The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Burl B2 vs Apogee Symphony Digital Converters
Old 10th November 2015
  #31
Quote:
Originally Posted by citrusonic View Post
thank you so much for helping me shake the 2200 dollar burden that the Burl was going to be. It's nice to get advice from people who aren't perpetuating the gear lust just for the sake of having a new piece, however cool and green the new piece may be!

going to be putting one of the Silver Bullet stereo tone amps in my 2 bus chain, and I think that will give me the majority of the saturation, or vibe I've been wanting. i could put the 2 grand to better use- like new monitors, or a nice compressor.

thanks for the heads up!
No problem. I'm certainly no pro, but I can churn out a decent mix, and I'm finalizing a release here now. I can certainly send you a private message on the release date/info (if that'll help you solidify an answer).

For my tubes and trannies, I went with a Sonic Farm Creamliner II. It's a great box if you want to add some density, and 'smooth' out a recording. It's the perfect match for the Apogee gear (IMO).

Another thing you may want to consider: adding additional converters means additional complexity. It's nice keeping things in one box with one clock (and not slaving a pile of stuff). You may run into unwanted 'jitter' issues, which is why some guys are using an extra computer to print off a higher rate mix.

I'd kill for a MADI setup involving a Burl Mothership.., but I'm more than happy with using a few choice outboard pieces, and my Ensemble Thunderbolt. I love keeping things simple and stable.

Cheers,

Phil
Old 10th November 2015
  #32
Lives for gear
 
CJ1973's Avatar
 

Thanks Phil for the insight! Refreshing. For me, I'm looking to demo the Burl first and see how it compares with Symph for tracking. The mixing is going to be done on an analog console and then I was looking at the Burl as 2 bus into another computer at higher rate. If you were in my shoes, would you take a different approach?
As an FYI, here is something I wrote and produced, mixed at my place on my console, but sent back into Symph and 2 bus was plugins. I'm looking to add a bus comp like Vari Mu and Burl for the mix. Thanks!
http://youtu.be/kg9enIuu_yA
Old 10th November 2015
  #33
Lives for gear
 
deuc647's Avatar
 

Quote:
Originally Posted by kmetal View Post
I've looked a lot into the connectivity options since I'm considering a MS (8x8), and I'm just about sold on the Dante option. 2ms latency, minimal cabling, the ability to route and dup channels digitally, and the fact that big companies like Yamaha and digico, as well as many others are Dante compatible (with the option card) it's the way to go imo.

Also, it can carry multiple sample rates simultaneously. And different formats. Video, audio, control surface data.

Ive read some success stories with it, on GS and elsewhere. Overall I think if burl is going to put their name on it, I have no reason to believe it doesn't work as prescribed. If the format, or the jack/card cheapened sonics burl would not use them on a product whose sole purpose is to handle and maintain audio fidelity.

Also, the advantage w the burl, is you simply buy multiple MOBO cards, if you needs various connection protocols. This is something not other high end converters do, or at least do as easily.

The minimal cabling, wide compatibility, and the high track count / sample rates sold me. I think as 10/100/500gbs speeds become more common, it's going to be easier to stay current hardware wise, a little longer. I hope to have mine powered up and racked in about 6mo.
Just a heads up on this, dante adds 2ms of latency to what you already have. It doesnt give you 2ms unless you get a pci/pcie card from what i understand
Old 10th November 2015
  #34
Lives for gear
 
Squawk's Avatar
Quote:
Originally Posted by CJ1973 View Post
As an FYI, here is something I wrote and produced, mixed at my place on my console, but sent back into Symph and 2 bus was plugins. I'm looking to add a bus comp like Vari Mu and Burl for the mix. Thanks!
http://youtu.be/kg9enIuu_yA
Cool track, nice work!
Old 11th November 2015
  #35
Quote:
Originally Posted by deuc647 View Post
Just a heads up on this, dante adds 2ms of latency to what you already have. It doesnt give you 2ms unless you get a pci/pcie card from what i understand
Yes I think so to. 4ms is the spec for standard RJ-45 ports/cards. It will run as low as your hardware will go and it seems like a 10gbs pcie card is about $500. I think this is the transfer rate that the Yamaha and focusrite cards are using achieve those specs.
Along with source connect remote recording / mixing / collaboration software, Burl with Dante makes a lot of sense and is actually sort of bang for your buck in the high end. It's super versatile with the (non hot) swappable mobo cards wich are about $500. Killer conversion, and compatible with any typical computer, and able to use any lan to transmit audio and data (video too I think...) it's the future imho. I'm excited anout this and am looking to pull the trigger this year.

I don't think Dante adds any latency per se, it runs off of an asio driver just like any other 'sound card'. If it added latency how would it stay in sync with another interface say usb connection? Maybe that's where the card comes into play?

I got the gyst that 2ms was with a Dante card, competing with the digink mobo 1.8 ms with PT native card. .7ms would be with mobo going to a PT hdx card.

I may have it misunderstood, there was a Dante thread with a rep here when it first was announced, it'd be a good question for them. Very curious about all this.

Last edited by Kyle P. Gushue; 11th November 2015 at 09:03 AM..
Old 11th November 2015
  #36
OK... This has the potential to cause some serious confusion - as I have mentioned elsewhere - please stop throwing numbers around as "gospel" as this leads the less informed confused, and potentially prevents others from finding solutions that suit them.

Latency numbers -

There is no "set in stone" latency setting for "Dante" - this is the truth. This has obviously been misinterpreted by the lazy in a variety of weird and wacky ways... So what does that statement really mean?

Dante is implemented both as software and hardware solutions (hardware is split into low-cost chip, and FPGA solutions).

An operating system running on Mac OS or Windows has other tasks to perform. In order to have totally reliable audio performance, a (relitively) larger latency buffer is used - this is typically set at 4ms on the virtual soundcard (a $30 64x64 channel piece of software that uses the network port of your computer to send and receive audio)

Conversely the PCIe card (as sold by inter alia Focusrite and Yamaha) can have its latency performance set to as low as 150uS (thats microseconds). The PCIe card is a Dante enabled hardware endpoint (FPGA) mounted on a PCIe card - which in turn presents itself to the operating system as a 128x128 channel (at 96KHz) soundcard (in much the same way as any other PCie soundcard).

This leaves Low cost chip (Ultimo) devices - 1ms

And Brooklyn II (FPGA) solutions - can be as low as 150uS, but typically 250uS.

So what does this latency number actually mean?

When we describe "latency" in Dante we are talking about the time it takes for audio presented to Dante to be presented as "audio" at the receiver.

For example - if we send a 4ms set Virtual soundcard to a 150uS set PCIe card audio coming out of the OS will appear 4ms later at the operating system of the device with the PCie card installed.

This latency is synchronised to +/- 1uS at any point in the network.

If I send a signal from a hardware mic pre to the same PCIe card - both set at 125uS the time taken from the output of the AtoD in the mic pre to the OS attached to the PCIe card would be 150uS.

To put this another way - if we could send the same signal at exactly the same time through the mic pre and the virtual soundcard - the audio from the mic pre would arrive (and be played out) 3.85mS earlier than that going from the Virtual soundcard.

In reality - the circumstance in real life where this may matter is sending the same signal from different network devices to the same device.

If the two sending devices are hardware identical and their latency settings are the same - then the playout at the receiver will be +/- 1uS.

If the latency is somehow increased on one vs another (inline processing etc) then there will be a timing difference.

Again please remember - this happens with analogue too - its just it is very much easier to attach numbers to this in a digital system that is telling you... Rather than grabbing an oscilloscope (physics doesnt care that you cant be bothered to measure something).

So hopefully that explains latency from a Dante perspective - in terms of how you use it and what you experience - this information is of surprisingly little value.

I see on this site that "MADI is fast" - ok...
I never see "Dante can be faster than MADI"

Both are true - both are a little bit pointless.

The significant contributor of latency is the time to deliver the audio to the Dante (or MADI) interface, and the time to take it from presentation at the receiver to the ultimate destination.

This overall latency figure is the interesting one - and for that - you probably need to get your measurement system of choice to get a real answer. Perhaps... If more people actually did the measurement of the whole signal path there would be fewer people in the world who believe that they have some kind of biologically unlikely aural sensitivity, that contradicts a lot of very carefully researched science
Old 11th November 2015
  #37
Lives for gear
 
deuc647's Avatar
 

It would be nice if you guys could do it since you guys are the company who puts out the virtual sound card.

This place is a site that buys gear based off the recommendation of other peers. I dont doubt that dante and the VS card are legit, but i would like a demo system or a youtube video or something to show what it does. I thinnk what most people(i think) want to see is a system with plug ins being tracked thru to see if you can get latency to a minimum without being audible. Basically people want a PTHDX system without having to be tied to PT. I had that thread in the music computer section about dante and i remember reading the responses and thinking, im just as lost as before i asked
Old 11th November 2015
  #38
Quote:
Originally Posted by deuc647 View Post
It would be nice if you guys could do it since you guys are the company who puts out the virtual sound card.

This place is a site that buys gear based off the recommendation of other peers. I dont doubt that dante and the VS card are legit, but i would like a demo system or a youtube video or something to show what it does. I thinnk what most people(i think) want to see is a system with plug ins being tracked thru to see if you can get latency to a minimum without being audible. Basically people want a PTHDX system without having to be tied to PT. I had that thread in the music computer section about dante and i remember reading the responses and thinking, im just as lost as before i asked
QED

The significant latency in the setup you describe is mainly in the DAW and plugins - we dont actually make those... So measuring some theoretical things will be even more potentially misleading.

Have you ever measured impulse group delay through different loudspeakers on analogue? Thats also interesting - comparing different loudspeakers can give group delay deltas in the 5-10 ms region... Again - I dont make analogue systems... So suggest you make your own measurements.
Old 11th November 2015
  #39
Thanks for your detailed response. So if I have this right, using the Dante connection as opposed to aes, digilink, sub, Ect, doesn't introduce any additional latency to the latency involved in the ad-da? And when using host based pluggins, the latency experienced would be typical of what the buffer size is, and whatever the pluggins introduce?

In other words when I compared the digi link, and Dante cards, it seemed to be a good buy, to live w dantes published specs, of 4ms, as opposed to the 2ms and under of the other options. Doing that would save. the cost of an HDX/HD native card, a madi card, for a later time if needed.

Does the connection/networking cost additional latency to what one would expect from any other typical Audio interface? Sorry if you feel you answered this already, I'm just trying to get it right.

Would 10gb network card perform better than say a stock 1gb card? Or are the higher performance specs related directly to the hardware support on the Dante cards?

I appreciate the time, as this is going to be a healthy expense, when the time is right. Want to make sure I make the right call. It's not about the price so much as generally expected performance. Not that I think it doesn't work, I believe it does. I'm building a whole computer/rig from scratch completely, so every little piece becomes so critical.

maybe vintage King, or burl has some tests they could show us or tells us about. They must have at least tested it with the coffee can Mac they have pictured on the website under the Dante section. You start to add up the cost of pcie cards, and snakes, and the Dante/burl option isn't as boutiquely prices as it seems on the surface.

It's not just being tied to avid or not, it's trying to figure out how every Johnny sore throats rig interacts. The idea of seamless plug and play is cool, just wondering what to expect when things are playing fairly nice in the field.

Last edited by Kyle P. Gushue; 11th November 2015 at 11:38 PM..
Old 11th November 2015
  #40
Quote:
Originally Posted by CJ1973 View Post
Thanks Phil for the insight! Refreshing. For me, I'm looking to demo the Burl first and see how it compares with Symph for tracking. The mixing is going to be done on an analog console and then I was looking at the Burl as 2 bus into another computer at higher rate. If you were in my shoes, would you take a different approach?
As an FYI, here is something I wrote and produced, mixed at my place on my console, but sent back into Symph and 2 bus was plugins. I'm looking to add a bus comp like Vari Mu and Burl for the mix. Thanks!
http://youtu.be/kg9enIuu_yA
This is a fantastic production.

I'm at a loss as to why you'd want a Burl when you're churning out stuff like this?

Look at getting a Sonic Farm Creamliner II. This will give you a wonderful way of adding density to your tracks.., but, after hearing that tune, I really don't think you need anything except to continue making music.

I'd say add hardware. You may inadvertently add unforeseen complexities if/when you start adding converters. Why not just expand your Symphony?

Keep up the awesome work!

Cheers,

Phil
Old 12th November 2015
  #41
Lives for gear
 
Squawk's Avatar
Quote:
Originally Posted by Palaver View Post
This is a fantastic production.

I'm at a loss as to why you'd want a Burl when you're churning out stuff like this?

Look at getting a Sonic Farm Creamliner II. This will give you a wonderful way of adding density to your tracks.., but, after hearing that tune, I really don't think you need anything except to continue making music.

I'd say add hardware. You may inadvertently add unforeseen complexities if/when you start adding converters. Why not just expand your Symphony?

Keep up the awesome work!

Cheers,

Phil
Funny enough, I've got a creamliner here today on loan from Sonic Farm. I'll be running a few things through it, but out from the Mothership D/A Will be interesting to hear what it does in that combo.
Old 12th November 2015
  #42
Quote:
Originally Posted by Squawk View Post
Funny enough, I've got a creamliner here today on loan from Sonic Farm. I'll be running a few things through it, but out from the Mothership D/A Will be interesting to hear what it does in that combo.
Cool.

Last weekend I did a test between the Burl and Ensemble Thunderbolt. As I stated earlier, the differences were negligible (IMO). I'm not going to lose sleep over subtleties in that department.

There was far greater impact when I added hardware.

The mastering engineer is using the tracks I've ran thru the Creamliner II for the most part.

I'd love to hear your thoughts on the unit. My unit has the iron tranny option. I'm mostly running it in Pentode mode with the transformers engaged. That may be a bit much for you considering your in Burl territory! Is it helping or hindering your work!?

Cheers,

Phil
Old 12th November 2015
  #43
Thanks for your detailed response. So if I have this right, using the Dante connection as opposed to aes, digilink, sub, Ect, doesn't introduce any additional latency to the latency involved in the ad-da? And when using host based pluggins, the latency experienced would be typical of what the buffer size is, and whatever the pluggins introduce?


First - thank you for some very good questions. I'll endeavour to give you honest answers... And of course i'm happy to clarify if needs be.

Comparing these (for want of a better word) "transport" latencies of various methods is far from transparent.

For example a TDM-based system such as AES3 or MADI concerns itself with a single sample capture style method. Depending upon how each end of a cable is buffered (digitally) it is likely that the latency on a point to point connection will be lower than a packet-based transport.

Lets look at 48KHz

A single sample is 20uS (and some change) - I need to capture a sample (it takes 20uS to do this - when I have prescient equipment I shall retire to my yacht! ) in the best case I capture it at the other end of the wire and pass it on (another 20uS) so my point to point wire connection on a TDM system theoretically could be as low as 40uS.

If I go through something like a MADI router - I need to extract and remap to re-transmit theoretically another 3 samples (60uS) then it is likely going to have to go down another wire - add 40uS. These numbers are theoretical to avoid major dispute... But add that up and you get 140uS or thereabouts.

In a Dante system we wait for several samples (this helps the network function efficiently) - if we wait for say 4 samples we have a capture time of 80uS. We then slam that in a packet... Encapsulate the packet in a frame and that gets transmitted to wherever the destination is (switched networks do this very nicely) at the same time we ensure that all devices have their presentation clocks synchronized to +/- 1uS.

Switches are not magic... They do introduce signal propogation delay (however a gigabit network switch will have a low latency Dante packet on its way within about 2uS). To get around the unknown number of switches between devices, and any minor traffic delays, we define a safety buffer. This means that if I transmit a packet on a system set for 150uS the packets are probably arriving at the end within 80-100uS. Think of it like going to a meeting - if you get there early, you have time for coffee (our packets havent heard of coffee so dont worry about adding a grinder to the kit list on their account ) when presentation time arrives (that 150uS latency value) the data is released from the buffer into the next stage (DAW/DtoA etc).

So to be fair there are circumstances where Dante Latency may be higher than other methods. Remember that the overtaking happens when more complex (dare I say "real" systems are built) this is because you only pay the "latency penalty" once on a Dante system, as opposed to at every router in a TDM system.

Regardless of this - I have made enough records using multiple MADI routers... And nobody has really noticed. I can also say the same about Dante. But as the whole subject of latency is quite personal... I would rather you formed your own opinion - I can only provide (hopefully honest) comparison information, and an explanation of how it works.

I agree - host- based plugins make the most difference in a "real" signal path - hence my resistance to making any claims in this direction.


Would 10gb network card perform better than say a stock 1gb card? Or are the higher performance specs related directly to the hardware support on the Dante cards?


The Dante PCIe card is effectively a dedicated computer on a chip, with a 1 gigabit network interface - this dedicated device is what achieves the low latency performance.

Conversely we have profiled DVS to work on a 1gigabit common network interface typical of a production computer.

In this respect - a 10gig interface will not give you any better performance from a Dante Audio perspective. The bottleneck is in the operating system - typically not the network interface hardware of the computer.

Obviously computers vary in performance quite greatly - I think few would find the idea that a higher spec computer will process host based plugins faster (as mentioned probably a greater overall latency contributor than the audio transport chosen).

If we extrapolate this theory, then you could well point out that theoretically higher performance could be achieved with the Virtual card- this is difficult to argue with... But in terms of dealing with the here and now - its probably not a worthwhile discussion... And there are going to be factors that cleverer people than I would consider in this respect. There are other interesting things in software to tackle first, and the PCIe card provides a great solution for high performance right now.

I hope this helps - please feel free to ask further questions.

Old 12th November 2015
  #44
Lives for gear
 
CJ1973's Avatar
 

Quote:
Originally Posted by Palaver View Post
This is a fantastic production.

I'm at a loss as to why you'd want a Burl when you're churning out stuff like this?

Look at getting a Sonic Farm Creamliner II. This will give you a wonderful way of adding density to your tracks.., but, after hearing that tune, I really don't think you need anything except to continue making music.

I'd say add hardware. You may inadvertently add unforeseen complexities if/when you start adding converters. Why not just expand your Symphony?

Keep up the awesome work!

Cheers,

Phil
Thanks, Phil. Really appreciate the kind words.
I am thinking demo-ing the Burl to see if there is a significant jump or otherwise. Else, Apogee it is.
I've noticed that a lot of folks mix into another computer on a very high sample rate (96 or 192). I won't be able to do that if I was to send data from the console back into my Apogee, which I do my sessions at 44.1 mostly. Any thoughts on this?
The other question is in regards to the Creamliner you mentioned. I checked out their website and it appears like a 'glue' machine? How does it compare to a 2 bus comp? Do you use another 2 bus comp and a Creamliner or are the roles totally different?
Thanks!
Old 12th November 2015
  #45
Quote:
Originally Posted by CJ1973 View Post
Thanks, Phil. Really appreciate the kind words.
I am thinking demo-ing the Burl to see if there is a significant jump or otherwise. Else, Apogee it is.
I've noticed that a lot of folks mix into another computer on a very high sample rate (96 or 192). I won't be able to do that if I was to send data from the console back into my Apogee, which I do my sessions at 44.1 mostly. Any thoughts on this?
The other question is in regards to the Creamliner you mentioned. I checked out their website and it appears like a 'glue' machine? How does it compare to a 2 bus comp? Do you use another 2 bus comp and a Creamliner or are the roles totally different?
Thanks!
Hi again,

I noticed that I was getting a bit of jitter when I was clocking to the Burl last week, so I stuck with the single converter (on my final prints). I was slaving to the Burl (I didn't try the other way around). I was using a word clock cable. I didn't thoroughly test it, and should've tried clocking it direct via SPDIF, etc. However, the cracks and pops were so subtle that I didn't notice it until the shootout was over (and I tested on my headphones). This is the main reason why I'm recommending you stick with a single clocked set of converters. I complicated the process, and got burned. But, luckily, the converters weren't a 'night' and 'day' difference. It was the hardware making the difference.

You have a Symphony, which is supposedly a step up from what I'm using. Even Apogee admit the Ensemble Thunderbolt is a step up from the Rosetta series (which are great converters), but it's no Symphony. The reason why I went with the Ensemble is simply for utility/workflow purposes. It has better headphone amps, and the layout is so much more conducive to the 'tracking/writing' process. It's brilliantly designed and has no latency. For me, the utility/tidiness/convenience factor (and lack of fan noise) won out over the Symphony. But do I want a Symphony? Yes. Do I want a Burl? Yes? I made my choice based on utility and cost.

The Symphony is world-class. Getting another converter isn't going to make or break your projects. If it works, then I'd stick with it. If you don't like it, then start running demos on other stuff.

If I had an unlimited budget, I'd be running a full Burl Mothership setup. But, unfortunately, I'm a normal dude with a day job and bills to pay. That's life.

As for the Creamliner, I haven't used it a ton. Mine is brand new. I just ran my tracks through it, and it compliments my Elysia Xfilter & Xpressor (which are ultra clean) quite nicely. It provides the 'density' which those units are missing. I gave the mastering engineer both raw unprocessed prints, and then another version where I 'creamed' my mixes. The mastering engineer went with my hardware versions. He has a beautiful room. If he's picking my hardware processed stuff, then I know damn well that my Apogee isn't the problem.

The Creamliner is similar to the InnerTube SumThang; the tubes really are doing 'something.' No pun intended! I guess you could call it a very subtle compression.., but glue would certainly be a great term for it. It's subtle, and you can push stuff harder. I believe the Sonic Farm stuff is the perfect compliment to a brand like Elysia.

I hope this helps?

Cheers,

Phil
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump