Help Please - RME Fireface400 Live Mode

[Stefan]

I've just received my new RME Fireface 400 which I have connected to my laptop (running XP) with SS in Live Mode but am getting awful problems.

I've read the Fireface manual (not an easy undertaking) and followed general instructions given here regarding buffers in Live Mode but I guess I am doing something wrong at a pretty basic level.

I started out by setting up SS Audio Driver Model = Standard Multimedia, Audio Device In1/Out1 = Fireface 400 Analog 1+2, then tried the number of SS In and Out buffers at 2,3 and 4 respectively at a size of 64 samples with RME Fireface Settings, Buffer Size = 64 samples, Limit Bandwidth = Analog, Phantom Power = Mic1 (everything else is at the FF400 default settings).

With a mic connected to the FF400 the output sound was either delayed by about 4 seconds and totally distorted (like a robot) - barely recognizable as the original voice, and then stopped after about 2 words, or didn't play back at all. This is the same as when I tried Live Mode with my old Soundblaster PCI card - but I assumed that the Soundblaster couldn't handle things and so bought the FF400. I guess because the two problems are identical that I am doing the same thing wrong in both cases?

Anyway, as an experiment I changed the number of SS buffers up to 8 and their size to 2048, leaving FF400 buffer size at 64, and the voice came out clear and undistorted (but delayed by about half a second - not surprising I guess with such huge buffers). I then gradually reduced the size and number of buffers with the resulting latency reducing slightly. But at around 4 buffers of 512 samples the sound started to distort again. I could not get both acceptable latency and a clear voice.

Can anyone please help me out here?

Thanks in anticipation,

Stefan

[Bob L]

Are you on XP SP2? If so, you will probably need to read the Firewire Fix info on the RME site or search the threads here for Fiewire Fix.

Your machine may only be delivering 100 speed until you do the fix.

Otherwise... there were many discussions also about video cards and PCVI latency issues that could easily stop the soundcard from operating at low buffer settings needed for live mode.

If you have a dual video monitor setup, Clone mode has been known to cause problems with many video cards. Try setting the video card to only one monitor for a moment and see if things get better.

Beyond that... system tweaks become important when trying to acheive low latency... read my Windows XP Tweaks doc in the Misc Downloads section of my site.

You may also want to look into one of the PCI Latency utility programs mentioned in the forum here (Search for PCI Latency)... by setting the PCI priorities with one of those utilities you may be able to stop other cards from stomping on the soundcard.

Bob L

[Naturally Digital]

Hi Stefan,

Start by trying the ASIO driver model. You may also want to try the hi-perf MME but I'm guessing the ASIO will work in this case. It's possible the MME driver isn't even supported on this unit. I'm not sure which driver other Fireface users are successful with.

Are you running any service packs on your XP install?

[DominicPerry]

Stefan,

We can help, but it will take some time. Don't expect it to be fixed tonight! Unless we strike lucky.

There are a bunch of things which you should check in Bob's XP tweaks document, which can be found on the sawstudio.com website.

On buffers. You need to be sure the control panel/tray icon for the fireface has the same buffer settings as you are using in SAW. Try 128 or 256 samples first. Try 4 buffers first too. Try a conservative sampling rate too. 24bit /48KHz perhaps.
There are some potential problems with certain versions of Windows and Firewire drivers. There are fixes, but it needs to be checked - details on the RME website. You may also like to try the ASIO drivers - I don't know what state the MME drivers are in for the Fireface400 - anyone help here with Fireface800 driver versions? I can never remember whether to set High Performance Multimedia or 24bit WDM compatible with High performance but you can have more than one item ticked at once.

You may need to shut down some of the internal windows crap - disable services and the like. Leave this 'til later, but bear in mind that anti-virus and other memory resident programs can screw audio.

Don't panic. It will be possible to improve things substantially.

Dominic

[Bob L]

Actually David's idea of the ASIO drivers is a good one... I am so used to the older RME Hammerfall cards with their excellent MME drivers, I never think about ASIO... it may just be that the newer firewire cards and nerwer drivers have already switched into the wdm model which will most likely never hold low latencies.

You may find the ASIO drivers save the day for you.

Bob L

[Stefan]

Thanks again, ASIO mode seems to be my saviour and works OK. No latency that I can notice in Live Mode and only the very occasional glitch (I guess I need to fiddle with the number of buffers and sample sizes - but I'll do that later).

But now I'm trying to have two mics in via the two inputs on the front of the Fireface 400, each going into their own channel in SS each with their own EQ and reverb. I've spent all day on this without success. I've read manuals and searched forums but am failing to understand their content as to just what it is I have to do to "route" mic1 to track1 and mic2 to track2. Both mics always end up on the same track. BTW, I can only get any output via Dev01. Dev02,03,04 produce nothing at all.

Please can someone put me out of my misery and point me in the right direction - again. I can't afford to loose much more hair!

Thanks,
Stefan

[DominicPerry]

Thanks again, ASIO mode seems to be my saviour and works OK. No latency that I can notice in Live Mode and only the very occasional glitch (I guess I need to fiddle with the number of buffers and sample sizes - but I'll do that later).

But now I'm trying to have two mics in via the two inputs on the front of the Fireface 400, each going into their own channel in SS each with their own EQ and reverb. I've spent all day on this without success. I've read manuals and searched forums but am failing to understand their content as to just what it is I have to do to "route" mic1 to track1 and mic2 to track2. Both mics always end up on the same track. BTW, I can only get any output via Dev01. Dev02,03,04 produce nothing at all.

Please can someone put me out of my misery and point me in the right direction - again. I can't afford to loose much more hair!

Thanks,
Stefan

Here's what I do so that I can record and monitor.

For channels I want to monitor AND record, I change the Input Source from Multitrack to Device. You have to select what device and what channel. Then you have to click the MONO button and click on the word 'mono'. This brings up a list and choose 'L only' for channel 1. Then you can pick the SAME device for the second channel Input Source, choose mono again and then choose 'R only'. For channels you want to record, but not monitor, leave the Input Source as MultiTrk.
Now click on the REC button on the MultiTrack window and up pops the record meters. (You may be asked to save a new EDL first - this only happens once). Click on the input window of the Rec meter and you get a list. Choose Mixer channel for the channels you want to monitor AND record (the first example, above) and choose Mono Devices for those that you ONLY want to record. Each Record button you click will bring up a new record meter. Pick and choose as you go along, or after. Sometimes you have to choose something you don't want to free up the channel for another assignment - you'll see what I mean.
NOW, you can click REC RDY on the little transport window, and your monitors will spring to life and you will get massive feedback. At least, I always do.

It's not intuitive until you understand it. Then it becomes plain and clear.

One last thing. If you want the EQ and effects to be 'heard' but not recorded, you have to change a setting in the Multitrack. In the 'Mixer' menu, choose Pre-FX signal flow, and tick those channels which you are monitoring and recording where you don't want the EQ etc to be recorded. I normally do this - the sound I want from the PA isn't the same as what I want recorded so I EQ the recorded version differently later. You can only access the menus in STOP mode. Once you are in REC RDY mode and you have set all your EQs etc, you can just press the REC button and it will sound the same but record all the tracks. When you stop it, the multitrack will populate automatically, but you won't see ity whilst you record unless you change another setting, and I can't remember where it is. - oh, MultiTrack menu - Record-Mark Display - it draws a pretty red bar against all the tracks you are recording, for reassurance.


It will all become clear. Honest. Glad ASIO works OK.

Dominic

[Bob L]

Set things up once and then save a mix template... include a record template in the mix template and you can start fresh... open a mix template for all channel setups and even named tracks and eq settings and even reverb assignments.... one more ctrl-click on a track's record button and you can popup 48 channels of pre-assigned record meters... and you are ready to record. :)

Bob L

[DominicPerry]

You always have an easier way, Bob.:D

Dominic

[Stefan]

After a further 20 hours of fiddling (no exaggeration) I think I have my live set-up, err, set up! Up to 4 mics in with 4 live tracks in SS each with Freeverb etc - I guess I need to buy the Studio Reverb now.

Thanks to everyone for their help.

I finally settled on ASIO drivers, with 2 buffers at 256 samples and have not noticed any glitches for some time. I also applied all Bob's XP tweaks I could. I did have a major problem that stumped me for ages - as soon as I engaged Live Mode I got an awful lot of hiss. The type of hiss would change along with buffer settings, but could not be removed. I re-read the posts here and saw Dominic's comment on trying modest sample rates of say 24bit/48KHz and realised that I had not made any setting here - it was running in the default 16 bit/44Khz. After lots of experimenting I found that 24-bit meant silence and 16-bit meant noise in Live Mode. I guess I don't need to know why, just that it works, but if anyone can shine any light on this I'd be interested. Thanks.

Also, after many many hours of head scratching and manual reading I think I have a handle on TotalMix and have managed to set up a few live scenarios - one for when there are just 2 of us, one for 4, etc. I have to say that this is a pretty nifty piece of software. I've also followed Bob's suggestion and saved some Record and Mix templates - very handy for going straight into Live Mode without having to configure everything each time.

Looking thro' the Fireface manual it mentions having ASIO, GSIF, WDM and MME and that devices without MME at the end of their name are WDM. I only have WDM, no MME listed anywhere. Does this mean that I don’t have the MME drivers installed properly – which might explain why I can’t get any success outside of ASIO? Also, when I look at the hardware device setup for the Fireface, it has 2 entries under Audio, 1 "working properly" the other "not installed properly". I uninstalled and reinstalled drivers but with the same result!?

All suggestions gratefully received.

Anyway, I’m off to admire my new virtual mixing desk.:)

[Bob L]

Should be no noise at 44.1k 16 bits compared to 24 bits... something is not right there... I have done most of my projects at 16 bits with the smoothness of analog. :)

Bob L

[Stefan]

After a little more looking around I discovered a known problem mentioned on the RME site regarding the maximum number of entries for devices in XP. I had to edit the Registry to remove duplicated entries to leave room for the new Fireface drivers. All OK now, with MME devices appearing in SS, also properties dialog states that driver is correctly installed and functioning properly. I now have the option to use MME as well as ASIO. MME is giving fine performance in Live Mode with 3x64 IN and 4x64 OUT.

I notice from the MT Load display in SS, in Live Mode with 4 tracks open, a reading of around 6% when using ASIO and around 8% with MME. They both sound the same to me. I guess I should use ASIO as it is somehow taking less load and offering a little more in the way of a safety margin? Anyone have a thought on this?

Also, I tried LiveMode with 16-bit and still had more noise than with 24-bit. Bob reckons something is wrong here. Any ideas?

Anyhow, it appears that all of my teething problems have been sorted. I have the option of either ASIO or MME, have a great little live setup with just a laptop and the Fireface without the need for mixing desk, reverb units and all the spaghetti cables that go with it. Thanks.

Now I need to decide what reverb to use/invest in for live vocal work - I need something that turns a sows ear into a silk purse. Having been so impressed with SS I reckon Saw Studio Reverb has to be at the top of my list - so I'll go and test out the demo version.:)

Thanks,
Stefan

[Bob L]

If you try out my reverb... start out with the Gymnasium HD chamber... then adjust the delay from there... it is a good overall chamber to get you going.

Bob L

[DominicPerry]

Stefan,

JMS Freeverb is FREE and works very nicely and has a bypass button which is sadly missing on the SAW Reverb. Freeverb also uses about 0.5% cpu so you can have it pasted in as many channels as you like. I also have the SAW Reverb and use it on a return channel, as it uses a bit more CPU and I like to be able to bypass, so I use the AUX send on/off. There are other SAW Native reverbs - Anwida springs to mind. Links to 3rd party Native plugs are on teh sawstudio website. The advantage of the native plug ins is that they are very solid compared to VST plugs (almost never crash) and they are usually fully automatable.
The best sounding verb from MY perspective is SIR, which is free, but is a VST. I've yet to try it in Live Mode, but will this week. It introduces some latency so it may not work live. Others here use Freeverb and SAW verb live.

Dominic

[Ian Alexander]

The best sounding verb from MY perspective is SIR, which is free, but is a VST. I've yet to try it in Live Mode, but will this week. It introduces some latency so it may not work live. Others here use Freeverb and SAW verb live.

Dominic
Hm, never done this, but unless the latency is huge, can't you just shorten the delay a bit to compensate? Reverb is supposed to come after the dry track, right?:cool:

[DominicPerry]

Hm, never done this, but unless the latency is huge, can't you just shorten the delay a bit to compensate? Reverb is supposed to come after the dry track, right?:cool:

Depends how long your pre-delay is. Theoretically the reverb could start within a few ms from early reflections. All that aside, big latency VSTs give the SAW engine a hard time, if I remember rightly.

Dominic