Fn+Delete works just fine. Didn’t use this in a decade or so
I am writing a convolution engine recently and as a dedicated Windows user since 90s, I will have to agree Mac OS (and Linux of course as both are Unix based) is orders of magnitude easier to deal with and cleaner for audio. Windows audio kernel is a joke, you need a security approval from Microsoft (expensive) to even use its API only to end up with not working drivers in the next Windows update.
Some do, others don't. What I know from my education, knowledge, and practical experience in research and development is important to me, but I'm not trying to convince anyone otherwise.
That does not seem to be an ASIO driver feature. Can you elaborate on what exactly this is?
The ASIO driver, which complies with the Steinberg specification, allows you to control some functionality of both the driver and the DAC.
No. ASIO lets you pick sample rate and clock source, not control the clock itself.
Fair enough. I would want all of the following in your product.
Only original drivers from manufacturers with signatures, certificates, and standard installation. It's crucial for me that my applications run smoothly, always and everywhere. I use only native Windows commands and assembler. I don't even use DirectX or the .NET Framework.
Again: ASIO does not offer what you claim.
So I'm just a wizard
Well, you’re the one believing in magic.. so I guess you’re right then.So I'm just a wizard
I know but that is the real issue, not hardware.
Well, exaSound and Merging will support sample rates up to 352.8 kHz (DXD) and 384 kHz as well as bitstream up to DSD256. The Okto DAC8 Pro comes close. Higher sampling rates have no significant repertoire support nor any proven audible value.
Thank you! Interesting list. But I'm afraid this is team work, while for me it's just a hobby. Nevertheless, I'll answer point by point:
Absolutely yes - everybody has preferences, and they are valid of course. Who am I to tell someone what he prefers? The point though is that a preference should be based on real audible differences, not on biased ones (that is non existent).
I think most skeptics here (including me) see a part of your project as a solution to a problem which doesn't exist in the first place (regarding the topic of DSD vs. PCM). This is AudioScienceReview and 80+ years of scientific research have shown beyond doubt that non controlled listening tests are no valid proof for any claim. Yes, for the people taking these tests the differences are audible beyond doubt, but for other reasons than they think.
You cannot get something better than what it already is; especially in the digital audio domain - you cannot create something out of thin air if it is not already there...
If you use only Windows native commands you are already on the wrong path. Good to hear, though, that you do not use neither DirectX, nor .NET (who in hell would ever use .NET for real-time audio processing will remain a mystery to me)
I think you've somewhat misunderstood the project's essence. It treats PCM and DSD formats as equal. When outputting via WASAPI, PCM is even preferable.
Guess what, I did ask ChatGPT. Gave the above quote and asked:As for the issues of jitter and delta-sigma DACs, there have been so many articles, studies, and discussions on these issues in the last decade that I thought everyone already knew everything.
Ask an AI about it; it will explain it better than I can
He’s mixing two things that often get bundled together in audiophile discussions:
---
## 1. **Jitter**
Timing variation of the sampling clock.
### What it actually does
- Moves sample moments slightly in time
- In extreme cases → creates **phase noise / distortion sidebands**
### In modern DACs
- USB async + internal PLL + buffering
- Jitter from the source is largely **rejected**
### Audibility (realistic numbers)
- You need **very high jitter** (or very specific correlated jitter)
- Typical modern DACs: **orders of magnitude below audibility**
Practical reality:
If you can hear jitter in a modern setup, something is seriously broken
---
## 2. **Delta-Sigma DACs**
Almost all modern DACs are delta-sigma.
### What they do
- Convert audio to a **high-rate 1-bit / few-bit stream**
- Push quantization noise **out of band**
- Then filter it out
### Common audiophile claims
- “Noise shaping affects sound”
- “DSD is more natural”
- “High sample rates reduce distortion”
### Reality
Within normal operating conditions:
- In-band distortion/noise is **extremely low**
- Differences between competent DACs are **tiny to nonexistent**
---
## Where the confusion comes from
### 1. Old problems that are now solved
- Early DACs → jitter mattered more
- Early filters → audible artifacts
People keep repeating outdated conclusions.
---
### 2. Misinterpreting measurements
- Seeing noise shaping → assuming audible impact
- Seeing jitter plots → assuming audibility
Without relating it to thresholds.
---
### 3. Sighted listening bias
No blind testing → everything “sounds different”
---
## The key point
Both topics are **real engineering concerns**, but:
In modern, competent DACs:
- jitter is **effectively irrelevant**
- delta-sigma behavior is **well below audibility**
---
## What he’s *likely implying*
Something like:
- jitter differences between transports matter
- DSD / high sample rates sound better due to DAC behavior
Which sounds technical, but:
collapses under controlled testing
---
## If you want a one-liner back
**“Yes, and all those studies also show it’s below audibility in any normal DAC.”**Well yeah. This is audiosciencereview. If you make some claims, we generally expect you to back them up with something tangible. Then you want advice and suggestions, but are not willing to share anything of substance once again. That is not how a discussion is supposed to work. If there is no substance, there is nothing to suggest or give advice on.
But I don't need advice or explanations on programming or physics.
Oppo 205?