There are two paths you can take with cameras.  For professional level quality with professional level budgets, there is the video camera and HDMI capture route.  Since you need a capture card for each camera (or an expensive multi HDMI capture card), the cost can be prohibitive.  There is a more affordable route which is using webcams.

It’s no secret that webcams use low end components – that is why they are affordable.  The main limitations with webcams are the framerate and low light performance, both of which can be improved with better lighting.

There are many webcams to choose from, but the Logitech C920 has stood the test of time and remains a great quality webcam at an affordable price.  There are a few tweaks that you can make to ensure a solid frame rate and lower latency.

A note on using multiple cameras

Due to a limitation or bug in the driver, it is not possible to configure multiple cameras independently.  Adjustments you make only affect one camera, regardless which camera you selected to configure.  Fortunately, these cameras work without a driver and in our opinion are better without it.  If you have the driver installed, uninstall it, we’re going to be setting things up manually to obtain the best quality.

Manual image controls

Most controls can be left on automatic, but we’ll be setting focus and exposure to manual.

One of the reasons webcams suffer from poor framerates is when the lighting is low and they try to compensate by adjusting the exposure.  By unchecking auto exposure, we can set it to -5 which seems to be the sweet spot for the best framerate.  If the image is too dark , you can compensate by adjusting the gain and brightness/contrast.  If after that it is still too dark, you may have to increase the exposure to -4 and live with the lower framerate.

Unchecking auto focus will stop the webcam trying to hunt around and focus which can cause the image to go blurry for a little while.  When it comes to player cameras, the players are always sat at roughly the same distance from the camera, so there is no need for auto focus.  We find that the default value is usually in focus for distance that players sit from a TV mounted camera.

Latency and sync

Just like with USB 2 capture cards, USB 2 webcams don’t have enough bandwidth available to transfer uncompressed high definition video, so they employ compression to do so.  If you remember from the capture card section, this causes the output to lag behind what is actually happening.  Fortunately, the C920 offers a few different output modes that help us combat that.

The C920 has H.264, MJPG and YUV modes.  H.264 and MJPG are types of compression, with MJPG having lower latency than H.264.  YUV is uncompressed, but because uncompressed video uses a lot of bandwidth, only certain resolutions are supported – up to a maximum of 640×360.

640×360 might not sound like much, but for windowed player cameras it’s enough, even for 1080p.  You may want to go with MJPG or H.264 if you want a scene with a camera in fullscreen for commentators as you will have access to the higher resolutions.

The Hookup

Now that we know what each component does and how to configure it, it’s time to move on to the hook up.

Just source, raw source

At the heart of any stream is a basic capture of the source material.  You can use this if you are just starting out or if you want a clean capture.

cabinets using rca in


, mixer +splitter cables, mic types, audio ground loop, webcams

LGP2 mjpg 1080p60

audio captured in real time but the video is delayed by the compression process.  combine at upscaler or scart block

Comments are closed.