Receptor Latency Testing Results

[Rick@MuseResearch]

Receptor Latency Testing Results
©2004 Muse Research Incorporated

Test conditions: we used a Tektronix digital storage oscilloscope model TDS350 to capture a MIDI event and the onset of an audio waveform from a plug-in in the case of measuring MIDI latency, and the onset of an input and an output waveform in the case of measuring throughput latency. In the case of audio input to output testing, an Audio Precision ATS-1 audio analyzer generated a pulsed audio test signal and the delay between the test signal and the output of Receptor was measured. In the case of MIDI to audio output latency testing, the 4Front E-Piano VSTi plug-in was used.

In the case of audio input to audio output, the line input was used, at nominal gain, with no effects instantiated so as to measure true A to D to A latency. In both cases the delta between the triggering signal and the audio output was measured using the TDS350 internal delta measurement tools, and worst case numbers were used. The unit under test, Receptor (serial number R40129030027) was pulled at random from production inventory, no tweaks or modifications were done to the unit.

In each case the measurements were made at all four different audio buffer settings and three common sample rates used in Receptor (1s = 1000mS)

Line in to Line out:

Code: Select all

Buffer setting    44.1kHz  48.0kHz  96.0kHz
-------------------------------------------
32                4.3mS    4.0 mS   2.0 mS
64                7.2 mS   6.7 mS   3.3 mS
128               13.0 mS  12.0 mS  6.0 mS
256               24.5 mS  22.7 mS  11.3 mS

As a frame of reference, sound travels at 1100 feet per second (at standard atmospheric pressure and temperature). This means an amplifier located 10 feet behind you on stage will have 9.09mS of “latency” between the time the sound leaves the cone of the speaker and the time it hits your ear. So if you perform through Receptor on-stage, at a buffer setting of 64 buffers at 48kHz, the latency is similar to an amplifier located 7.4 feet behind you.


MIDI input to Line out:

Code: Select all

Buffer setting    44.1kHz  48.0kHz  96.0kHz
-------------------------------------------
32                3.4 mS   3.1 mS   2.0 mS
64                5.1 mS   4.3 mS   2.7 mS
128               9.0 mS   7.5 mS   4.6 mS
256               18.0 mS  16.0 mS  7.5 mS

By way of comparison, we will compare Receptor's latency with two common, comparable configurations: a laptop with an external audio/midi USB interface device, and a desktop pc with an internal audio/midi solution. The results for the desktop system are forthcoming, the lap top configuration test results follow:

For the laptop comparison, we repeated the test using an HP Pavilion model ze3500 lap top computer (2.66Ghz P4 processor, 448MB Ram, running Windows XP) using an M-Audio Audiophile USB audio / MIDI interface. The Laptop was configured to run Cubase SX, with only one VST instantiated (4Front E-piano) with project sample rate set to 48kHz. Device settings were configured in Cubase so that the “lower latency” radio button was checked.

The latency settings were then changed on the M-Audio driver panel, and each time Cubase was rebooted so that those new settings were used.

Code: Select all

Latency setting   48kHz
-------------------------
“Very low”        14.6 mS
“Low”             21.6 mS
“Medium”          25.3 mS
“High”            36.4 mS
“Very high”       59.2 mS

In this particular test case, using this particular I/O box, Receptor’s worst latency is nearly as good as the laptop’s best case latency.

Of course, the latency of a desktop computer system with an internal sound card should (theoretically, at least) be better than that of a laptop using a USB interface due to the absence of the latency contributions of the USB interface layer. And due to a variety of reasons, no two computers are exactly the same in terms of performance, so computer tests can vary significantly. We will be doing some comparison tests with a desktop system soon and will post the results when they become available.

If you’d like to share your own test results, we’d be delighted to hear them.

[negativeinjury]

The results for the desktop system are forthcoming, the lap top configuration test results follow:

For the laptop comparison, we repeated the test using an HP Pavilion model ze3500 lap top computer (2.66Ghz P4 processor, 448MB Ram, running Windows XP) using an M-Audio Audiophile USB audio / MIDI interface. The Laptop was configured to run Cubase SX, with only one VST instantiated (4Front E-piano) with project sample rate set to 48kHz. Device settings were configured in Cubase so that the “lower latency” radio button was checked.

The latency settings were then changed on the M-Audio driver panel, and each time Cubase was rebooted so that those new settings were used.

Code: Select all

Latency setting   48kHz
-------------------------
“Very low”        14.6 mS
“Low”             21.6 mS
“Medium”          25.3 mS
“High”            36.4 mS
“Very high”       59.2 mS

In this particular test case, using this particular I/O box, Receptor’s worst latency is nearly as good as the laptop’s best case latency.

This is a completely useless comparison, distinctly skewed to make receptor look very much better than a laptop. As Muse Research well know , you are ALWAYS going to get terrible latency through USB! The comparison should have been done with a USB2 or Firewire device.

The usb audiophile isn't advertised for having low latency, so it is hugely incorrect for Muse Research to compare their Receptor to it.

Now, I am not suggesting that receptor is a bad piece of kit (far from it, even workstation keyboards have a responce time that's greater than 4ms!) but just that this is a truly shameful abuse of statistics, and is extrememly misleading.

Regardless of how much MuseResearch say "In this particular test case", they cannot escape the fact that an audio interface, representative of what is available for use with a Laptop, should have been chosen when making a claim that "Receptor’s worst latency is nearly as good as the laptop’s best case latency."

A device with a mininum latency of 14.1ms is the worst case scenario for a laptop, and as such is not representative of what is avaibable. I can understand a degree of reluctance from a company to search down the lowest latency portable audio interface and use that in this kind of comparison (i'm not that idealistic, lol... ), but a "middle ground" unit should have been chosen.

-Neg☼I[/quote]

[wonderwanker]

Now, I am not suggesting that receptor is a bad piece of kit (far from it, even workstation keyboards have a responce time that's greater than 4ms!) but just that this is a truly shameful abuse of statistics, and is extrememly misleading.
-Neg?I

I agree. I was and still am excited about the receptor as it looks to be something that is almost exactly what I have been searching for but such a blatently skewed presentation just turns me right off.

Operating at 44.1khz A PC with an RME Multiface/Cardbuss running logic with a 64 sample IO buffer setting has a latency of 190 samples or 4.3 ms.

The receptor operating at the same buffer setting has latency of 7.3 ms which tells me it has about 128 more samples worth of buffering going on somewhere.

I see the receptor has a 32 sample buffer setting which lets it operate at a 4.3 ms latency. That is an acceptable amount of latency for me as a player to deal with. 7.3 is more than I want to deal with.

What plugins can I use at the 32 sample IO buffer setting?

Are there an convolution reverb plugins that will work at that setting?

How fast is the receptor CPU?

Is it up gradeable?

[SpIdErWeB]

Hi,

It's a Athlon XP2500+
I tried to change it with Athlon XP3200+ but the FSB200 don't really works correctly, even the mobo is FSB200 Ready.

I tried with many type of RAM, don't change anything.

This is one of my reason I resell my receptor now.

To much close, non evolutive, really so bad because the concept is good.

For start, the day who they will exit the close OS, I will perhaps change my opinion.

Bye
Spid

[negativeinjury]

Operating at 44.1khz A PC with an RME Multiface/Cardbuss running logic with a 64 sample IO buffer setting has a latency of 190 samples or 4.3 ms.

The receptor operating at the same buffer setting has latency of 7.3 ms which tells me it has about 128 more samples worth of buffering going on somewhere.

I see the receptor has a 32 sample buffer setting which lets it operate at a 4.3 ms latency. That is an acceptable amount of latency for me as a player to deal with. 7.3 is more than I want to deal with.

the figure they quoted is how long it takes to get through the input buffer and the output buffer. Are you certain the figure from your RME is not just the output buffer, ie the latency for playing virtual instruments. Their figure concerns how long it takes a sound to get into receptor and back out. Try their measurement suggestion.


Neg ☼ I

[DanTimis]

I agree. I was and still am excited about the receptor as it looks to be something that is almost exactly what I have been searching for but such a blatently skewed presentation just turns me right off.

Operating at 44.1khz A PC with an RME Multiface/Cardbuss running logic with a 64 sample IO buffer setting has a latency of 190 samples or 4.3 ms.

The receptor operating at the same buffer setting has latency of 7.3 ms which tells me it has about 128 more samples worth of buffering going on somewhere.

I see the receptor has a 32 sample buffer setting which lets it operate at a 4.3 ms latency. That is an acceptable amount of latency for me as a player to deal with. 7.3 is more than I want to deal with.

We do not claim that our latency is better than a desktop system, it is probably more or less the same. But, Receptor is portable so our first "official" test was with a laptop. We will perform other tests and post them.

We do not have currently an RME Multiface/Cardbus to test with. If we can get our hands on one we will test it and post the results. But, that is a more expensive system - $299 for the Cardbus and $649 for the Multiface at sweetwater. That is only $450 less than Receptor just for the interface without the laptop.

We were very clear what was the equipment we tested. It happens to be a laptop and USB interface similarly priced to a Receptor. Yes, a more expensive portable system or a desktop system has better latency than the laptop with a USB interface. But you'll have to pay more for the first and the second is not portable.

You cannot have your cake and eat it too, not even with Receptor. Receptor has more than decent latency, it is portable, and rugged. A similarly priced laptop with a USB interface allows you to run a sequencer and read your email, but has worse latency and is less rugged. A more expensive laptop with a more expensive interface has very good latency, allows you to run your sequencer and read email, but it is not rugged and it costs more than Receptor.

For every problem there is a solution and we know that there are situations where a computer might be better suited. But, we hope Receptor will be able to help musicians in situations where the computer comes up a little short.

[edoosx]

To much close, non evolutive, really so bad because the concept is good.

I totally agree. Please...

[SpIdErWeB]

Right,

But the hardware isn't not the most important thing now, because we can have bigger version with bigger HD and more RAM...

The biggest problem (IMHO) is the system for the plugin.... really too much restraining when you already own a lot of plugins on your PC....

Even with the best efforts of Muse Research, there's always plugins which don't work, always plugins you need to have a special installer, etc... And this is mostly due to the protection... and everybody knows that the protection are mostly restraining for the honest guys, not the hackers...

I have a lot of plugins (with licence), but I had unlocked/hacked my Receptor to don't have problem in future.... I know that's not a real and good solution, but now that's a little bit better for me...

my 2 cents...

Bye
Spid

[thesoundsmith]

Guys, the world's not perfect. I do not have a Receptor yet, but may be getting the E/W one very shortly. But when I compare Receptor's latency with most of my machines, they're comparable. And my big, bad (last year) 3G dual-core box that I poured my heart and soul into, along with a few bucks, won't run my two primary apps - one will not run HTed and the other must. Nothing is perfect, and this is a product that does a job. If it does your job, use it. If it doesn't, don't buy it. But beating up on the programmer doesn't solve any issues (I know, I had 20 years of people POed at me because when I did a 6-month coding job in three, they were unhappy it wasn't 6 weeks.

Give these guys some respect. They're not out to rip you. From what I've seen (and I've been watching for a while now) Dan and the Muse team are going to do everything they can to make the Receptor as useful and fast as is possible within its hardware class. But good code and good processes take time to develop, and in the meanwhile, here's a portable solution that works just fine for some people for some situations. If it works as is right now for you, great. if not, how close is it? It's not a computer, it's a tool, an appliance even, a hardware/software module. You'll never (I hope...) run Excel or 'Age of Umpires-Sports Wars' It's a dedicated machine, and if you can make it do your tricks today, it will probably do the same tricksin ten years, But in five years it probably won't do new ones, they'll all be 64-bit, or require 20G RAM and 2T storage...

The Receptor can't be all things to all people. Don't worry about it, just use it for what it is, and they'll take it as far as it can go, as they can find ways to do so. I read their ads from the get and saw nothing that makes me think there was any misleading prose. But I always look harshly at ads, no matter what the source. And ask stupid questions. And tick people off sometimes.

Sorry if this response does that to anyone, I'm just kind of tired of reading all negative stuff.

And FWIW, I have an almost identical laptop to the one tested. I've done no formal testing, but with the Audiophile, I have a moderately tight but not crisp feel playing the B4, haven't used the piano he's talkling about, but his specs for the Receptor are definitely better than mine.

[kara]

I pretty much agree with this.
I don't own a receptor but have followed his evolution for a long time now.
I think it is a very good solution for on stage performance and in his price range it is unique.
Since I don't need this on stage (I'm a piano player) I didn't buy it,but... build my own studio 'receptor'.
I went only for quality hardware and guess what ? It ended in the same price of Muse Receptor
Still I have a couple advantages :
- It's easy upgradable
- It runs all VST's and VSTi's
- No ilock protection

But if I ever would use such a thing on stage, I would buy the receptor

Rony

[Jazz+]

Interesting

[TomDM]

I found this comparison very useful in spite of the criticism for comparing Apples and oranges.

I didn't appear that the comparison was trying to assert that the Receptor beat all the other options. I took it as a simple test between one set of specific options vs. another set of specific options.

In this context it was helpful, to me, since it showed that I can at least expect performance equal to that of a comparably priced computer and audio interface combination.

Thanks!

Tom

[janfree1]

interestingly enuff there were no responses to:

are there any HEAVY vsti STUDIO users that are just thrilled with the efficiency of this thing? as compared to having done the same thing on a reasonably robust desktop>

[thesoundsmith]

janfree1, I doubt you will find such a thing - as Dan said above it's not necessarily FASTER or lower latency than a desktop. Look at the Receptor for what it is - a dedicated VSTi player, similar to a hardware synth in that what you get is what you get, and nobody ever released a hardware upgrade for a Wavestation or Triton. It is a dedicated device, so it is engineered to perform certain tasks as efficiently as possible, and to not care about other possible uses.

I tried most of the other under-$4K options, and eventually bought my Receptor, and all I can say is that I wish I had done so BEFORE I invested in the laptop rig, which is only about half as useful as Receptor, even when working perfectly (about 1/3 of the time.)

It is useful in the studio, but you can build faster, more flexible (more VST plugs will work with the desktop unit) PCs, but you won't easily carry them to the gig...

[Atardecer]

It is useful in the studio, but you can build faster, more flexible (more VST plugs will work with the desktop unit) PCs, but you won't easily carry them to the gig...

True that, you gotta hand it to the receptor for bridging the gap between the studio and the stage. Also the receptor makes a great "second computer" in my studio, taking the load off the main desktop by running most of my VST's. Try doing that with 2 computers using FX Teleport. Absolute pain the ass.

regards,
james