Apogee Symphony Users... 1.6 ms of latency? Hype, or the real deal?

Hi. Didn't read all the other posts. In short my recording latency adjustment in logic is at -82 samples for 96Khz, it's 52 for 44.1 Khz. I feel the latency is definitely 1.6ms or better, and to me the calculation is round trip. The second thing you get with Symphony is that you can actually do full recordings with the Buffer set to 1024 and not hear any latency at all. Symphony seems to handle that.

I HIGHLY recommend those cards hands down.

Contact me if you need more details, you know how to find me 🙂

R

Latency as it pertains to the Symphony card is always the same regardless of buffer setting in Logic. I've tested that over and over, and that persists through restarts. 52 samples off at 44.1 and 82 samples off at 96.

R

Huh? I see better latency at 44.1 hahaha better check something on my setup 🙂 As I mentioned in my other post, I also noticed that the buffer setting in Logic does nothing to the latency for Symphony, so setting it at 1024 doesn't affect latency. Am I alone out here and have some kind miraculous setup?

R

I use the Rosetta with Symphony. Track count on average of AUDIO tracks is about 24-26 tracks. Plugins wise I have as many as 30-25 plugs used, but not all different plugs some repeat, unique plugs I'd say about 6-7. I do record at 96. Simultaneous recorded tracks at one point was 8.

That being said, I am much happier at 88.2 because of the down sampling that happens when I need to bounce. I feel like the tools I have to sample down from 96 are not as good and I lose quality. 88.2 with the Rosetta 800 has been phenomenal as well.

R

i am wondering if you have factored in plug-in delay compensation?

i feel a bit like a ford escort owner exchanging notes over the specs of a lambourghini, but if i understand it correctly, part of the whole really good latency with symphony is down to being able to keep latency down through the D/A A/D conversion and running through PCIe which doesn't incur latency in the same way as firewire.

actually this may explain why my experience of latency with a buffer set to 512 is acceptable - i am running through a PCI system. hmm....getting out of my depth now.....

1024, 512 or what ever in between down to 32, latency is same on Symphony in my experience. So I don't set it to 32, the only time I jump down from 1024 is if I am tracking in BFD, and I set it at 128 there. Then once I'm done tracking, back up to 1024 it goes. The rest of the time with all audio tracks, I stay at 1024. The fact that I use UAD and Liquid Mix as my main plug helps a lot too of course, and the fact that I use the UA DCS to help a guitar player or singer record while listening to the music in headphones off of the DCS helps tremendously.

R

Rohan,

My friend, there is no latency running signal through an analog desk. If such were the case, having speaker cables where one was a few inches or even feet longer than the other would theoretically cause L/R to be slightly out of phase. We're talking lightspeed here, which is a much higher baud rate than digital audio. 😉 Those few inches/feet just don't matter. (You're probably saying the same thing about my pwecious widdle 1.6 ms LOL!).

Second, natural latency (sound through air, orchestral players separated over the span of a stage, pipe organists) cannot be overcome (other than to lower the humidity and raise the temperature in the room. Or is it raise the humidity and lower the... ) 😉 It's a given. My argument is that additional, artificially induced latency is undesirable. It's like spreading those orchestral musicians further apart on stage. Per your example of the click, the drummer's live sound -- as monitored -- will always be behind the click no matter what, unless the audio of the click and his generated audio have no latency-inducing thingees (technical term) in the middle.

ITB monitoring on any system is not up to the performance of a simple analog monitoring desk. It is not an improvement. It's an artifice. Direct monitoring systems, as found in Lynx and RME, come much closer, if not as close as can be reasonably expected, to the ideal of analog monitoring.

Speaking of long arms, have you ever seen Sparks in concert?

sorry iS but you are wrong in one fundamental aspect - and that is that the latency is added to existing latency.

for example, in my example of the orchestra, the time it takes sound to reach one side of the orchestra to the other can be beaten by a native system. therefore any discussion of latency at that sort of speed is virtually moot. you could in theory hear the sound of the clarinet through your cans before you heard it through air. to say that the latency is "artificially" induced seems beside the point in terms of monitoring.

secondly, (i should quickly point out i am paraphrasing word of mouth here but i believe this is in essence correct) analogue desks do induce a very small amount of latency through the capacitors used in gain staging. it is not as fast as light.

but again its beside the point. i think once you get below 5ms of latency you are in the realms of as near to zero as makes no difference. 5ms is still only 1.7m away. if you are telling me a musician can tell the difference between 2 sounds sounding simultaneously at say 0.1 m and 1.8m (ie a 1.7m difference) i am likely to say i don't believe you....so don't push me.... 😉

ITB monitoring on any system is not up to the performance of a simple analog monitoring desk. It is not an improvement. It's an artifice. Direct monitoring systems, as found in Lynx and RME, come much closer, if not as close as can be reasonably expected, to the ideal of analog monitoring.

but how are you going to monitor with FX? you would need hardware (which inevitably has latency as well) for that.

sorry iS but you are wrong in one fundamental aspect - and that is that the latency is added to existing latency.

I think I made exactly that point, referring to it as "artificially induced" latency.

capacitors...

Nope. A capacitor may take time to charge and discharge, but it doesn't need to fully charge before it passes signal. And all capacitors have a specified tolerance. Tantalum capacitors typically have 1% to 5% tolerance in their rated capacitance. Electrolytics: typically 10% to 20% tolerance. According to your theory, if you constructed a board's circuitry using electrolytics, the differences in rated capacitance (due to tolerance) would cause any two channels to have differing amounts of delay. And were you to take the same signal, split it, run it into two channels and then sum the outputs you'd have some kind of phase cancellation (less in a board made with the lesser-tolerance tantalums) but this is simply not the case. It's kind of like saying that you could make an analog delay line out of a few miles of cable coiled on a spool. It just doesn't work that way.

True, there is some delay in analog electronics, i.e., the amount of time it takes for an electron to enter the input and come out the output, but it's measured in near-light speed, and is, for all intents and purposes, a moot measurement when compared to the speed of sound traveling through air. Or digitally induced latency.

clarinet...

Finally... if you have a clarinet player 100 feet away from a horn player, sure, they'd learn to play in time, because humans are good at adapting to that kind of thing, within reason. If there were a mic on each player and each was monitoring eachother, the latency induced by the monitoring system would indeed be a moot point; their performance timing would be better in that situation. But wouldn't they play more in sync if they were seated next to one another? I think that's my point -- that reducing (if not eliminating) latency to the barest possible amount (ideally zero) is best. Anything else is futzing with half-baked techology.

By the way, is this the £5 argument or the £10 argument? 😉

EDIT:

but how are you going to monitor with FX? you would need hardware (which inevitably has latency as well) for that.

Yes. Hardware. Hardware could save 10's of thousands of utterances of "I'm hearing myself out of time in the headphones", "I'm playing through GuitarAmpPluggy and everything I play feels late". And so on... The technology is great, I love it, I use it, but it's behind the curve when it comes to the performance of analog electronics.

Message was edited by: iSchwartz

Rohan Stevenson1 wrote:
but again its beside the point. i think once you get below 5ms of latency you are in the realms of as near to zero as makes no difference. 5ms is still only 1.7m away. if you are telling me a musician can tell the difference between 2 sounds sounding simultaneously at say 0.1 m and 1.8m (ie a 1.7m difference) i am likely to say i don't believe you....so don't push me.... 😉

Hey, I'm not iS (not even close... just check my number of posts for confirmation 🙂, but I'll certainly confirm that not only can musicians hear this, pretty much anybody can hear this. If you set up two stereo speakers right next to each other and run a click track through them, then start moving one of the speakers farther and farther away, you'll find that even just a few inches will cause some phase artifacts if you listen carefully. And move one of the speakers 2m away, and it will be obvious that there are two sound sources (i.e. your ears/brain will resolve the sounds individually).

(I experienced this effect while fine-tuning the position of my stereo speakers in my living room. The goal, of course, is to form an isosceles triangle between the speakers and the listener. An identical sound originating from both speakers should reach the ears at precisely the same time at precisely the same volume. Any deviation from that weakens the "phantom" image between the speakers. My experience was that when those distances were more than a few inches off, the difference was audible.)

Also, consider that the ear/brain's ability to identify and localize sounds depends essentially on latency cues (typically called the "interaural time difference"). It has been well-researched that the ear can distinguish a time shift between the two ears when it is a mere 10 MICROseconds. Not exactly the same thing as feeling the latency of a DAW, but the point is that the ear is amazingly sensitive to time and phase, so unless you've actually tried a blind A/B test yourself, don't make the mistake of underestimating the ear! 🙂

James

FWIW, Take in consideration the close integration between Apogee and Logic, and I would also take into consideration the company behind the product. I can tell you after having used the Symphony and Rosetta 800 together for over a year now, and calling in for support a couple of times, the Apogee staff and engineers are stellar. Lastly, the combination of card and converter from same manufacturer usually guarantees better interoperability, and when you ever need to troubleshoot anything, there's no finger pointing from one side to the other.

HTH

R

ps. you can also grow your converters by using more Apogee box combinations and more symphony cards as you need it.

iSchwartz wrote:
Finally... if you have a clarinet player 100 feet away from a horn player, sure, they'd learn to play in time, because humans are good at adapting to that kind of thing, within reason. If there were a mic on each player and each was monitoring eachother, the latency induced by the monitoring system would indeed be a moot point; their performance timing would be better in that situation. But wouldn't they play more in sync if they were seated next to one another? I think that's my point -- that reducing (if not eliminating) latency to the barest possible amount (ideally zero) is best. Anything else is futzing with half-baked techology.

I agree completely. And, I think, so will anyone who has ever been in a band that was used to playing in small clubs, then suddenly got a spot on a large stage. The physical feeling of being even 10 feet farther away from the drummer than you're used to is very disconcerting.

It is a fair point that playing in the studio is inherently different from playing live. For instance, if I'm used to hearing the snare drum from 15 feet away, and now I'm hearing it close-mic'ed through headphones, I may actually be hearing it earlier than I'm used to. In the same way, if I'm used to playing a snare drum with my head 4 feet away, and I put a mic 1" away from it and run the analog signal through my headphones, I may actually hear the drum sooner than I would acoustically. And when the acoustic bleed reaches my ears, it may phase slightly with the headphone sound. Theoretically, a very small amount of latency might actually sound better, to make the monitored sound line up precisely in phase with the acoustic sound I'm hearing.

But, my experience has been that the amount of latency you're typically considering in a good digital system (1-5ms) is more distracting than anything else, especially when it comes to phasing between a headphone mix and the acoustic sounds that bleed through. And try singing using a fully analog signal path, then introduce even 1ms of latency... your voice will sound and feel very different.

I've worked with many, many people who never complained (and presumably never noticed) 4 ms of latency in their headphones. However, try this sometime: start out with an analog monitoring path, then switch mid-session to a digital path with some latency... I'll bet you'll get some complaints then.

James

Guess I am coming in right at the back end after following this thread I am so Glad I monitor through my 32 8 2 Mackie desk as I have always done I use my tried and trusted Hammefall Multiface RME with a buffer of 64 on my G5 dual 2gig I may be using dated technology but I am comfortable and have found my seat to work in. The musicians and Vocalist I have worked with have never complained. and coming from a session vocalist background I have not had any problems on my rig its just a breeze and a joy to work with. I am also recording well over 48 trks to boot with a good amount of effects and this my friends is on logic pro 7.1.1 on tiger 10.4.11 with 3.5gig of ram and 2 terabyte of HD space. yes sometimes we record at 96k pending on the projects requirements and yes we can hear the difference and we love it that not all our projects are recorded at this sample rate as this brings a different flavor to the project.

After all is said and done I just love making good productions and working with great talent well peace to you all got to get back to the board all this talk is creating my own latency in getting back to my project at hand.

well, i suppose it boils down to this:

1) an argument over whether to monitor through software or hardware. personally, i have been really happy to monitor through hardware on my dodgy old desk etc etc. there is a principle that some engineers use, is that they want to hear what is being recorded sound like it will when it is being mixed. if the mix is going to be done ITB with those effects they will need to software monitor. also, it quite possible and practical these days to do away with a desk altogether. no external gear at all. in order to put some effects on the mix going back to the artist or artists, you would need to do it this way.

2) whether or not latency, or the differential is perceivable at distances of under 2 meters. i am deeply skeptical about that - that it makes any sort of difference at those levels. and this is probably why i sound defensive of software monitoring.

as to analogue desks - i should point out i have no particular theory of my own, i was possibly inadvisably spouting off something second hand which could quite easily be wrong, and it was being made as part of broader point to me that i came to accept that there is delay - latency - everywhere. that zero latency just doesn't exist.

so i guess it comes down to whether you need to do it via software or via hardware. recently i have been doing it via software and no one i have recorded could tell the difference (including myself) and this is at much higher latencies.

as to my orchestral analogy, that really is to try and demonstrate that latency exists anyway. the same clarinet player may just have had to be playing in a wind quintet and then dealing with basses sitting 10m away. what difference does it make if the drummer hears what he plays fractionally early (zero latency) exactly on time (low latency software monitoring) or fractionally late (higher latency software monitoring)? it's everywhere all the time, that very little extra latency makes no difference - is what i am trying to say in my round'about way.

god we haven't a meaty issue to chew on for ages. it's very nice - i need the distraction....