1 (edited by TNM 2017-09-18 10:36:49)

Topic: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

Hi, I think the 802 is probably right for me.. The reason is, that i have the last imac before firewire was dropped, so i have a native FW800 bus not being used for *anything*.. and by using it i will lave both thunderbolt busses free for drives (many drives don't have pass throughs unfortunately, especially the silent SSD types).

I use pro tools 12, so i am stuck to a maximum of 32 inputs for non avid hardware... This is why i am after something that can give me 32 analog inputs or close to it.. I have TWO focusrite octopre mk2 adat interfaces already for 16 inputs.. so 16 + 12 on the 802 equals 28 analog inputs just like that.. which is pretty close to a 32 input analog mixer..

Basically what I want to do is use total mix like an analog mixer, with ALL the analog inputs live in realtime, at all times. I will never monitor through the DAW itself.. I have a hardware guitar box so do not care about the lack of amp in total mix, i just need a reverb, delay, comp and EQ.. which it has.. without at all taxing my 5.5yo imac.

That way i can leave my PT buffer even at 512 and there will be no issues.

BUT, i have literally searched the last few hours and simply can not find what the total i/o latency through total mix is, when using dsp EQ and compressor.

Also, can you confirm that i can use these hardware fx as well as the reverb send on the adat inputs?

I would be very interested in 44k RTL - I just don't see any benefit of recording romplers and digital hardware 16 bit synths at 96K, however it would benefit my analogs.. SO, just for comparison, i would also like the 96K RTL through the DSP, and if that affects compressor/eq count at all.

Thanks so much, my first post.. looooong time lurker, and if this has the right latency for me, will be my first RME interface!

Right now I am using an apollo, but the low latency through the mixer is a fallacy once you start adding uad plugins.. using a standard eq/comp chain in apollo right now and a reverb, at 44k, i am monitoring through almost 8ms latency! and yes, through apollo console! Cheers and thanks!

PS, if the UFX+ figures are significantly different, i wouldn't mind knowing those also.. I like the durec idea, and i also like thunderbolt.. unfortunately i do not like the lack of TB pass through, and it's twice the price of the 802.. remember i can't make use of any more than 32 ins with pro tools, so the massive amounts of inputs isn't a drawcard, as i am not going to change my daw. Only significantly higher FX counts and lower latency would be a drawcard.

2

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

TM FX latency is about 1 sample. Yes, we render everything in real-time. BUT you do not have access to these 1 samples. Because you need to get the signal to TM FX, either by analog I/O (see manual for a few samples more), or via Core Audio (play/record from the Mac), which involves the DAW set buffer size plus Safety Offset. I would expect around 4 or 5 ms RTL with a buffer size of 64 samples (can't check this right now).

Regards
Matthias Carstens
RME

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

that's impossible kind Sir.. any ad/da converter would have more latency than that. Great that the fx are only one sample, but i am asking what the roundtrip latency from input to output is through the dsp mixer. i.e i plug in a guitar into the hi Z..how long from strumming to hearing the sound? NOT through the daw.. once again, i am not interested in monitoring through the DAW, I just want a low latency dsp mixer.

4 (edited by TNM 2017-09-18 18:15:28)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

Sorry, I just realised i may not be explaining this clear.. to give a perfect example of what I am after..

the figure in the apollo console hardware mixer would be 1.1ms a/d to d/a through the dsp mixer with no plugins at 96K, and 2.23 ms at 44.1K.

Avid pro tools HDX would be 0.7ms at 96K..

and so on.. I hope this clears it up. What i am after is the total monitoring latency just through the hardware mixer itself.. Again, i will always bypass the DAW.. Even if i do not buy the product for my main studio room, i am very much looking for a second interface to use with my laptop.. in that case i might get the thunderbolt one though.. I may stick with my apollo in my studio, even though the latency time in console is abysmal when using FX.

So, in a nutshell.. I have all my synths, a vocal and a guitar plugged into the analog ins and monitor them in total mix and use total mix FX where needed.. This is simply replicating an analog desk for me and the way i am used to working.. When i want to use native fx and mixdown time has come, i simply record all those tracks dry to audio in one pass. What i am talking about here with total mix is purely for analog style monitoring purposes.. if the delay is under 2ms at 44K, that's what I am after.

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

This is what Cubase reports from the RME driver, RTT latency:

http://www.tonstudio-forum.de/index.php/Attachment/1931-UFX-UFX-RayDAT-Latencies-jpg/

Full article here: http://www.tonstudio-forum.de/blog/inde … 8-RME-UFX/

I use a combination of 2 UFX+ .. one as main interface and one directly for a recording rig, see here:
http://www.tonstudio-forum.de/blog/inde … arren-Rig/

The RME driver latency is so small that you can plug the guitar into the Instrument input of an UFX or UFX II or UFX+
and then play through an virtual amp inside of the DAW. For this the signal has to travel from the interface to the DAW, will be processed by the VST, goes back to the recording interface to either speaker or phones.

It was possible to use a ASIO buffer size of 256 to play the guitar through the virtual amp.
Better take 128. But with the UFX+ its also possible to use only 32 ASIO buffers.

This is all easily possible ... and as you say you dont mean the whole Round Trip Time ...
You mean mixing on the recording interface ... this little AD/DA conversion is really very small,
the handbooks show you the details.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

Ok i am afraid that perhaps I can't be any clearer.. wow this is a little frustrating. what i am asking has nothing to do with the latency of the asio driver through the DAW, or the core audio driver! I want to know what the latency is when bypassing the daw, from input to output, using only the dsp mixer! I know what the DAW latencies are because TAFKAT has posted them at gearslutz! This is so far from what I am asking.. please if no one understands what I am asking, just say so, and i will try to explain even further.

7 (edited by TNM 2017-09-18 18:23:00)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

To get the sort of latency i would need monitoring THROUGH the daw, i would need 32 samples at 44K or 64 at 96K, and my 2011 imac can not handle that! No way! I just want an ultra low latency (2ms or under) hardware dsp mixer that i can monitor my outboard gear in in realtime, with a bit of compression or eq here or there, and maybe some monitoring reverb for a vocalist..

When it is time to mixdown i will transfer them all to audio tracks and THEN use native plugins! for monitoring i ONLY want to use RME  DSP plugins!

Ok, here's an easier question..

what is the latency for a vocalist monitoring through total mix FX reverb? Is that more coherent? I am not sure what I am explaining wrong. Again, *please* forget asio or core audio drivers and forget absolutely any latency figure that any daw reports. peace smile

And to be quite honest, i can't even believe RME doesn't say what this figure is.. all they say is near zero latency. Steinberg are the same with their DSP mixer for the UR824. The UR824 has much better fx but you can use less of them.. and even though it has terrible latency through core audio or asio, it is of no relevance to me as i won't ever be *live monitoring* through that pathway..So even the UR824 is an option cause of it's HARDWARE mixer.. the daw latency is of no relevance to me as i use all midi hardware synths...the reason i am gravitating to RME is sound quality and firewire vs USB for the UR824.

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

I understand you but I don't know. It is converter latency plus the one sample MC mentions. I have seen the converter latencies on this site or in the specs.

Vincent, Amsterdam
https://soundcloud.com/thesecretworld
BFpro fs, 2X HDSP9652 ADI-8AE, 2X HDSP9632

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

vinark wrote:

I understand you but I don't know. It is converter latency plus the one sample MC mentions. I have seen the converter latencies on this site or in the specs.

I thought MC meant the FX had 1 sample.. did you take it to mean the entire dsp mixer only had 1 sample?

Still, i'd like an official RTL from rme. I think that's a reasonable request. Cheers and thanks for trying to help. smile

10 (edited by cigame 2017-09-19 07:49:47)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

In my measurements, it's 34 samples at 44.1 kHz

And now, can you tell me exactly how much NOX is getting out of my VW Caddy?

UFX+, FireFace 802 FS, Digiface USB
12 Mic, M1610 pro, Micstasy
MacBookPro M1
Logic Pro X

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

Like mentioned before, the manual lists the exact AD/DA latency. Add them together, put another 1 sample on top of it and you know your latencies.

12

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

That is not fully true. He asks for stand-alone operation (one simple word to understand the question...). In that case some unexpected samples from internal routing FPGA to/from DSP will be added. I can measure this tomorrow.

Regards
Matthias Carstens
RME

13

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

Thank you, as i thought MC.. and THANKS very much smile I just didn't think 34 samples was possible at 44k, i mean that's better than a 20 thousand dollar hdx system at 96k! Looking forward to your reply, but i found a fireface 802 in stock and i am pretty sure i'll get it as long as the RTL through the mixer incl effects is not over 2ms. Cheers

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

What about Durec, could you require this ?

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

15

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

ramses wrote:

What about Durec, could you require this ?


It would be wonderful but the ufx+ is 4 thousand here and the 802 is 2 thousand.. the difference is too dramatic.

And ultimately all i can use are 32 ins in pro tools..

I found a little spdif to aes converter box so i will also be able to use my virus Ti in the aes input so i'll be able to use all 30 inputs of the 802 which is just 2 shy of the maximum pro tools allows without HD avid interfaces anyway.
I am trying to find videos on durec without success.

I searched "rme durec"...

What i am trying to find out is whether it can punch in and out to record so everything is grid aligned, otherwise there'd be no purpose for my style of music anyway..cheers

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

I mean the UFX .. its one of the rare occasions where the former flagship interface is still available where new products hit the market
UFX+ - € 2699
UFX II - € 1936
UFX - €1599
802 - €1394

All three UFX* have Durec. The very little price difference of around €200 between 802 and UFX fully justifies to take the former flagship interface UFX.

With the UFX you get on top of the 802:
- Micstasy Mic Preamps with parallel converter technology (2 converter per channel)
- 65db Mic Gain
- +12dBU max input level
- 2x MIDI
- Durec
- Digital Gain Controls which you can store in snapshots and workspaces
- Autoset
- Fully standalone via Color Display (the 802 needs the old ARC to switch between the 6 standalone profiles)
- Standalone profiles accessable for Display

As you want to work in standalone mode and maybe want to switch between up to 6 setups, the 802 is not recommended IMHO anymore as you really require the old ARC to switch between the standalone profiles. The old ARC is not available in the market anymore.

For a device in this price range I would say the UFX, which costed before the same as the UFX II now, its a present to get the UFX now for only €200 more and it makes IMHO no sense to invest into the 802, if you can get all the listed features for only €200 more.

See my blog article here
http://www.tonstudio-forum.de/blog/inde … B-MADIfac/
And an overview about features of the different RME interfaces:
http://www.tonstudio-forum.de/index.php … 17-01-pdf/

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

17 (edited by TNM 2017-09-20 08:59:05)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

the arc remote control? I will look into it i am sure i'll find a second hand one if i really need it.

However, i will check out australian prices of the UFX then, thank you.

Edit, i just checked it out, you are right it looks like a much better interface...

It also has a nice big mains monitor on front which lacks in the 802..

I am a bit disappointed the UFX inly has FW400, but it's nice it has usb backup for win machines.. my win laptop i sometimes but not often use, is usb 2 and 3 only (2 ports of each). 90% of the time I am on mac.

My FW port is not being used for *anything*, as i said it's a completely free buss as it's not shared with any of the other busses, totally independent of usb and thunderbolt.

So, the question is, can firewire 400 handle 30 channels of audio being inputted simultaneously at 96k?


PS wasn't durec improved in ufx+ etc? is the first revision still useful?

will it work with a fast sandisk usb3 flash drive?

18 (edited by ramses 2017-09-20 09:46:17)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

> I am a bit disappointed the UFX inly has FW400, but it's nice it has usb backup for win machines.

1. USB2 is not only for windows machines. Apple doesnt build machines with Firewire interface anymore.

2. There is absolutely no reason to be dissappointed. 30 channel interfaces do not require more than USB2 and FW400! This has been mentioned very often by RME in the user forum. And also the 802 has "only" USB2 and FW400. RME only mounted a FW800 connector to make it easier to connect to FW800, but actually the FW800 adapter then needs to fallback with speed to FW400, because the 802 is a FW400 interface.

And even the UFX II has only USB2. Why .. Because it also has only 30 channels and its sufficient.
UFX II btw has only USB2, no 2nd interface anymore, because even Apple doesnt build Laptops with firewire anymore
and to get Firewire to run can be challenging at times and also with the new Win10. I read reports about that in this forum.

> So, the question is, can firewire 400 handle 30 channels of audio being inputted simultaneously at 96k?

Yes and yes. BTW with 96k you reduce the amount of ADAT channel because of channel multiplexing.
At the end the "load" stays the same. Either all 30 channels in and 30 channels out @44.1/48
or a reduced numer of total channels because of channel multiplexing at higher sample rates up to 192kHz.
But even then there is no issue with USB2 and FW400.

> PS wasn't durec improved in ufx+ etc? is the first revision still useful?

For years people used Durec with UFX. You simply had to look for a proper Disk or USB stick that can deliver the throughput sustained without interrupt. I have such a 64 GB USB3 stick, that ran perfectly with Durec/USB2. Here is a thread in user forum which details which devices are proven to work reliably with the total amount of Durec channels.

The improvement with UFX II and UFX+ is, that the newer devices have a real time clock, so that you get timestamps on the recorded files. And the compatibility to some drives has been improved. But this is at the end no big issue.

Of course, if you want, you can get the UFX II, if you want to have the latest and greatest.

BUT then you are back to the old "flagship" price in the range of €2000.

All that I tell you is .. if you want to go now for a 802 and are not willed to pay €2000 for UFX II or €2700 for UFX+.
Then you definitively SHOULD look at the UFX as long as it is available in the market.
You cant get any better interface in terms of features and functionality.

When UFX II and UFX+ were not yet available in the market, then it was the flagship interface superior to all other interfaces (802, UCX, etc ...). So .. you do nothing wrong by getting it now for a much reduced price.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

19 (edited by TNM 2017-09-20 10:21:55)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

1) I wasn't saying that.. But i did write it incorrectly.. What i meant was windows has usb more commonly than firewire.

You can always firewire perfectly with apple with the thunderbolt to firewire. it works with RME.

I was talking out loud to myself and thinking of my case, where i have apple native firewire but only usb on windows.

Sorry about that

I'd never use USB on the mac.. i have like 15 synths all going via USB, drives, and so on. The 2 usb busses are absolutely maxxed out. Since RME don't have Tb pass through, i doubt i will ever buy their TB stuff.

I noticed the UFX loses the FW pass through of the 802.

20 (edited by TNM 2017-09-20 10:27:34)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

Ok i read the rest.. durec is useless for me then.. at least i need the timestamps. The way it is now, i would be having to chop every audio file i imported to pro tools to line up perfectly and i hate doing that.

I don't do any live situation where durec might be useful.

But still, If i can find a UFX, of course it makes sense to get it over an 802 with what you have said..and the fact it is only firewire 400 on the 802 after all, well, all good then. If i can not find a reasonable priced ufx in australia within 10% of the 802 price, then i might still get that. It's hardly going to be a bad interface..I do one vocal and one guitar but everything else is synths.. most of them old.. they can only sound as good as their own DAC and the the 802 ADC is way better than any of my synths as it is.

Of course these points are all moot if MC comes back tomorrow and the throughput dsp mixer latency is no better than apollo.. i'd then stick with that. It all comes down to the total mix latency for me, i need it to be as close to analog (0) as possible.

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

How much is the apollo latency anyway?

Vincent, Amsterdam
https://soundcloud.com/thesecretworld
BFpro fs, 2X HDSP9652 ADI-8AE, 2X HDSP9632

22 (edited by ramses 2017-09-20 11:16:12)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

TNM wrote:

Ok i read the rest.. durec is useless for me then.. at least i need the timestamps. The way it is now, i would be having to chop every audio file i imported to pro tools to line up perfectly and i hate doing that.

No, this is not true. You dont know the workflow.

The Durec File is ONE file containing all tracks multiplexed.

So after a recording of lets say 4h of 20 tracks, you have only ONE big file.

This file you can either directly drag and drop to your DAW (Cubase allows for this)
or
You can extract it with RME multi-channel batch processor and then import the single tracks to your DAW.

Of course all the files of one Durec Recording start at the beginning at the same timestamp.

With timestamps I only meant, that on the Durec Medium (Stick, Disk) the big Durec files have no Date and Time.
On UFX you can differentiate the BIG multiplexed files only by its sequence number in the file name.
On UFX II and UFX+ there is a clock inside the device, there the Durec Files have a timestamp.

BUT ... if you do one or more recordings ... you can easily differentiate them by the number in the file and by content when playing back which is also possible.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

23 (edited by TNM 2017-09-20 11:22:06)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

vinark wrote:

How much is the apollo latency anyway?


through console it is 2.23 ms without any FX...

This is where i got fooled and so many others..

at 96K it is 1.1 ms, which is great,. and 90% of UAD fx use the same dsp at 96k that they do at 44k.

This is because most of them are upsampled internally to 192K.

unfortunately, with UA bragging about apollo and using UAD fx with them without any additional latency, they fail to mention unless you look in the manual, that it's only 10% of the uad plugins that fall into that category.

This is why they have an extra feature called input delay compensation, to keep phase coherence amongst all apollo console tracks when any plugins with latency are used. Even on the smallest setting of 100 samples, that already adds another 2.5ms latency at 44K.. and most of the good effects are 55 samples or up, so by using just two (one say, pultec eq and one la2a mk2 comp) you'd already break the 100 sample IDC.. it's ridiculous. For my needs i need the medium setting of 200 samples, and i even break that sometimes.

So, my monitoring chain is anywhere from 7ms to 9 or 10ms! through apollo's console!

The console advantage is it's all DSP and the uad effects are very good. But really, 96K is where it's at with apollo console if you want low latency.. it can be kept under 3ms even with a few good uad plugins  being used.

OR, 44K is ok if you use just the old compressors and EQ's which don't add extra latency. That's the only workaround.

I use it as a live analog mixer right now and it's out of time with my daw project playback.. you can clearly hear those 8 ms.

Presonus quantum is another option, a pro tools guy has measured it and it's 2.5ms RTL in PT at 44k, and he said cpu was very stable.. it's under 2ms in studio 1 (mac - win driver has higher latency).

That's if i was ever to go native..

HOWEVER... the RME is intriguing cause it guarantees eq and compression on every channel.. I know we are not talking uad stuff here but i have read that the latency is very very low.. someone i know from forums said it's under 2ms with comp and EQ engaged, at 44k, which is why i am interested.. I just wanted RME to confirm with real figures.
And this still gives me the case of an analog mixer style but all digital, with no native cpu load. My imac is old, it's almost 6 years old.
If i get the RME i only need the one interface.. i got an incredible freak deal on my apollo8 and twin duo so i can sell them and probably get my money back in australia as i got them from the US, they are very very expensive here, and keep my uad satellite for daw mixing, so i won't lose my FX.. that would leave me a nice bit of change after the RME (put it this way, the apollo 8 is 3699 here on SPECIAL, so the RME almost 2k cheaper), and means i could do everything, 30 inputs with one interface which is just 2 shy of pro tools vanilla 32 maximum anyway. I currently have 28 between apollo 8, apollo twin and 2 adat 8.
I could still use my focusrite adats perfectly with the RME.

it's all about latency for me whilst all my gear is playing live so the timing can be tight.. I often don't need to record stuff to audio because all the synths levels and fx mixing is done on the synths themselves these days, in many cases (like the virus, nord, etc)..But with the apollo i have to record everything to audio so i can have tight timing, as the driver accurately reports it's latency to pro tools depending what IDC setting is, and records to the sample, perfectly on the timeline. But when playing live midi with uad fx, timing stinks. All due to the latency of the uad fx.

I need something with real LOW i/o latency and a dsp mixer is what it boils down to.

NOW IF, the RME total mix latency is no better than apollo, i will get a second apollo 8 instead.. this way i can run 32 channels in, the pro tools max, and use both adats at 96k with 8 channels per apollo, as both the focusrites and the apollo can do 8 channels at 96k each.. so that would be 8 on board ins x 2, and 8 adat ins x2.. and i'd run at 96k and the latency through console would be acceptable.

but i'd have to spend another 3 and a half grand.. if i can spend 2 grand and get 3 and a half grand instead for both my twin and apollo 8, i will not only not need to spent another 3 and a half grand but i'll have 1500 in the bank.

It all boils down to whatever figure MC gives me tomorrow smile

24

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

ramses wrote:
TNM wrote:

Ok i read the rest.. durec is useless for me then.. at least i need the timestamps. The way it is now, i would be having to chop every audio file i imported to pro tools to line up perfectly and i hate doing that.

No, this is not true. You dont know the workflow.

The Durec File is ONE file containing all tracks multiplexed.

So after a recording of lets say 4h of 20 tracks, you have only ONE big file.

This file you can either directly drag and drop to your DAW (Cubase allows for this)
or
You can extract it with RME multi-channel batch processor and then import the single tracks to your DAW.

Of course all the files of one Durec Recording start at the beginning at the same timestamp.

With timestamps I only meant, that on the Durec Medium (Stick, Disk) the big Durec files have no Date and Time.
On UFX you can differentiate the BIG multiplexed files only by its sequence number in the file name.
On UFX II and UFX+ there is a clock inside the device, there the Durec Files have a timestamp.

BUT ... if you do one or more recordings ... you can easily differentiate them by the number in the file and by content when playing back which is also possible.


i read the manual and you are right, i really do not understand durec at all.

What i would need is something that can start recording only when the audio playback starts, perfectly on time, so i could insert right on the grid in my daw.. if it can't do this, i won't use it. From your explanation i still don't understand whether it can. The way i read the manual it says i have to hit the record button manually, so no matter how fast i am, it still won't be sample accurate to the beat when dragged into daw.. if i can do this, let me know! Pro tools is well known to give errors when recording even with fast hard drives, so a backup system is not a bad thing!

25

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

At 48 Khz, analog in to out via TotalMix FX, input and output channel have EQ, LC, Dynamics and AutoLevel active. Rev/Echo do not change these values:

48 kHz: 0.7 ms
96 kHz: 0.33 ms
192 kHz: 0.16 ms

The manual says at 48 kHz AD 13 samples, DA 7 samples. One has to add some samples for the internal routing FPGA to/from DSP plus one sample TM FX, which  results in about 33 samples at 48 kHz. Half and another half at 96 and 192. Please note that at 192 kHz the number of available FX channels is significanty reduced due to limited DSP power.

Regards
Matthias Carstens
RME

26

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

Hi, thanks!

is 44k the same latency as 48K All were done but the main one i asked for lol! But surely it couldn't be much higher!

Which interface is this? The 802 manual shows higher  figures than that for the ad da..

BUT, i can still work it out, cause i can add 33 samples to the result.

it's exactly 1ms at 44k for the 802'w A/D/DA rountrip, so add another 0.75ms to that for the FPGA..

basically it's 1.75ms at 48k but i wonder if it's still under 2ms for the 44K.

Still, even the first gen UFX would be faster than this yes?

As it stands, even the 802 looks like it is no higher than 2ms complete RTL at the 44 sample rate.

This is huge news, and means rme dsp AND the converters are MUCH faster than the UAD.

If the figures you just gave, cause you didn't specify which interface, are valid for the UFX, it would mean the UFX RTL with FX at 48K is only 1.45 MS.. this is as close to analog as we will ever get, and faster than all digital mixers i know of.

If 44K on a UFX first gen (this is as high as i will spend and i want firewire as said) is 1.45 ms RTL, WOW, this is amazing. Vocalist would never complain about monitoring, i can put a bit of dsp verb for him or her, and a compressor, and all my midi synths will play in perfect time with my daw project playback, unlike the apollo dsp mixer.. it means i can stay in the midi domain instead of having to record everything to audio all the time like i do with the apollo. For a bit of fun i might even invest in a little patch bay and an external lexicon 400 as well, still keep me cheaper than a second apollo 8.

the scarlet octopres has word clock out but the octopre mark 2 has only in.

So this is how i do it on apollo and i presume on the RME.. i would use RME as master clock, take bnc out into scarlett octo pre, then out of the scarlett pre into the octopreMK2, then the adats of both into the RME via optical cable. This works perfectly with the apollo as master.

Tell me, does the AES EBU digital input of the RME gear have a "resample" option at input, just for that channel? This is important to use the virus via digital, it causes clicks and pops otherwise. If not, no matter, i am still happy with 28 analog ins and could buy a 2 channel a/d to aes sometime in the future to make use of all 30 ins, if i needed to.

Thanks for your work MC

27

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

Oh and please understand.. the only Vi i use is kontakt, and all my libraries are 44k..
All my audio loops are 44K...
none of my digital synths are higher than 44K.

Only my dave smith true analog could possible benefit from recording at 96K, but to use all that drive space and cpu when there won't be any benefit for anything else, would be a waste.

I tested a virtual aax synth yesterday at 44k, it was 4%, and at 96K at the same latency and midi file on the same imac was 24%. This is a huge difference. I want to keep CPU as low as possible, which is why for mixing i will still use as many uad fx as possible with my octo and quad satellites, and then when out of DSP i would use fabfilter which are incredibly light, as well as a couple sonnox.. Just really well optimised stuff. Currently, even a 96 track project, the way i work, shows at peak only 15% on the pro tools meter, on an imac that basically has a 2600k inside (3.4ghz i7). That's the beauty of DSP and using all hardware synths..

see, changing to RME does not affect quality of my mixes in any way.. if i really need to use a uad effect i will track it to audio, but i'll be now able to keep many of the midi synths live as the latency is so low.

I will lose the ability to "track to tape" like i do with Apollo, but i can live without that! Low latency is more important than anything for me.

You would think, when UAD released the mark 2 apollos, they would have improved ad/da and dsp latency, huh? Or even, why not OFFER non upsampled versions of the newer FX.. that don't add latency? Just give us the option! but they don't.

Only the 2007 effects from uad have no latency, and some plugin alliance ones. that's it.

28 (edited by vinark 2017-09-20 13:30:53)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

TNM wrote:
vinark wrote:

How much is the apollo latency anyway?


through console it is 2.23 ms without any FX...

This is where i got fooled and so many others..

at 96K it is 1.1 ms, which is great,. and 90% of UAD fx use the same dsp at 96k that they do at 44k.

This is because most of them are upsampled internally to 192K.

unfortunately, with UA bragging about apollo and using UAD fx with them without any additional latency, they fail to mention unless you look in the manual, that it's only 10% of the uad plugins that fall into that category.

This is why they have an extra feature called input delay compensation, to keep phase coherence amongst all apollo console tracks when any plugins with latency are used. Even on the smallest setting of 100 samples, that already adds another 2.5ms latency at 44K.. and most of the good effects are 55 samples or up, so by using just two (one say, pultec eq and one la2a mk2 comp) you'd already break the 100 sample IDC.. it's ridiculous. For my needs i need the medium setting of 200 samples, and i even break that sometimes.

So, my monitoring chain is anywhere from 7ms to 9 or 10ms! through apollo's console!

The console advantage is it's all DSP and the uad effects are very good. But really, 96K is where it's at with apollo console if you want low latency.. it can be kept under 3ms even with a few good uad plugins  being used.

OR, 44K is ok if you use just the old compressors and EQ's which don't add extra latency. That's the only workaround.

I use it as a live analog mixer right now and it's out of time with my daw project playback.. you can clearly hear those 8 ms.

Presonus quantum is another option, a pro tools guy has measured it and it's 2.5ms RTL in PT at 44k, and he said cpu was very stable.. it's under 2ms in studio 1 (mac - win driver has higher latency).

That's if i was ever to go native..

HOWEVER... the RME is intriguing cause it guarantees eq and compression on every channel.. I know we are not talking uad stuff here but i have read that the latency is very very low.. someone i know from forums said it's under 2ms with comp and EQ engaged, at 44k, which is why i am interested.. I just wanted RME to confirm with real figures.
And this still gives me the case of an analog mixer style but all digital, with no native cpu load. My imac is old, it's almost 6 years old.
If i get the RME i only need the one interface.. i got an incredible freak deal on my apollo8 and twin duo so i can sell them and probably get my money back in australia as i got them from the US, they are very very expensive here, and keep my uad satellite for daw mixing, so i won't lose my FX.. that would leave me a nice bit of change after the RME (put it this way, the apollo 8 is 3699 here on SPECIAL, so the RME almost 2k cheaper), and means i could do everything, 30 inputs with one interface which is just 2 shy of pro tools vanilla 32 maximum anyway. I currently have 28 between apollo 8, apollo twin and 2 adat 8.
I could still use my focusrite adats perfectly with the RME.

it's all about latency for me whilst all my gear is playing live so the timing can be tight.. I often don't need to record stuff to audio because all the synths levels and fx mixing is done on the synths themselves these days, in many cases (like the virus, nord, etc)..But with the apollo i have to record everything to audio so i can have tight timing, as the driver accurately reports it's latency to pro tools depending what IDC setting is, and records to the sample, perfectly on the timeline. But when playing live midi with uad fx, timing stinks. All due to the latency of the uad fx.

I need something with real LOW i/o latency and a dsp mixer is what it boils down to.

NOW IF, the RME total mix latency is no better than apollo, i will get a second apollo 8 instead.. this way i can run 32 channels in, the pro tools max, and use both adats at 96k with 8 channels per apollo, as both the focusrites and the apollo can do 8 channels at 96k each.. so that would be 8 on board ins x 2, and 8 adat ins x2.. and i'd run at 96k and the latency through console would be acceptable.

but i'd have to spend another 3 and a half grand.. if i can spend 2 grand and get 3 and a half grand instead for both my twin and apollo 8, i will not only not need to spent another 3 and a half grand but i'll have 1500 in the bank.

It all boils down to whatever figure MC gives me tomorrow smile

Thanks! I agree 9ms is much to high. And wow those are great figures for TM FX! On top of this, if you ever play your Kontakt libraries live, you will notice an improved low latency experience with RME. In my humble experiences it is night and day, usable 64 samples buffer under heavy load!
And as a side note...I am still on UAD1's

Vincent, Amsterdam
https://soundcloud.com/thesecretworld
BFpro fs, 2X HDSP9652 ADI-8AE, 2X HDSP9632

29

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

the output latency is all that matters when playing virtual instruments, and that's very very low on the apollo thunderbolt.. THAT, i can't complain about, not even a bit. it's super responsive. at 128 samples it's like 4ms and at 64 it's like under 3 from memory..nothing...

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

TNM wrote:
ramses wrote:
TNM wrote:

Ok i read the rest.. durec is useless for me then.. at least i need the timestamps. The way it is now, i would be having to chop every audio file i imported to pro tools to line up perfectly and i hate doing that.

No, this is not true. You dont know the workflow.

The Durec File is ONE file containing all tracks multiplexed.

So after a recording of lets say 4h of 20 tracks, you have only ONE big file.

This file you can either directly drag and drop to your DAW (Cubase allows for this)
or
You can extract it with RME multi-channel batch processor and then import the single tracks to your DAW.

Of course all the files of one Durec Recording start at the beginning at the same timestamp.

With timestamps I only meant, that on the Durec Medium (Stick, Disk) the big Durec files have no Date and Time.
On UFX you can differentiate the BIG multiplexed files only by its sequence number in the file name.
On UFX II and UFX+ there is a clock inside the device, there the Durec Files have a timestamp.

BUT ... if you do one or more recordings ... you can easily differentiate them by the number in the file and by content when playing back which is also possible.


i read the manual and you are right, i really do not understand durec at all.

What i would need is something that can start recording only when the audio playback starts, perfectly on time, so i could insert right on the grid in my daw.. if it can't do this, i won't use it. From your explanation i still don't understand whether it can. The way i read the manual it says i have to hit the record button manually, so no matter how fast i am, it still won't be sample accurate to the beat when dragged into daw.. if i can do this, let me know! Pro tools is well known to give errors when recording even with fast hard drives, so a backup system is not a bad thing!

Simply also record your audio playback as well by using loopback then you have all tracks together in the Durec file.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

31

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

but can durec start precisely when audio playback starts?

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

TNM wrote:

but can durec start precisely when audio playback starts?

Does it really matter ?

If you have all relevant tracks in the Durec recording (which I suggested, to record also the playback track)

then you can import this to your DAW

And there you can use cut the recording

More important is that all the tracks have the same time base which is the case by recording them all.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

33

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

yes it matters, i've been repeating that for multiple messages.. if i can't use durec to fall exactly on the midi grid, it's not for me.

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

TNM wrote:

yes it matters, i've been repeating that for multiple messages.. if i can't use durec to fall exactly on the midi grid, it's not for me.

Maybe RME can comment to this.

As Durec recording, stop, playback can be controlled by the ARC USB I would guess that its either possible or maybe not so tricky to implement this via MIDI.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

Why use durec at all in your (and mine) case at all? It adds a layer or 2 that add complexity and room for error. It is very useful for live recordings (also in the studio as safety backup with live takes) though.

Vincent, Amsterdam
https://soundcloud.com/thesecretworld
BFpro fs, 2X HDSP9652 ADI-8AE, 2X HDSP9632

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

Simply forget Durec for the time being then .. its a nice feature, but maybe not for you now.

Nevertheless the UFX is compared to the 802 advanced.
Its, as I said, a nice thing to be able to get it now for only a few bucks more.
With or without Durec ... doesnt matter.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

37

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

TNM wrote:

is 44k the same latency as 48K

I wrote 33 samples. 1/44100 x 33 = 0.75 ms. Also when using stand-alone you can use whatever sample rate you like.

TNM wrote:

The 802 manual shows higher  figures than that for the ad da..

Not my manual. UFX+ is the same as 802. The old UFX has higher values as its DA has more samples in latency (28 instead of 7).

Regards
Matthias Carstens
RME

38 (edited by TNM 2017-09-20 18:50:43)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

no u wrote 0.7ms + 33 samples.. it's the initial figure im asking about in 44k

so 802 has more up to date converters than old ufx?

i'll read the online manual again, cheers, must have read it wrong

39

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

vinark wrote:

Why use durec at all in your (and mine) case at all? It adds a layer or 2 that add complexity and room for error. It is very useful for live recordings (also in the studio as safety backup with live takes) though.

agreed

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

You can find it also in my PDF under "Wandler Verzögerungen in samples AD/DA"

http://www.tonstudio-forum.de/index.php … 017-01-pdf

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

41 (edited by TNM 2017-09-21 17:38:19)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

ok so the UFX is every so slightly faster than the apollo RTL.. negligible. Pointless in my case... Better to stay with what i have in this case for similar latency, and use the uad plugins which do not have any additional latency like the old pultec/1176/la3a/cambridge etc.. which are more exciting than the totalmix fx anyway...

the only figures that make sense for a latency change and switch are the 802 or ufx+, and latter too expensive, so 802 it is, as thread began. smile

EDIT, i was totally wrong, ufx is faster than apollo, almost 50%, the ufx+ is 3x faster (monitoring latency through dsp).

42 (edited by ramses 2017-09-20 19:07:19)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

You forgot the UFX II ...
Its the UFX+ without
- MADI
- USB3, Thunderbolt

Otherwise the same and USB2 only because of the 30 channels.

The UFXII / UFX+ got many enhancements compared to the UFX.

You can read them here:
http://www.tonstudio-forum.de/blog/inde … 8-RME-UFX/

Should not be too hard to read the german and translate it into english, I substracted here the UFX+ bits and pieces:

- Überarbeitung der analogen Schaltkreise und des DSPs
- verbesserte SNR und THD Werte
- Direct USB Recording (DuRec): bis zu 76 Kanäle können gleichzeitig aufgenommen werden (ie 12 analog, 64 Madi)
- Interne Real Time Clock, ermöglicht Zeitstempel in Dateinamen für DuRec
- Erhöhte Zuverlässigkeit bei Verwendung langsamerer oder "multi partitioned" USB thumb devices
- Mikrofon Eingänge mit +18dBU max, kein PAD
- Mikrofon Eingänge haben 75 (!) dB Gain Range, das sind 10dB mehr im Vergleich zu UCX/UFX und 15dB mehr im Vergleich zum 802.
- Kopfhörer Ausgänge mit nur 2 Ohm Impedanz, mit +19 dBU max output Level
- Neue Remote Control "ARC USB" (Details siehe unten) für regulären und stand-alone Betrieb
- Die beiden Drehgeber rechts vom Display regeln nun individuell die Lautstärke von Phones 9/10 und 11/12.
- Fully stand-alone. Es gibt im Vergleich zum UFX zusätzliche Einstellungsmöglichkeiten.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

43

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

nope thats 3200$ here, not an option. also usb only.

802 is perfect in my case.. 1.5ms RTL at 44k, 0.7 at 97K.. 44k figure is faster than avid hdx (1.9ms).

i do not need durec and the 802 has all the features i need plus fw through. smile

what a shame the ufx is 2ms RTL and not worth it vs apollo, as i found a store with 3 in stock same price as 802..

if it was 1.5 like 802 i'd buy it right now

44 (edited by ramses 2017-09-20 19:37:27)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

Suggestions, ask your dealer. Get the two (802, UFX II), decide then. USB only is not a real reason to say no to an UFX II.

BR Ramses - UFX III, 12Mic, XTC, ADI-2 Pro FS R BE, RayDAT, X10SRi-F, E5-1680v4, Win10Pro22H2, Cub13

45

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

TNM wrote:

no u wrote 0.7ms + 33 samples.

No I did not. But no problem, I give up.

Regards
Matthias Carstens
RME

46

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

don't give up, i misunderstood.

so it's 0.7 TOTAL rtl at 48K?

wow that's incredible which means even the ufx is only 1ms total not 2 like i thought.

ufx it is then for extra features.

47 (edited by TNM 2017-09-21 17:39:03)

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

edit, not needed

48

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

ramses wrote:

Suggestions, ask your dealer. Get the two (802, UFX II), decide then. USB only is not a real reason to say no to an UFX II.

huh? i have checked the best australian dealers by phone since 9am this morning.. ufx 2 is at least 1 grand more than ufx or 802/

i said i want to have some money in the bank.

btw, found the 802 remote in stock for 200

regardless, the ufx is a better buy as you said.. i misunderstood the latency figures.

To have that much nicer screen and extra midi is a good thing.. also the 802 misses the monitoring knob as i said as well as the detailed level meters.. just noticed that.

UFX is a 7yo interface though.. i might try hear them side by side to my apollo.. the apollo d/a seems to be better specced, although RME measure unweighted, apollo is a weighted.. i need to do some research on this..

usb is no good in my case, as i said both busses are already struggling to cope with the bandwidth i have asked of them. Adding 30 inputs of audio in realtime is impossible . I know this, because i tried a babyface pro of my friends to listen to the rme compressor, and it couldn't cope with 12 live inputs as usb bandwidth was exceeded. All i got was clicks and pops.. I have already 2 x 10 hubs on two of the 4 usb 2 outputs (only 2 busses actually though).. and on each of the other usb ports has a motif streaming audio and data and a virus streaming 6 channels of audio and midi.. on the hubs i have drives being maxxed out, dongles, and another 8 midi devices, as well as my 5x5 motu usb midi. Forget it.

I'd need to buy a thunderbolt hub with a pass through, to access another usb buss.

The unused firewire bus is working perfectly, i tried a motu interface on it and it was great.. rme can only be better.

Finally, the new problem i may have discovered is that the combo jacks on the front of the ufx and 802 are XLR or unbalanced TS only for the line ins, so i presume for guitars or microphones.

Does this mean my synths won't work in those inputs?

anyway, the total recall of the pre levels in UFX, the bonus that i *may* use some day of durec, the screen and meters and extra midi port means that will be the one i get, and make do with the extra 0.5ms output latency.. it's still only 1.22 total RTL at 44.1K which is incredible and almost half the apollo at 44k using no effects.. but the rme measurement is through the eq, comp and i can use verb and delay as well. On apollo one must add yet ANOTHER 32 samples to use the aux buss.. the devil is all in the very hidden details. Uad console latency sucks.

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

I gotta say.... these posts are as good as watching " Breaking Bad ". While I actually find myself understanding the information, I'd never be able to do it but...I am anxiously awaiting the next posting

Ramses:  what a great job detailing the benefits of the UFX !! You and Vinark never fail to get my attention

Babyface Pro, UFX+ via Thunderbolt, Win 10, Cubase 9.5 Pro, Asus Z270 i7700k Guitarist-1961

50

Re: Trying to find out TotalMix Latency (not daw monitoring) 44K, 802, UFX

Hmm, you paid your monthly subscription fee already? smile

Regards
Matthias Carstens
RME