Always random Pops n Crackles

An area for people to discuss Scope related problems, issues, etc.

Moderators: valis, garyb

User avatar
dante
Posts: 5043
Joined: Sat Nov 24, 2001 4:00 pm
Location: Melbourne Australia
Contact:

Re: Always random Pops n Crackles

Post by dante »

Can't see how Scope alone could fix it - if an ASIO audio stream is delivered to Scope with a click already in it. Cant see how a Soundblaster could fix that either. Other than latency inducing buffering.
User avatar
valis
Posts: 7306
Joined: Sun Sep 23, 2001 4:00 pm
Location: West Coast USA
Contact:

Re: Always random Pops n Crackles

Post by valis »

The issue with the wave driver is NOT the same as ASIO clicks, though if you have the latter you'll likely have more issues with the former.

Scope has a FIXED samplerate and buffer size. Windows' wave driver functionality changes constantly, at first from service pack to service pack but now it can be any number of variations on the API based on what you have for the hardware side, what version of Win7/8/10/11 you're on, and which driver is in use.

When we dig in (I HAVE and WILL again do this, as a gamer and audio professional) we find that the Windows internal samplerate for the older 98/XP era was 96kHz (potentially, if internal SRC was needed to combine multiple samplerates in playback and/or recording) and 192kHz (can be higher now, I think) for the modern HDAudio spec.

If you have multiple different applications (say, two browsers and an audio player) each feeding different buffer sizes (imagine they are NOT evenly divisible, simple math/jitter image in your head), how do you think the combination of a realtime SRC and varying non-related buffer lengths is resolved by SCOPE, which has a fixed buffer size and latency?

Herein lies the issue, with the case that ASIO is working and the WAV driver also (mostly) functioning. It's such a simple thing to use another machine in this era, anyone who is still complaining about a 20 yr old Scope card not working in Youtube surely has something more than the P3 they once had...

On this late note, this is my firm PERSONAL opinion that some just want to complain when there's work to get on with.
User avatar
ARCADIOS
Posts: 1339
Joined: Tue Aug 02, 2005 4:00 pm
Location: Glyfada, Athens-Greece
Contact:

Re: Always random Pops n Crackles

Post by ARCADIOS »

It's not about complaining.
For me its a 20 year old system (not the pc, the cards)
But happens the same to xite with newer systems.

In the perfect scenario of asio click free systems, is that achievable in 3ms ully also?? ( I mean every system may have it's limit that the when passed the clicks will appear).

It's about knowing clearly pros n cons of our system.

Perhaps the solution could be to add a raise of the latency of ully even more by Soniccore, or to add a latency setting to wave modules.. so to perhaps cover a wider range of pc - Soniccore matchings of the chaos of differences in each and everyone's equipment.
User avatar
garyb
Moderator
Posts: 23248
Joined: Sun Apr 15, 2001 4:00 pm
Location: ghetto by the sea

Re: Always random Pops n Crackles

Post by garyb »

i use 4ms @44.1k
User avatar
valis
Posts: 7306
Joined: Sun Sep 23, 2001 4:00 pm
Location: West Coast USA
Contact:

Re: Always random Pops n Crackles

Post by valis »

Again, it's because you have multiple difference applications that maintain connections to these devices. The best way of avoiding this that I have found (even with RME, MOTU, M-Audio, Focusrite, Turtle Beach, Soundblaster, Realtek, Azalia, Nvidia/AMD 'HDAudio' 'ala HDMI etc) is twofold:

1.
Assign ONLY your professional applications to Scope (or your Pro audio card). For WAV, make sure that in windows this is NOT assigned as 'communication device' or 'default device' anywhere. You can do this both in the modern Win10/11 control panel area ( System > Sound or right click on the Volume icon in the tasktray area and choose "open sound settings") AND the separate still necessary "Sound Control Panel" app (where I prefer to always check) that under current versions of Win10/11 you bring up System > Sound and choose "Sound Control Panel" in the top right.)

Once the Sound applet (it just says "Sound" still once open) is up, on the Playback tab choose your onboard audio or a consume oriented chipset as both Default Device, and Default Communications device. For the SCOPE Wave driver, bring up its properties, set the samplerate match Scope, and set the bit-depth to match Scope and your settings in Device Manager (where you set the number of WAV driver devices for Scope as well).

You need to also do this for the RECORDING tab, and the samplerate is the part that will cause you issues. Make sure the playback, recording AND scope samplerate settings are identical.

2.
This applies to consumer soundcards as well! I have measured an increase in gaming engine latency from 150ms (including audio playback) to sometimes as high as 300-400ms (sound gets out of sync and/or the game engine has problems maintaining framerates due to the internal time alignment buffers necessary). So these SAME settings, depending on your choice of devices, can actually impact ALL of your Windows software using them! For my gamer friends, usually this happens with a combination like a Webcam (while not technically 'in use' but not disconnected from the system, and often only truly 48kHz in the device with all audio/data sent COMPRESSED to the host PC, not PCM or PCAM audio), a midrange $50 gaming headset which uses USB and an 'onboard DAC' (and thus is typically fixed at 48kHz) and then some onboard playback as well (like keeping a twitch window open on a second monitor not assigned to their headset). Usually in this scenario, people will have massive performance issues and once I get them to set the audio settings properly and DISABLE unnecessary things, it all goes away.

The SAME thing happens when multiple browsers are open!

Now, the reason that consumer devices (aside from the fixed samplerate USB device) work fine in this case is because Azalia & Realtek are VERY simple chipsets that handle PIO audio and delivery to that chipset via a software codec that handles encoding the datastream via the operating system's driver for it (this is oversimplified, but there is NO hardware accelleration for onboard audio! It's all PROGRAMMED INPUT & OUTPUT meaning entirely serviced by the CPU, which is not so bad in the multicore era). So when you use a realtek chipset, Win10 internally keeps the SRC engaged, and will as much as 350-500ms of latency (and then time align video playback if necessary) to keep all the different datastreams aligned.

Think this applies only to Windows? On iOS (iPhone, tablet) I can have playback going in any audio device and visit most Conde-Nast owned websites (like Ars Technica) and the audio playback STOPS in my other application with the web browser in the fore! Why? They are running scripts that actually sample your background audio as part of their ad platform! (Something Google started doing about 2004 ish). So you *think* you only have 1 stream going for Youtube, when in reality you have the API hooks open to ANYTHING that is allowed to make those device calls (remember, we are given OS level controls now, but clearly those can also be bypassed)
User avatar
valis
Posts: 7306
Joined: Sun Sep 23, 2001 4:00 pm
Location: West Coast USA
Contact:

Re: Always random Pops n Crackles

Post by valis »

So in the above long text, what you can find is that onboard soundcards are typically PIO mode devices where EVERYTHING is handled in the software stack, and the datastream encoded via the onboard chipset (and its codec) and delivered for DAC playback (goes OUT the selected onboard output). The software stack itself manages the ENTIRE process of time aligning and samplerate converting all active and passive (API hooked but no data fed) software inputs and outputs to the onboard device i/o's.

Contrast this with SCOPE, where the nature of the DSP environment is working at a fixed latency/buffer to the host OS (ULLI) and fixed samplerate which CANNOT be adjusted on the fly to meet the demands of opening a new browser or starting another tab of audio playback alongside all of the other already existing API (OS-level) datastreams present. HERE is why the wave driver is not really comparable to your typical onboard device. Perhaps there's a programmer out there talented enough to help Holger with this? I can say that even Realtek had so many issues over the years (well known to gamers) that to think that I learned all of this just because of SCOPE itself would be utterly silly.

By the way, Macs avoid some of these issues by ALWAYS having an SRC for OS level API's, and ALWAYS having a 'safety buffer' that cannot be bypassed by the audio driver stack. RME has documented this as the reason why 1. Digicheck does not work under Mac with the same low latency operation it has on Windows (where it can directly access the FPGA and doesn't need to collect bit statistics from the OS level of the datastream) and 2. why buffer sizes on Mac with the SAME (PCIe or USB/Firewire/etc) soundcard are always higher.
User avatar
ARCADIOS
Posts: 1339
Joined: Tue Aug 02, 2005 4:00 pm
Location: Glyfada, Athens-Greece
Contact:

Re: Always random Pops n Crackles

Post by ARCADIOS »

Fully agree.
1. About the settings you mentioned above, these have been set this way always to me.
2. I mentioned previously a perhaps simpler mothod to handle this situation, by extra adding raising latency (more than what exists now) for asio and wave(not existent). Perhaps it could help with the cost of more "latency".
User avatar
valis
Posts: 7306
Joined: Sun Sep 23, 2001 4:00 pm
Location: West Coast USA
Contact:

Re: Always random Pops n Crackles

Post by valis »

The extra latency not only gives more time for the ULLI buffer, but less need to PREbuffer in the OS in order to time-align things. I think I will repeat the point though that I have seen this concern (wave driver issues) across EVERY device, from $0.12 onboard chipsets to our beloved legacy Scope cards. I have NOT yet had a chance to use an Xite...soon!
Post Reply