Frequency Band Splitters Shootout

This forum is for discussing Rack Extensions. Devs are all welcome to show off their goods.
User avatar
dioxide
Posts: 1781
Joined: 15 Jul 2015

07 Jan 2016

I've been comparing how the different Rack Extension splitter behave when set to the same setting. To be honest I was surprised at how much they vary.

Here I am running white noise through the splitter with only the high band activated. All devices are set to split at 1kHz on the device controls. Song file is attached as a Reason 8.3 file.

Edit: Added Splex to the GIF (but not the Reason song file). Note that Splex's high band only goes down to 1.6kHz as opposed to 1kHz, so for Splex it isn't a direct like-for-like comparison.
Splitter-Shootout.gif
Splitter-Shootout.gif (597.31 KiB) Viewed 3541 times
Products in the shop
https://shop.propellerheads.se/product/ ... -splitter/
https://shop.propellerheads.se/product/4dyne/
https://shop.propellerheads.se/product/ ... -splitter/
https://shop.propellerheads.se/product/ ... ompressor/

Here are the split types on the various devices. Not to say that one device is better than another, but you may find different setups work better for different uses, mastering compression, multiband effects etc.

Yoko Linkwitz-Riley
User manual: http://www.unfilteredaudio.com/YokoOperationsManual.pdf

4Dyne Perfect (see manual) or Analog (Linkwitz-Riley)
User manual: http://www.flowerau.com/4Dyne-manual.pdf

Elements Splitter sum-to-zero crossovers (couldn't find any more detailed info in the manual)
User manual: http://airraid.ch/pdf/Splitter.pdf

Splex Linkwitz-Riley (12db, 24db or 48db) or BSL – this is a special filter type, constructed with combinations of complementary Bessel filters specifically designed to guarantee that the recombined output signal is in-phase with the input signal.
User manual: http://dlogb.altervista.org/splex/SPLEX ... Manual.pdf

MClass no further details
User manual: http://docs.propellerheads.se/reason83/ ... .38.3.html

https://en.wikipedia.org/wiki/Linkwitz–Riley_filter
https://en.wikipedia.org/wiki/Bessel_filter
Attachments
Splitter Shootout.zip
(105.58 KiB) Downloaded 78 times
Last edited by dioxide on 09 Jan 2016, edited 6 times in total.

User avatar
dioxide
Posts: 1781
Joined: 15 Jul 2015

07 Jan 2016

Unfortunately it looks like 4Dyne doesn't behave properly at different sample rates. The filter is dependent on the sample rate and halves when you switch to 96khz.

Edit: Flower Audio have responded already and have confirmed the issue and committed to fixing it as soon as possible. More info here:
http://www.reasontalk.com/viewtopic.php ... 00#p239700
4Dyne-bug.gif
4Dyne-bug.gif (378.04 KiB) Viewed 3726 times
Last edited by dioxide on 08 Jan 2016, edited 1 time in total.

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

07 Jan 2016

Interesting comparisons - at first I was wondering how a splitter could be affecting things in any way, since that term is historically used to denote an audio signal which is sent to more than one place (see the Props own Audio and CV merger/splitter as an example, particularly the one in your video labeled "SPLIT"!) ;)

What you're talking about is a crossover - is "splitter" a DJ term maybe? Anyway, crossovers are a set of complementary filters, and as such there are as many variations with crossovers as with filters. With a crossover, the more important thing IMO is how they combine, which is to say how they "color" the sound when the audio bands are re-combined. Ideally there should be no change, with the Props own Stereo Imager setting the bar with absolutely no coloration and a completely flat signal when the bands are re-combined. There are plenty of examples of crossovers in multi band units that do not re-combine as accurately as one may like, with the McDSP ML-4000 being a well known example (though I still find it useful myself).

There are some who feel steep slopes are more desirable than gentle slopes for things like multi-band compression, but I much more prefer the gentle slopes for the same reasons I prefer crossfades on edits over butt-splices in most cases. This is because there are very few sounds that occupy ONLY a single band of frequencies and can therefore be controlled by steep bands - most of the time a sound occupies a wide range of frequencies, especially a sound that moves in pitch (melodic or chordal). In my experience, the steeper the slope the more you "hear" the processor working, which is almost always a bad thing IMO.

Ideally IMO, a variable slope is best for mult-band type processing since you can dial in the best slope based on the sound. As far as I'm aware, the Waves C4 is the only multi-band with this feature (called Q in their product), but there are many new products that I'm not aware of that may offer this useful feature.

Great video, very useful to visualize the differences when comparing similar processors!
:)
Selig Audio, LLC

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

07 Jan 2016

dioxide wrote:Unfortunately it looks like 4Dyne doesn't behave properly at different sample rates. The filter is dependent on the sample rate and halves when you switch to 96khz.
4Dyne-bug.gif
That's odd and certainly not desirable. I don't use the 4Dyne so I can't comment further or confirm this.
Selig Audio, LLC

User avatar
dioxide
Posts: 1781
Joined: 15 Jul 2015

07 Jan 2016

selig wrote:What you're talking about is a crossover - is "splitter" a DJ term maybe?
I don't know, is there a time when DJs use frequency splitting? I can't think of anything for DJs that does this? For me a crossover refers to a single instance of this kind of circuit as you find in a speaker.

I use the term 'splitter' as a shortened term used for this kind of device, a device that uses multiple crossovers - a band splitter or frequency splitter. Two of the four devices here include the term splitter in the device name, and a third has it on the faceplate. It's only the PH device that uses 'crossover' but then it's a device that uses only a single crossover to create two frequency bands. Either way I think it's fair to say that this term is in common use and there are more examples of this term being used for this kind if device out there in VST land.
splitter.jpg
splitter.jpg (30.76 KiB) Viewed 3703 times
For the purposes of this thread a 'splitter' is defined as an audio device that's main purpose is to split the sound into multiple frequency bands using multiple crossovers. I agree it is confusing though and the term 'frequency splitter' or 'band splitter' is more specific.

As you say another useful test would be to see how well the split bands recombine once reassembled. This is really out of my level of expertise as it probably involves looking for phase problems and the most I could do is see if things cancel out when phase inverted. I'm assuming here that all the devs have already checked that as part of their own test and in any case I don't have the tools or the knowledge to do the phase checks. Feel free to have a go at this if you are able to, it is all useful info for the community.

I have contacted Flower Audio to see if they can comment on the sample rate issue. It would be nice if it can be fixed if it is confirmed as a bug.

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

07 Jan 2016

You can have a 2 or 3 way crossover, so the term doesn't only apply only to one. Besides that, a single crossover IS a band splitter by that definition, right? Further complicating things, googling band and frequency splitter give all sorts of radio related devices rather than Audi crossovers. I've only recently seen the term splitter used for a crossover, and only in software.
But I'm just an old fart and tend to stick to the original terms rather than confusing things by introducing new terms for existing technology!

EDIT- found an interesting thread in gearslutz from back in 2011 that not only corrects the term "splitter" as I did here, but also mentions using the name YOKO for a plugin! Seems the term may have originated in an Adobe audition plugin, as Adobe is not an audio company they probably choose a word that made sense to them. Then others copied it, etc. Just conjectured on my part here...
https://www.gearslutz.com/board/masteri ... itter.html
Selig Audio, LLC

User avatar
dioxide
Posts: 1781
Joined: 15 Jul 2015

07 Jan 2016

No worries. I know you're from more of a studio background, but most here aren't and we use the terms the industry feeds us. Frequency/band splitter seems to be in common use now also for better or worse ;)
https://helpx.adobe.com/audition/using/ ... itter.html
https://kilohearts.com/products/multipass/

User avatar
dioxide
Posts: 1781
Joined: 15 Jul 2015

07 Jan 2016

I'm guessing here but maybe crossover refers to a specific type of circuit configuration. So a frequency band splitter would employ a 2-way crossover circuit to split the audio into two frequency bands or a 3-way crossover to split into three frequency bands. But the complete device itself (containing other elements for different functions) is called a frequency band splitter. Note that only the knob on the MClass is labelled crossover. The freq splits on the AirRaid device are also labelled as crossovers, but the device itself has other functions (such as panning) so the complete device is a frequency band splitter.

Actually here I am showing the differences between filter roll off of one of the crossover filters (HPF), so we're both right :D

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

07 Jan 2016

dioxide wrote:I'm guessing here but maybe crossover refers to a specific type of circuit configuration. So a frequency band splitter would employ a 2-way crossover circuit to split the audio into two frequency bands or a 3-way crossover to split into three frequency bands. But the complete device itself (containing other elements for different functions) is called a frequency band splitter. Note that only the knob on the MClass is labelled crossover. The freq splits on the AirRaid device are also labelled as crossovers, but the device itself has other functions (such as panning) so the complete device is a frequency band splitter.

Actually here I am showing the differences between filter roll off of one of the crossover filters (HPF), so we're both right :D
hey, you're not the one who chose that term. But yes, crossover is industry standard, and splitter is a more recent and "simplified" term. Both are talking about exactly the same thing, hence my frustration that new terms are invented when the old terms still perfectly apply. Not anything I'm loosing sleep over…
;)
Selig Audio, LLC

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

07 Jan 2016

dioxide wrote:No worries. I know you're from more of a studio background, but most here aren't and we use the terms the industry feeds us. Frequency/band splitter seems to be in common use now also for better or worse ;)
https://helpx.adobe.com/audition/using/ ... itter.html
https://kilohearts.com/products/multipass/
Adobe, not being an audio company, may have actually started this new trend…
Selig Audio, LLC

User avatar
dioxide
Posts: 1781
Joined: 15 Jul 2015

07 Jan 2016

Does crossover get used in most studio settings in your experience? Pretty much everyone I know is from a MIDI + synths background so most of us have never had the need to go to an actual recording studio.

User avatar
dioxide
Posts: 1781
Joined: 15 Jul 2015

07 Jan 2016

selig wrote:
dioxide wrote:No worries. I know you're from more of a studio background, but most here aren't and we use the terms the industry feeds us. Frequency/band splitter seems to be in common use now also for better or worse ;)
https://helpx.adobe.com/audition/using/ ... itter.html
https://kilohearts.com/products/multipass/
Adobe, not being an audio company, may have actually started this new trend…
I'm not sure, Adobe don't have enough authority in the audio world to create their own terms and I'm betting they know this. I'm leaning towards the theory that crossover is an electronics term and the earliest of these devices were homemade type affairs and so the term crossover was used in studio environments where this stuff was first in use. As I say I think the term crossover is only the actual circuit that does the splitting. It is a part of a larger device in the same way that a LPF might be part of a synthesizer or some other device.
http://www.electronicshub.org/active-au ... r-circuit/

An interesting detour if anyone is interested:
https://www.gearslutz.com/board/masteri ... essor.html

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

07 Jan 2016

dioxide wrote:Does crossover get used in most studio settings in your experience? Pretty much everyone I know is from a MIDI + synths background so most of us have never had the need to go to an actual recording studio.
A crossover is a crossover, whether it's in a pro or home studio (or whether it's called a band splitter or a crossover!).
Like I said previously, it may be the term band splitter came from Adobe, since it's only recently become popular (someone correct me if I'm wrong on this). I can't find a use of the term "band splitter" on any product before Adobe introduced it in 2006 (again, someone please correct me if I'm wrong here!). Audition, fwiw, was considered a home studio and later broadcast application, so it may have been more free to use it's own terminology rather than stick to industry standards. Again, I'm not loosing any sleep over this, it's just a curiosity IMO.
Selig Audio, LLC

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

07 Jan 2016

dioxide wrote:
selig wrote:
dioxide wrote:No worries. I know you're from more of a studio background, but most here aren't and we use the terms the industry feeds us. Frequency/band splitter seems to be in common use now also for better or worse ;)
https://helpx.adobe.com/audition/using/ ... itter.html
https://kilohearts.com/products/multipass/
Adobe, not being an audio company, may have actually started this new trend…
I'm not sure, Adobe don't have enough authority in the audio world to create their own terms and I'm betting they know this. I'm leaning towards the theory that crossover is an electronics term and the earliest of these devices were homemade type affairs and so the term crossover was used in studio environments where this stuff was first in use. As I say I think the term crossover is only the actual circuit that does the splitting. It is a part of a larger device in the same way that a LPF might be part of a synthesizer or some other device.
http://www.electronicshub.org/active-au ... r-circuit/

An interesting detour if anyone is interested:
https://www.gearslutz.com/board/masteri ... essor.html
But there's no evidence the term Band Splitter was used before Adobe used it, right? Adobe wouldn't need any "authority" to use whatever terms they want to use, so I don't follow that logic.

Sticking strictly to software rather than hardware, again I can find no use of the term band splitter before Adobe used it in 2006. Any evidence to the contrary would be appreciated!
Selig Audio, LLC

User avatar
dioxide
Posts: 1781
Joined: 15 Jul 2015

07 Jan 2016

Post moved to first post.
Last edited by dioxide on 09 Jan 2016, edited 2 times in total.

User avatar
dioxide
Posts: 1781
Joined: 15 Jul 2015

07 Jan 2016

selig wrote:
dioxide wrote:
selig wrote:But there's no evidence the term Band Splitter was used before Adobe used it, right? Adobe wouldn't need any "authority" to use whatever terms they want to use, so I don't follow that logic.

Sticking strictly to software rather than hardware, again I can find no use of the term band splitter before Adobe used it in 2006. Any evidence to the contrary would be appreciated!
I'm sure it would ;) But as it's your theory then the burden of proof is on you to prove it, not for me to prove you wrong :D
https://yourlogicalfallacyis.com/burden-of-proof

Personally I just don't see Adobe's marketing department as being stupid enough to start making up their own terms when they're trying to get their foot in the door to a market where they don't have any kind of decent reputation. I know people love to hate marketing people and I've met no end of stupid marketing people in my time (oh yes I have ;)) but IMO something like that would be commercial suicide. Anyway, as I say the burden of proof is on you now to prove your theory.

Looking at the GS thread someone-on-the-internet thinks that multiband compression originated in broadcast, which kind of sounds right to me. Multi-band everything-else would have probably come after that. It is an interesting detour that I'm going to look into, seeing as I'm in science lab mode and it's something I've not thought about too much. But if we can keep discussion here to the devices available in Reason that would be more in tune with the theme of this thread: band splitter devices available in RE format and how they compare. ;)

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

07 Jan 2016

I'm not interested in proving anything to anyone here, just saying what I've seen and making an educated guess based on my years of experience (and a little research to back up my theories). Besides that, I don't see you backing up your theories either, which I don't agree with myself. As for the intelligence of marketing departments, I don't agree they are all beyond making up their own terms, which is hardly suicide! There are plenty of examples of bad marketing decisions made by major companies, which proves they are not beyond this sort of thing fwiw.

Anyway, I apologize for the detour and thank you for indulging mean this discussion!
Selig Audio, LLC

User avatar
Bonkhead
Posts: 335
Joined: 18 Jan 2015

08 Jan 2016

I guess we need the reverse the question, from "are these splitters ? " to "aren't these splitters ?".
Since they split one audio channel into multiple channels, they are. Same as the spiders or the Reason 1 standard modules.

No matter if it is about crossover filter behaviour or not.

User avatar
Bonkhead
Posts: 335
Joined: 18 Jan 2015

08 Jan 2016

dioxide wrote:Unfortunately it looks like 4Dyne doesn't behave properly at different sample rates. The filter is dependent on the sample rate and halves when you switch to 96khz.
As a beta tester I've seen this happen often for different devices of different developers (and also being corrected after mentioning it to the dev). I guess it's a simple looked over development error when creating RE's. I mentioned it last week for a new device which is in beta, for instance.

But indeed, this being still part of a relative long time released RE is a bit weird.

User avatar
LABONERECORDINGS
RE Developer
Posts: 401
Joined: 16 Jan 2015
Location: UK
Contact:

08 Jan 2016

Ok another angle..... crossovers, to us, are commonly in equipment like speakers, car amplifiers etc, where the crossover splits the signals but have no user control for the frequency points (some sub channels may have a frequency tuner, plus may be LPF, HPF or full range... we have one that does this)

Splitters seem to be more towards the audio production process, so we would 'split' a signal to get the parts we want.

MClass seems more of a tuneable crossover (all in one unit) until you change it's function to split low and high frequencies (then becomes a 'splitter')

Just our little spin on things :D call it what you want, one man's splitter is another man's crossover LOL

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

08 Jan 2016

Bonkhead wrote:I guess we need the reverse the question, from "are these splitters ? " to "aren't these splitters ?".
Since they split one audio channel into multiple channels, they are. Same as the spiders or the Reason 1 standard modules.

No matter if it is about crossover filter behaviour or not.
They could also be called band dividers, or band separaters since that's what they do: divide or separate an audio signal into separate bands. The issue remains that they ARE crossovers, that are being called "splitters" by some developers (as a way of simplifying the idea to newbs maybe?). To put it another way, Adobe created a process using crossovers that they called a "Band Splitter" for whatever reasons. Bottom line: there is no way to "split" or "divide" (or "isolate", a DJ version of a crossover) the signal into bands other than to use a crossover, full stop.
Selig Audio, LLC

User avatar
dioxide
Posts: 1781
Joined: 15 Jul 2015

08 Jan 2016

Well Flower Audio responded very fast to my email, which is a good sign! They say they are going to fix this issue as soon as they can.

From the email:
- It does appear that at the 88 and 96k sample rates, the frequency at which the splitter splits the audio is double what it should be. Thank you for pointing this out. We will fix this issue and send out an update as soon as we can.

- The slope of the low-pass signal in 4Dyne is always 12 or 24db per octave as in the label, but the high pass slope varies slightly based on the frequency. This is normal behavior and is part of the perfect reconstruction splitting algorithm in 4Dyne.

- In the 'perfect' modes in 4Dyne, if you add the low and high signals back up, you get a signal identical to the original signal (except that there is a small amount of latency). So the base-line for 4Dyne is completely neutral on the sound in this sense.

User avatar
K1TTENM1TTEN
Posts: 315
Joined: 16 Jan 2015

08 Jan 2016

Technical terms aside, can we all just agree that the Yoko Band-Splitter is possibly one of the greatest names ever for a device? I bought that device because

1) It was on sale.
2) It did something I wanted to have easily done.
3) The pun was gold.

</end random interjection>

Back to topic, thank you for posting this! I have been curious why some devices change my signal's sound despite the crossover device being init.

User avatar
Aikmofobi
Posts: 39
Joined: 17 Jan 2015

09 Jan 2016

K1TTENM1TTEN wrote:1) It was on sale.
2) It did something I wanted to have easily done.
3) The pun was gold.
My reasons for buying it as well. Brilliant marketing.

Ostermilk
Posts: 1535
Joined: 15 Jan 2015

09 Jan 2016

Splex is another RE band splitter (or variable 3-way crosoover if you prefer... ;) ) as well since you can bypass it's compressor stages.

Just thought I'd mention it.

Post Reply
  • Information
  • Who is online

    Users browsing this forum: prefab sprout and 26 guests