We had a lot of fun last weekend recording a piano recital in three parallel formats. 96k, 192k and quad-rate DSD (11.2MHz sampling rate).
Anyone else tried quad-rate DSD?

Hi Tony, what converters did you use?

We used Merging Horus converters and mic pre's. The system uses Gigabit ethernet to avoid any USB bottleneck.

How did it sound? Was there a clear winner among the formats, or were the differences more subtle?

There's been a lot of discussion about the recent Boston Baroque Haydn recording 5/4 did in 11.2 MHz DSD. Has anyone heard it yet?

Does quad DSD push the ultrasonic noise even further out of the audible range? Would you even need a filter for it at that point?

We were backstage in a dodgy small control-room but what I heard on the monitors and over headphones was brilliant. Quad-rate DSD was my favourite. The noise with 64fS DSD rises steeply north of 25kHz, so 256fS should push it a lot further out of the way. I have made no measurements yet.

Merging Technologies Horus and Pyramix are the only converter and DAW capable of recording and playing back quad DSD.
Dennis

Hi Tony,
Ronald Prent (now Wisseloord, Netherlands) has done some testing a few years ago at Galaxy Studios with Philips prototype equipment @ 256FS. IIRC he liked it a lot. You could contact him for more info. Unfortunately I didn't have a chance to hear the setup myself.

Does the Horus allow for all three sample rates at once to different destinations? I would love to get my hands on one of those.

I am a little unfamiliar with DSD workflow, so just so I am straight on working with DSD files, does Pyramix require a conversion to "DXD" format to work with the audio files? If so, have you ever noticed any change in the sound quality after converting to PCM and back to DSD for SACD?

You can have Horus connected to 3 destinations but the sample rate would need to be the same. When working with DSD and Pyramix only the sections of audio that is manipulated is converted to DXD. So if you just are editing with no plug ins or mixing just the cross fades are converted. I'll let others comment on the sonic change.
Dennis

From the Merging FAQ:
On one RAVENNA network, is it possible to have one Horus running at 44.1kHz for example and the other one at 96kHz and use them both for different DAWs or in turn?
Yes. RAVENNA allows for concurrent multiple Sampling rates on the same network.
Note: While RAVENNA does indeed provide the possibility to simultaneously convey signals at different sampling rates over the same network, Horus is however designed to operate at a given sample rate and will only accept RAVENNA streams that are presented to it at the proper matching sampling rate. This said if, during this summer, we find out there is some FPGA real-estate left, Merging may add a few channels of real-time SRC in the architecture of the internal Horus router to provide such capability on a limited number of audio channels.

I guess in that case sound differences would be irrelevant since there would be no change without some kind of intentional manipulation. A good argument to utilize analog processors instead of digital ones if you want to remain "pure." Assuming the original recording isn't perfect in every way. Good to know, thanks.

So to Tony, were there multiple Horus' in the recording or were other devices responsible for the PCM part?

Hi Tony:

What's your take on the Horus microphone preamps compared to boutique outboard models such as Millennia, Grace, DAV etc......? I've heard different things. Nothing bad at all mind you, just varying opinions.

I would imagine they are quite excellent given my experience with Merging's design philosophy thus far, however I'm just curious if people are finding them to as good as the best stand alone units out there. I'm seriously thinking about adding a Horus to my remote kit after the New Year.

Best,

Marlan

We had no chance to compare pre's, but what I heard was certainly excellent and very clean. My usual favourite in-house pre's are EAR (tube), Peach (tube) and Audient ASP008's. The Merging pre's were part of the Horus package in a collaboration with my good friends at Classic Sound. Green Room had been asked to supply three multitracks running in parallel at 96k, 192k and DSD. The job was beyond what Green Room could manage on our own because of the complexity of the project within a tight time-frame for set-up and take-down (there was video crew as well covering the same event), and we had insufficient in-house gear for the number of formats and track-count required.

Thanks for the info.

This setup is quite unclear to me. The OP says there were three parallell recordings. So where there three horus units ? How many mic setups ? Where the mics splitted ?

I ask, because I recently tested a DAD AX32, a Horus, and our own Grace m802 plus AD option.
The differences were small, but the list above is in order of best to worse.
The DAD makes me want to upgrade, the horus not (there is some color present, but it is very good anyway, the touch screen is awful and not good enough even for a 10 dollar made in .... smartphone), the m802 probably has had its time (mainly the AD section, however good it is).

Anyway, the differences were very small, and we needed double, identical mic setups. Any difference between two takes were much bigger than the sound between the units.
Several orders of magnitude smaller were the differences between 44.1 and 88.2 Khz.
Which makes me wonder how on earth one could hear a pronounced difference between 192KHz and quad dsd (almost every ADC chip is bitstream anyway, I have heard straight from manufacturers implementing DSD purely for marketing reasons, every digital signal starts as a bitstream, so what is the point if the delivery is CD ?)

Can't help wondering, but has Tony F(aulkner) returned to GS?

The touchscreen is not important on the Horus and is actually unnecessary. All important work is done through the http interface. The Horus is typically deployed remotely to the monitoring position.

I agree, the touchscreen is nice if you absolutely need local control or for a quick local check of settings and status. It is, however, not necessary.

The unit works superbly when used with the remote, particularly from Protools or Pyramix

We´ve also done a couple of recordings in DSD256 using Horus pres and conversion, and it sounds fantastic.

Yes, it pushes the ultrasonic noise; in a DSD256 recording, the 40-100K frequency range stays approximately at -90dB, against the ca. -30dB in DSD64.

We´re really happy with the Horus, it´s a superb unit. A great thing about the latet Pyramix version is DSD render function, with which only crossfades are processed in DXD and the rest remain in the originla DSD format. Recordings made in DSD256 sound great that way!

I bet it does. For me the big questions are:
1) does it sound different,
2) if so, does it sound better and
3) is the DSD256 format responsible for this.

Hi Marlan,

Claude of Merging let me inspect the Horus mic input card at AES show. We had a good shop talk regarding the actual circuitry design of the card. I was very satisfied with all their thoughts and design went into the product. Here are some high lights if you are interested;

1 The entire analog chain is fully balanced, there is no phase summation or splitting going on at all. All the gain management is done in balanced fashion. (Differential)

2 The gain change is done by using low turn-on-resistance digitally controlled switches. The actual gain is set by discreet resisters in the feedback loop. The mic pre is not based on any off the shelf products.

3 The analog amplifier gain is switched at 5dB per step. The in between gain adjustment is done in DSP after the AD converter. The gain adjustment shows up in your application as 0.5dB per step. This is to simplify the analog stage design and reduce cost, I am assuming. Frankly, with today’s highly accurate and low noise AD converter chips, 5 dB gain mismatching does not really affect the audio quality. We both agree on this.

4 If I remember correctly, there is only one analog amplifier stage in the audio chain. Same amplifier is also used for line level input. Very simple circuit thus short analog path. Perhaps not the very best suited for line level but this card was design to be used for microphone inputs.

5 The only capacitors used in the analog circuit is at the input to block phantom power and is not bypass able. The input capacitors and shunting resisters form a low cut filter @3Hz.

6 The AD converter chip is made by ARDA technologies. Very new stuff. According to Claude, this is the best AD chip he can find.


Best regards,

Da-Hong

Thanks a bunch Da-Hong!

As always, you are a veritable fountain of knowledge.

All my best,

Marlan

Credentials of ARDA CEO:
Not bad.

How are they doing DSP on a DSD signal? Does the digital signal originate as DXD or some other high sampling-rate PCM and then get converted to DSD? Or are they starting out with a multi-bit signal (5-8 bits, perhaps?) at DSD rate and then somehow reducing word length to 1-bit DSD?

In any case, I'm sure it's an excellent device and whatever solution they came up with probably works very well. I'm just curious.

I was standing next to Claude and Da-Hong when the conversation was going on at AES. It was pretty awesome to hear them talk shop between 2 people who really "get it".
Dennis

The fact that I can have the Horus as a front end for Pyramix, Sequoia AND ProTools is very appealing to me. Right now I love the preamps and conversion in the ULN-8, so a simple high end AES/EBU sound card may be the answer for the time being before I invest a significant chunk o' change on a bunch of new gear.

Correct me if wrong, but only PMX allows full integration of Horus mic pre gain into the mixer controls. This is a compelling reason to use PMX as your data acquisition DAW when using Horus.

Not to mention ravenna/horus into masscore allowing for hundreds of bidirectional channels vs the 64IO of the asio driver.
This and the ability to use a secondary soundcard for monitoring and talkback without.

Hi,

I didn't mean to imply I was going to use the ULN-8 to control Pyramix. It simply can't function that way and I am aware of that. I meant the easiest way for me to experiment going back to Pyramix with my current set-up is to use my Metric Halo gear as stand alone mic preamps and a/d converters directly to an AES sound card running Pyramix Native.

Once I upgrade to a Horus, that particular system would be self contained and the Halo gear would function in a Mac only environment.

Sorry, David, your assumption is not correct. The Horus box is controlled through a web browser, it has its own web address. So, if the Horus is connected to your computer via a network, you can actually control it using another computer that is also connected to the same network. Horus shows up as a webpage. And any other computers on the network can also record from the same inputs, pretty neat. Merging supplies a regular ASIO to allow you using the Horus with any other host that supports ASIO system. You don’t need Merging to record. But if you want to use their secondary ASIO system you need to use Merging software. (To monitor audio with a different sound card and use your laptop microphone for talkback.)

Best regards,

Da-Hong