H9000 as main audio interface over USB QUESTION

[Kamurah]

Hello all,

I am new to this forum and am strongly considering purchase of an H9000R.  I was wondering if anyone here is successfully using the H9000 as their main audio interface over USB in a recording environment, and if so, how is the latency performance of the system?

 

I would initially be running this unit over USB, but would eventually be transitioning to a Dante system and incorporating the Dante expansion card for the box.

 

My current setup is a Mac running Logic Pro and have had success with both RME and Focusrite interfaces, but was curious if the H9000 as an interface could provide similar (latency) performance?

 

Thank you in advance!

[John Baylies]

Here are some details on what the H9000's latency depends on:

1) What IO you are using

• Using analog IO will add about 1ms of latency due to the converters.
• Digital IO will have much lower latency, probably a sample or two.
• Network audio will add some latency.  For Dante, this is configurable.
• There is an additional but very small latency due to the internal H9000 audio router, which is 3 samples in and 3 samples out, so a total of 6 samples, or 0.125ms at 48k.

2) The structure of the FX Chain that you are running audio through

• A single algorithm FX chain, or an FX chain in which the algorithms are in parallel, will add about 1.3ms of buffering latency.  Each additional series algorithm will add another 0.7ms of buffering latency.  So four algorithms in series will add about 3.3ms.
• The internal FX chain routing compensates for buffering latency, so you will not hear comb-filtering artifacts due to running e.g. 2 series algorithms in parallel with one.

3) Whether any of the algorithms add latency as part of the processing (e.g. a lookahead limiter).

• We don't currently have information on the processing latency added by the algorithms.
• In many cases it will be zero.
• Reporting and compensating for processing latency where necessary is something we'll consider implementing in future software versions.

Keep in mind that the H9000 does not have any microphone or Hi-Z inputs.

[2mmi]

I use a MADIface Pro with my H9000R (with MADI card), its a very nice combo.

The driver software for the H9000R is lacking (to put it mildly.)

Let’s hope Eventide transform their Emote plugin with Elekton-Overbridge-like integration with audio via USB

[John Baylies]

Here's what I get when I compare my H9000R's USB to my RME Fireface800's firewire on my 2012 MacBook…

I'll probably be buying the RME MADIface soon. I'm curious about how the MADIface's USB latency compares to the H9K's.

[Kamurah]

Thank you so very much for taking the time to post this.  Very useful and just the comparison I was looking for!  

[joeydego]

can the buffer size go below 64 samples? 

[Kamurah]

I know it’s obvious from the screenshots, but it is worthy of repeating…..Buffer size on its own is not the sole component to characterize low latency performance.  If you look at the comparisons above, INPUT latency on the H9000 via USB is considerably higher than the RME via Firewire at the same buffer setting (output latency appears to report pretty close).

[Kamurah]

John Baylies wrote:

Here are some details on what the H9000’s latency depends on:

1) What IO you are using

• Using analog IO will add about 1ms of latency due to the converters.
• Digital IO will have much lower latency, probably a sample or two.
• Network audio will add some latency.  For Dante, this is configurable.
• There is an additional but very small latency due to the internal H9000 audio router, which is 3 samples in and 3 samples out, so a total of 6 samples, or 0.125ms at 48k.

2) The structure of the FX Chain that you are running audio through

• A single algorithm FX chain, or an FX chain in which the algorithms are in parallel, will add about 1.3ms of buffering latency.  Each additional series algorithm will add another 0.7ms of buffering latency.  So four algorithms in series will add about 3.3ms.
• The internal FX chain routing compensates for buffering latency, so you will not hear comb-filtering artifacts due to running e.g. 2 series algorithms in parallel with one.

3) Whether any of the algorithms add latency as part of the processing (e.g. a lookahead limiter).

• We don’t currently have information on the processing latency added by the algorithms.
• In many cases it will be zero.
• Reporting and compensating for processing latency where necessary is something we’ll consider implementing in future software versions.

Keep in mind that the H9000 does not have any microphone or Hi-Z inputs.

Hello John, and thank you for the response.  It is much appreciated, and I have read that information previously….but what I am really looking for is the performance of the H9000(R) over USB (internal FX processing aside).  

All interfaces are not created equal…and as we have seen there can be a wide variance between USB implementation and how this translates into real world performance.  As an example, I currently use an RME Babyface with my portable setup.  I can state with confidence that the RTL performance of this interface exceeds other USB 2.0 interfaces I have experience with, and gets very close to PCIe figures….both in ASIO and Core Audio configurations.  Not only that, but RME’s driver stability is exceptional.

I would love to be able to use the H9000 as my main interface in the studio…but if it is unrealistic to expect similar performance from the Eventide, I would like to know before purchase so I can plan out a different implementation (MADI, Dante, or AES).

Thank you!

[John Baylies]

joeydego wrote:

can the buffer size go below 64 samples?

32 samples is the lowest I can set it in Ableton's preferences.

[Author]

Posts