RME introduces the HDSPe AIO Pro PCI Express interface card - Page 6 - Gearspace.com
The No.1 Website for Pro Audio
RME introduces the HDSPe AIO Pro PCI Express interface card
Old 15th March 2021 | Show parent
  #151
Gear Maniac
 
DogCake's Avatar
 
🎧 10 years
I guess you may find better support from users or moderator in the RME forum ...
Old 15th March 2021 | Show parent
  #152
Gear Addict
 
🎧 5 years
Quote:
Originally Posted by pablo1980 ➡️
Testing with only the onboard graphics, nothing improved, I don´t know what else to do
Did you put the AIO Pro in the graphics card slot?
Old 15th March 2021 | Show parent
  #153
Gear Maniac
 
🎧 10 years
Quote:
Originally Posted by soupiraille ➡️
Did you put the AIO Pro in the graphics card slot?
Just tried it to be sure...the same,now even at 4096, it cannot play the projects that were working fine with motu usb
Old 15th March 2021 | Show parent
  #154
Lives for gear
 
🎧 10 years
Quote:
Originally Posted by soupiraille ➡️
Great, thanks Michal. Just out of curiosity, where did you get those streaming mode millisecond values? I did a rough and quick search and did not find those detailed numbers you give. Thank you!
I had those values somewhere in my notes. Many years ago I've did something with XMOS MCUs and their class compliant firmware for audio. Thesycon driver was used for Windows and I've had documentation. I've seen even some driver build by other USB interface vendor, which had this ms values in pull down menu instead of those names.

Michal
Old 16th March 2021 | Show parent
  #155
Lives for gear
 
🎧 10 years
Also I'd maybe check some software things.. unless you've already tried that, I'd test MMMCS option for RME ASIO driver (at global tab in control panel), try it both enabled and disabled, maybe bit will behave better.. IME its effects depends on particular DAW software. Also after you toggle that option, I'd restart DAW between attempts.

Finally although I mentioned, I'm no expert in Cubase/Nuendo and definitely didn't follow all software changes among its recent versions, I'd try maybe couple of things there. For start comparison between ASIO guard on and off. Generally ASIO guard should help with and give you better overall performance (eg. more processing headroom), there can be situations with certain plugins/instruments and workloads, where it might be worth of test without that turned on.

Also I'd look at utilization distribution among CPU cores (normally in Windows task manager with enabled per-core performance graph), when running your heavy project. Maybe there's some really uneven situation, where one core will be completely choked. As touched before, the critical thread in a DAW is usually the one, which processes master bus and send data to/from ASIO driver.
If such thread is scheduled on that choked, overloaded core, then you'll get dropouts.
As MOTU driver has that huge safety buffer, it might overcome that kind of issues, which will be showstopper at RME with 32 samples of safety buffer.
In that case, I'd try to relief some strain in project.. like with bypass of some heavy plugin stuff on master bus (some Ozone, Acustica etc. are most common candidates), if you use it.

Michal
Old 16th March 2021 | Show parent
  #156
Lives for gear
 
🎧 10 years
Quote:
Originally Posted by pablo1980 ➡️
Thanks Michal, but the problem is, I can get it to play without zero dropouts at 4096. 2048 has lots of them. It cannot be working correctly if the motu could handle it fine at 1024.

I cannot accept that this is the intended behaviour of a card that is a standard reference for the kind of work I do (I have seen other composers projects and they are a lot bigger than mine, with similar computer specs).
I see what's your practical issue. And yes you're true that many of those composer rigs employs RME interfaces. It will be just wild guess, projects can be a bit different, so it doesn't have so much load at main thread responsible for master bus and streaming to ASIO driver. They might use shorter buffers in general, as I've mentioned. They didn't compare the same projects at interface with full double buffering at driver.

However only technically related difference, I could describe, between MOTU and RME is that mentioned safety buffer length. 1024 (or what's your working buffer) vs fixed 32 samples is quite huge difference and yes that itself can "iron out" disturbances during audio streaming to hardware, when there's some system disturbance due to scheduling of other processes, drivers in DPC queue or when system is really loaded.
Further increase of working ASIO buffer length (eg. to 2k) in DAW can relief some processing load there (although usually just to up to certain point from which any further increase won't bring any meaningful improvement). However that isn't same thing like increase size of additional hidden buffer in driver, which you can't change.

Quote:
Something surely must be wrong, something I missed in my config?
According to previous posts and your attempts for a remedy, at first spot I haven't seen any obvious flaws. You've checked your DPC latency, assuming you have all usual stuff regarding Windows power plan etc. No wireless, no antivirus etc.

As you've touched slots and PCIe lanes. Yes your CPU has 16 direct lanes, but it shouldn't be necessarily limiting for your setup. All of additional peripherals, SATA, network, NVMe, USB ports etc. aren't included there as those are connected off the PCH, similarly like short PCIe x1 slots at most boards.
I've successfully used RME PCIe cards in those short slots going from PCH and haven't really experienced any issue nor any repeatable measurable performance difference (latency, efficiency) compared to direct CPU slots.
And you've already tried the card in x8 slot, in that case it used direct CPU lanes (graphics then ran with reduced 8 lanes).
Furthermore even if you decide to leave it in such setup and use graphic card running in x16 slot with 8 lanes, I don't think, you necessarily need go to purchase new computer. In vast majority of cases with PCIe 3.0, you won't notice any difference in graphic card performance when running with 8 lanes.
Lower models of graphic cards (like AMD up to RX 570) even runs with only 8 lanes.
I'd be concerned about move to higher HEDT platform with more direct PCIe lanes only in case of some other internal peripherals, like buch of NVME SSDs at expansion cards or so.
So personally I don't think, the culprit will be there or in hardware generally.

I can recall one additional hardware tweak, which helped me with specific issue and glitching at some setups was to completely disable C-States at CPU (you do that typically in BIOS). But it is rather rare for me to do, it usually has some power consumption and temperature implications. So if you would like to try it, watch your CPU temperatures.
Also the issues were prevalent with shortest working buffers, the main point there was best and smoothest performance with lowest latencies. So it might not be applicable to your woes.

Michal
Old 16th March 2021 | Show parent
  #157
Gear Addict
 
🎧 5 years
Quote:
Originally Posted by pablo1980 ➡️
Just tried it to be sure...the same,now even at 4096, it cannot play the projects that were working fine with motu usb
Ok then your issue might not be due to a lack of PCIe lanes. Dropouts like I said can have so many, so many causes, you'll have to be meticulous as to discover the exact cause.
Btw RME is doing live support on Youtube every Wednesday, get your chance with it, even though I wouldn't bet all my money that a solution will arise live.

Otherwise yep, just browse the whole internet and make your own diagnostic. Reach for RME's support. That's a long and lonely journey but you have to know what you want!
All I might say is that your Motu shouldn't have such an advantage over your AIO Pro just because of the double buffering, the cause I think is elsewhere.

I just remember that Pete Brown from Microsoft did present a method which he described as the proper way to record and analyze DPC latency, with Microsoft tools and not LatencyMon that everyone is using. That could be a starting point. I just can't remember in which thread it was but it was here on GS.

I haven't had your specific issue, and never owned a PCIe audio interface, so I might not give you more specific directions other than those I gave you so far; but what I can do is wish you best of luck, and most importantly: REPORT BACK THE SOLUTION IF YOU SOLVE YOUR ISSUE! Jut don't give up
Old 16th March 2021 | Show parent
  #158
Gear Maniac
 
🎧 10 years
I won’t give up yet!!

Michal and Soupiraille I will read very carefully all that you have written tomorrow morning and make more tests, I think something is off, and I hope I can find it.

I have already written to support and a couple of RME forum members are helping me too.

I cannot express how much I thank the time you are taking to help me. Hope I can repay it foward
Old 17th March 2021
  #159
Gear Maniac
 
🎧 10 years
small update, disabling hyperthreading has helped somewhat. I am testing now with keeping the cores at 4.9 (already disabled all c states and parking with park control).

My cores micro fluctuate between 4.898 and 4.901 (aprox numbers), is that ok? or should be locked to 4.900?

I have been updating drivers, BIOS, and will make another pass at latency mon today
Old 19th March 2021
  #160
Gear Head
 
Guys, it's just a statement: These converters are incredible! I have never heard anything more accurate and deep and open and just beautiful than these converters. Great job RME! 🥳
Old 22nd March 2021 | Show parent
  #161
Gear Nut
 
🎧 15 years
Quote:
Originally Posted by Jansemann4433 ➡️
Guys, it's just a statement: These converters are incredible! I have never heard anything more accurate and deep and open and just beautiful than these converters. Great job RME! 🥳
Nice..have you compared them to something like the ADI-2/ADI-8 ? Anyway I'm super happy with my RayDat/BF/ADI-2 combo for now...till I start getting paid :D
Old 26th March 2021
  #162
Gear Head
 
After several hours of working in a session I'm getting micro dropouts. After session restart it's working fine again (for several hours). What could it be?
Old 4 weeks ago | Show parent
  #163
Gear Nut
 
🎧 15 years
Quote:
Originally Posted by soupiraille ➡️
Beware folks: Low Latency Performance is not High Buffer Size Performance! RME performs well in the first case, but as soon as you set the maximum buffer size in the control panel, RME will always perform worse than the double-buffered audio interfaces.

RME is better for tracking and playing live (even though the Quantum has better latencies), but not for more or less "intensive" mixing situations where the buffer size is cranked up.
This is not really true dude, you're overly exaggerating an issue that's usually not there. Once you go up to 2048 or 4096 you're at a point where the latency is so high, that you'll wanna freeze tracks that start becoming an issue anyway, as opposed to having double or more latency added on top of that already huge latency.

RME IS PERFECTLY FINE AND CAPABLE AT LOW LATENCIES AS WELL AS HIGH LATENCY MIXES ON PAR OR BETTER THAN ANY OTHER BRAND OUT THERE DUE TO THEIR AMAZINGLY STABLE DRIVERS AND CLOCKING.
Old 4 weeks ago
  #164
Gear Addict
 
🎧 5 years
Well, common sense and experience just prove the opposite of what you say regarding high buffer sizes and RME+high buffer sizes. However loud you shout it out.
Anyways..
Old 4 weeks ago
  #165
Gear Maniac
 
🎧 10 years
I am waiting for a break at work to reinstall windows, if that doesn’t help I must say that soupiraille seems to be right. Now matter the buffer size this card performs worse than the usb one (in fact it is now back in the box it came in)
Old 3 weeks ago | Show parent
  #166
Gear Nut
 
🎧 15 years
Quote:
Originally Posted by soupiraille ➡️
Well, common sense and experience just prove the opposite of what you say regarding high buffer sizes and RME+high buffer sizes. However loud you shout it out.
Anyways..
Prove it then, wise a..
📝 Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 4615 views: 609333
Avatar for kire
kire 6 hours ago
replies: 97 views: 36309
Avatar for mowmow
mowmow 15th September 2010
replies: 1016 views: 357911
Avatar for QueenSisi
QueenSisi 20th August 2018
replies: 14162 views: 1583533
Avatar for maartenl945
maartenl945 3 hours ago
Topic:
Post Reply

Welcome to the Gearspace 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…

Forum Jump
Forum Jump