Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Scott Holtzman]

Do you or anyone else on the thread have experience with these x18s?  Are they a real solution for breakout room or general AV applications over the old standby Mackie mixers?

Yes I several xr's.  Big step up from the Machie's in features.  Quality, they are both well build and sound fine.

The Mackie doesn't offer remote control of the pre's.  Just digital trim.  The xr, just like the x series have full control over the preamp.

[Justice C. Bigler]

Corey, I'm sure by now you've just gone and automated the levels in your project and sent it off.

But next time, I think what you need is the Leveler module from the RX5 Advanced bundle by iZotope. It's umm...only $1,200. 

[Corey Scogin]

Corey, I'm sure by now you've just gone and automated the levels in your project and sent it off.

But next time, I think what you need is the Leveler module from the RX5 Advanced bundle by iZotope. It's umm...only $1,200.

That sure looks like a useful dialog editing tool. They certainly are proud of it.

Between multiple other plugins, I already have 80% of that functionality. The way I read it, it still doesn't quite fix the specific issue that I was having as I think an automixer would by taking into account multiple sources and attenuating those not in use.

[George Dougherty]

So, it wouldn't handle reducing channels not in use, but if you can get any sort of gating set up that would allow a quick mixdown of all channels, the Levelator may do what you want otherwise.
http://www.conversationsnetwork.org/levelator
It is designed to take a rough mix, scan for varying levels of speakers, normalize regions then do some light compression to even out peak levels.  It's my first step in post processing spoken audio all the time and it's great in HOW scenarios where you get soft prayers, normal speaking voices and then periodic points of loud emphasis.  Even better is that it's free.

I'm personally looking to get an x-air just for talking head stuff as it's the least expensive automixer I've found and one I demoed worked well enough for a small back and forth session I did the other week.  It would be fairly inexpensive in this case and might work if you fed it line level inputs and ran them through the automixer function then recorded to USB or back into your DAW from the outputs.

[George Dougherty]

Do you or anyone else on the thread have experience with these x18s?  Are they a real solution for breakout room or general AV applications over the old standby Mackie mixers?

The only downside to the XR series is thinking you can depend on the built-in wifi.  It's not intelligent enough to auto-select a less crowded channel and you have to connect via PC to modify the settings.   In a crowded wifi environment I get about 10-15 ft of range.

[Corey Scogin]

If anyone uses Reaper for post production and would find an offline AutoMixer useful. I built a simple one in JSFX. More details and source code at the Reaper forums here: forum.cockos.com/showthread.php?t=173289

[John Roberts {JR}]

If anyone uses Reaper for post production and would find an offline AutoMixer useful. I built a simple one in JSFX. More details and source code at the Reaper forums here: forum.cockos.com/showthread.php?t=173289

You can perform AM with linear math.... If individual channel is say 90% of the sum of all level, you multiply it by that 90%. Division is not very fast, but perhaps faster than log conversion depending how you do either in code...I'm pretty sure I used linear math inside the AM I coded up (that never made it to market).

JR

[Corey Scogin]

You can perform AM with linear math.... If individual channel is say 90% of the sum of all level, you multiply it by that 90%. Division is not very fast, but perhaps faster than log conversion depending how you do either in code...I'm pretty sure I used linear math inside the AM I coded up (that never made it to market).

I'm doing linear math in the code posted over on the Reaper forums aren't I?  I started out using log values but changed it yesterday.

It's all a bit quick-and-dirty. I spent most of the time learning how JSFX works and figuring out how to code the RMS averaging.

Is there a way to do the calculation without division?

[John Roberts {JR}]

I'm doing linear math in the code posted over on the Reaper forums aren't I?  I started out using log values but changed it yesterday.

I must not of looked hard enough, I saw a few posts but not executable code that I recognized.

It's all a bit quick-and-dirty. I spent most of the time learning how JSFX works and figuring out how to code the RMS averaging.

I don't know what JSFX is so I can't help you.

I am not sure that RMS is needed... I've used simple average weighting for several AM designs.

Is there a way to do the calculation without division?

Which calculation? I actually coded up true RMS for a meter application but didn't use it because it wasn't different enough from average to justify the computing over head. That RMS calculation involved taking the square root of the integral of the signal squared. In machine language squaring something is a one clock tick multiply. The integral is just a weighted sum over time so relatively few clock ticks. The square root was the hard part and I wrote a sub routine that took maybe 3 clocks per significant bit of resolution... you don't need a 16b result to be usable.

To perform AM gain sharing you need to determine what fraction is the channel level compared to the "sum of all" level... Again I wrote my own routine to do this using fast multiplies, and compares in an iterative loop, instead of a formal divide.

For example to determine if the top bit is set to define the fractional value, multiply the "sum of all" level by Ox80 (0.5x) and compare the result of that multiply to the individual channel level. If it is higher the top bit is set, if it is lower that top bit is cleared. Then you set the next bit (making either 0xC0 =0.75x or 0x40=0.25x) depending on result from first trial, multiply and compare like before. Again you don't need more than the top several bits to get adequate gain sharing resolution (top 4 bits will give you down to 1/16th or more than 20dB of attenuation).     

In case this isn't what you were asking.... never mind. But I prefer iterative multiply and compare to doing a formal divide which probably takes more clock ticks at machine level.

I hope this makes sense...  I know more about AM and microprocessor coding in machine language than high level console control languages.

JR

[Corey Scogin]

I hope this makes sense...  I know more about AM and microprocessor coding in machine language than high level console control languages.

It does make sense.

Your first post made me think that you looked at the code I posted over on the Reaper forums and noticed something that could be improved.

JSFX is just Reaper's C-style scripting language. It is, of course, higher level than machine code but lower than what I'm used to working in. My only experience with machine code was in school but I understand what you describe.

Thanks for taking the time to respond. Your insight into the lack of need for a highly accurate RMS reading is valuable.

[ProSoundWeb Community]