My Digital Audio Journey (So Far)

That is mostly fine, but Why does it show as ASIO for "Generic Device". Did you install the official driver form your device manufacturer? I think it should typically show as "<Manufactuer/Device> USB ASIO driver". For me it shows as "Apogee USB ASIO driver".

Give it a shot with these settings but meanwhile check it the drivers are from the official manufacturer. The official driver will have a control panel where you can adjust the buffer, latency priority, individual channel volume (this will send the volume info as a control signal, while the software mixers will use CPU to scale the levels).

 

FYI: Minorityclean runs as a background service. Double click it and you'll see it in the task manager. Once it's up and running it'll do its job. Just restart your music player and enjoy.

 

Oops good catch, forgot to screenshot it with the official driver I use with my DAC:



So that's all that needs to be done for settings in Winyl?

I skimmed around your original post, my eyes glaze over this kind of stuff, nothing personal.

 

Yep. Good to go. First just try comparing winyl and other players. Then try comparing winyl with and without Minorityclean. To disable Minorityclean, just remove the task from windows task manager and restart your music player. Once you're done, play around with buffer settings in the control panel for the dac to see if you notice changes. I just keep mine at max buffer and safe mode because all I do is music playback. Beware, there seems to be a bug in winyl when playing back wav files (I explained it in posts in first page), when playing back wav files in ASIO, please keep the volume low and your arms in a position to stop playback asap. Sudden burst of noise can happen.

It's okay to be skeptic. I used to be one, and still am. But it's also important to be open and acknowledge that we are not computer audio geniuses (yet).

 

Yeah I'm a total skeptic until I hear it. On headphones right now thru Yggdrasil WASAPI event, and yeah Winyl is definitely clearer than foobar2000. Just a tiny bit, also just a minuscule bit of static reduction. It's a tiny bit smoother and barely more reverb. Cymbal crashing is tiny bit less choppy. Splitting peach fuzz hairs here. Playing the same tracks back to back in both players. My foobar2000 is an old version, v1.4.2 so I'm almost 99.9% sure the ASIO/WASAPI is not going through the Windows Mixer like the new release as Psalm said earlier.

Is there a setting to stop the fade in and out of playing different tracks? Will give that executable file a try shortly.

BTW: not sure I can evaluate any changes in softness, as Yggdrasil has a hardness to it, especially in the bass. I have a feeling this is from the Gen 5 USB cause Bifrost Multibit had the same problem for me.

 

I am unsure how to get that fade in/out off in winyl. I use winyl since it is plug and play and I prefer its UI. The fade in/out doesn't bother me much at the moment, but HQplayer and Xmplay both don't have that issue. You can try the same tracks in HQplayer (demo version usable for 30 days) or xmplay. Both sound identical to me, as is winyl.

Regarding foobar, I've tried version 1.3 and 1.5, (with all the sub revisions in between, lost track of all of those). Not sure of 1.4 but different versions sound different and none sounded anywhere near as good as winyl or even Jriver for me. It doesn't seem plausible that the middle version (1.4) would have been any better, but I could give a try if I can find that executable.

 

LOL don't worry about it. I'm ditching foobar2k as well, this is a nicer application. When you increase the buffer on ASIO to max, what benefit do you get from that? Lower latency? I've seen that buffer length setting in foobar and default set at 1000ms. My ASIO driver doesn't give me a buffer option.



Here is the generic ASIO driver:

 

Buffer - a person or thing that reduces a shock or that forms a barrier between incompatible or antagonistic people or things. The buffer we mean here is IO buffer. It is a stack/queue which stores data before being fed to the dac.

The advantage of a higher buffer is giving the system enough scope to be able to have less data loss since the device will pre-fetch and keep the data for the next time cycle. If you think in terms of system response to your input to play/pause the thing, a higher buffer will make it more laggy. Higher buffer will cause higher latency.

If you want real-time output, like in gaming, you are better off reducing buffer and aiming for faster latency, else you'll have audio-video sync issues. If you want to play music, where input lag isn't a big deal, keep it at high buffer (which however will mean worse latency).

I am unsure of dangerous music convert's driver structure. Maybe psal could help you with it. I am not sure why your ASIO control panel forces a specific bit depth. The asio driver for all 3 dacs I have switches bit depth and sampling rate on the fly as per the input music. It is communicated to the DAC in the control frame (the first frame that is sent before sending data).

The above are for IO buffer in dac. The buffer setting in foobar is just how much of the data is pre-fetched into the RAM. IMO, once you go ASIO, such concepts are mostly useless and you're better off flushing everything to RAM and play from there (JRiver and HQplayer have these options, I'm not sure what winyl does).

Sequence is:

Storage Device -> RAM -> USB controller Buffer/DMA -> Cable -> DAC Buffer -> DAC USB to I2S decoder -> the Acutal Dac chips.

There are buffers and caches everywhere in this link. And every single of these data transfer will be controlled by instruction sets fed to the CPU or instructions given to the DMA in these devices.

 

Gotcha, thanks for the explanation!

You can ignore that second screenshot, it was just showing you what my generic ASIO driver looks like. I don't use it for playback. For the XMOS driver, I have to manually set the bit rate/depth on the converter itself, then double check in my properties to make sure the default format matches what I set the converter too. This is because my converter has a clock mode where my PC source slaves to it. Any other clock mode I can just switch on the fly in sound settings.

 

I long since abandoned Windows. Insofar as sound matters to me any longer (@dietwater, I'm the deaf guy around here; I just stick around for the adult content) I use Linux tweaked to increase the priority of audio processing. This is probably available in several distros now, but I use Linux Mint with added features from KXStudio. I wouldn't be surprised if most, or even all, are built on rtirq, written by Rui Nuno Capela.

 

Nope but most players support direct output to alsa which is equivalent to asio/wasapi. Years ago i tried oss which is an alternative to alsa and i found it did sound a bit better but even then it was very buggy. I bet it's pretty much dead now for linux, maybe Unix support is better but then also comes the question of hardware support. Oss is quite limited compared to alsa in that regard.

 

I can’t speak for the softwares here, as I use a Mac. But I’ll definitely agree on the Supra cable. For $35, it is bloody good. In my testing, it proved to be noticeably better than every cheapo USB cable I have lying around here.

I’d still recommend a good SPDIF source over USB. But if you have no choice but to use USB, then I’d highly recommend the Supra.

 

An update on Foobar2000 vs Winyl. A friend of mine tried to measure the two players with his oscilloscope and has used his function generator as golden. 1Khz Square wave. He first tried 5khz wave and then skipped it (for a reason I will explain in the following posts).

We already had some measurements with 1khz sine waves but his oscilloscope wasn't able to register the low frequency signals properly so we discarded it (the same reason my oscilloscope sucks in audio band - not enough bit depth to do reasonable fft in low frequencies). It is important to note that the test scenario is still not without variables. He used a 32bit 384khz file on a 24bit 192khz dac. He compared all in WASAPI, but ASIO on Winyl measured the same as Wasapi on Winyl. He didn't bother with trying ASIO on foobar, but I can confirm ASIO and WASAPI sound almost identical on foobar (and they both are bad).

Here are the measurements.

https://imgur.com/gallery/50P4hRJ

I repeat that these measurements are not to be taken as gospel. There are still variables to iron out. I am posting it only to boost confidence for people with better gear to go ahead and make measurements, and expect results.

Edit:: the Images are of 5khz wave.

 

Diverging a little from the post, I just wanted to share this sweet little thing I just got. I swapped the V6 Vivid op-amps on my Burson fun to Sparkos ss3601 yesterday. All I needed was some 10 minutes of listening to come to a conclusion. This setup is probably the first piece of audio gear I've heard that I could recommend to anyone without reservations. Typically I write long passages what something does good and bad and its compatibility, etc. (you might have seen an indication of how I write already). Unfortunately I don't see any scope of nitpicking here. Nothing sounds withdrawn but nothing sounds forced either. There's that sheer sense of palpability and liveliness and the feeling of being there with the artist/the scene happening around in real life. Make sure the rest of your setup is not compressed or off, and given your power requirements fall within 2 watts peaks, you've struck gold.

The setup is a solid 9/10 in my book. -1 points for the poor documentation/manual provided by Burson Audio and some random power cycling.

 

The simple explanation I can give is that the sparkos update made the bandwidth better - which means better bass, better highs and better power handling. With the V6 vivid, the rest of the circuit was probably trying to extract too much gain out of the opamp making it harder to maintain bandwidth. Everything else was perfect - coherency, noise, etc but you can feel something was missing. It sounded somewhat dry and lacking in vividness (despite the vivid name). Sparkos is much higher open loop gain while being able to keep up with an even higher bandwidth. The setup sounds literally flawless now and there's nothing I can think of that it misses (comparing with how I perceive sounds in real life). Pretty much every headphone I throw at it sounds wonderful - SRH1540, HD800, HD700 or Urbanite xl.

For those who didn't get what I meant by higher bandwidth - https://en.wikipedia.org/wiki/Gain–bandwidth_product . Yes opamps have practically very high gain, and can be considered infinite for a lot of applications, but depending on your circuit, it may not be true for audio. I haven't tried tube amps yet, but I believe bandwidth would be pretty amazing on those as well considering how much of a heavy workload they are designed to perform. The short thing is, opamps are not the mythical creations that can fix all your audio amplification woes.

Does it mean the sparkos will rule the roost on all parameters. Maybe may not be. There is always a catch. You gotta get a nice balance between Bandwidth, Gain, and Noise to match your circuit well. Maybe the noise floor might be higher on sparkos, but I'm unable to hear it. There's also the concepts of biasing though I think most audio op-amps can be just plug and play.

Some rudimentary measurements I made with my Hantek 6022 BE oscilloscope. This just like the player measurements, is not to be taken as a weapon of justification, but rather something to give confidence to others with better gear to measure. I used a 5khz Square wave from this website : http://onlinetonegenerator.com/ . Yeah despite all these pedantic stuff I made about audio players, I had to resort to an online generator since I was too lazy to code my own tones. I made few other measurements just to confirm what I had been seing. I forgot to scale the thing when making the snapshot but I have done comparsions with equal scaling as well.

https://imgur.com/a/ZDNKXTJ

Take a look at the shape of square wave on either of them. You can see that the V6 vivid dips a little bit harder. This is most likely an effect of high frequency rolling off (but I can't confirm since there is more variables). Try simulating Square waves and then adding a roll off on top and you'll get what I mean. (check stuff like Gibbs Phenomenon as well, they have a nice visualization of how the shape changes as you add/remove harmonics). Bass bandwidth is also audibly better but my gear can't measure it.

An independent measure by another person at audiosciencereview:
https://www.audiosciencereview.com/...aked-pics-some-technical-thoughts-a-s-o.3958/

Not sure which opamp he used but it surely isn't sparkos. See that there is a phase shift in 20Khz sine waves and some blunting in 20Khz Square waves. He was attributing it to Low pass filter, but I can now assure that it is because of bandwidth limitation in other op-amps when pushed to the gain requirements of the Fun. If he re does the measurements with sparkos it will show less of that phase shift and blunting in square waves.

 

Back to the original post, You can check 5Khz square waves on above post to know why looking at 5Khz square waves is hard in audio applications. We don't have a real golden curve to target, since square waves are practically infinite bandwidth and trying to get your gear to have infinite bandwidth will have consequences in gain and noise.

1Khz square is still inifinite bandwidth but most of the necessary harmonics fall within the 20Khz spectrum and makes it easier to approximate.

A simple rule of thumb for measurement accuracy is - for lower frequencies or slower slope, your bit depth matters more. for higher frequencies or higher slope, your sampling rate matters more.

 

Dietwater, aka the infamous manueljenkin from Reddit, banned from every discord audio server. Says it all really.

For context: https://www.reddit.com/r/headphones/comments/etqutu/finally_found_perfection_almost/


Some nuggets of gold if you really want an aneurysm:

"Srh1540 : the brightest and most detailed of the bunch. Has a sharp spike around 10khz. Also the hardest to drive of the bunch. If my calculation is right, I'm feeding near 1.5 watt peaks for high dynamic range music."

"Hd800: second most resolving of the bunch. Pads were flaking so i threw them away. Was a bit too bassy for my tastes with pads on so I haven't bothered to replace them and now i typically listen without pads. Very sensitive of positioning in the head even more so without pads. Second best depth and detail of the bunch. Incredible bass (even without pads) and sounds very full with most music (not as dry or ruthless as srh1540 could be to a lot of music)."

"Hd700: softest and smoothest can I've tried with this setup. Very source picky and can go into murder mode if your amp is compressed. Sounds every bit as horrible as people claim on the geek out but not on this setup."

"Urbanite XL: the second widest sounding headphone I own. Has a very sharp 10khz spike and is very very sensitive to amps or dacs that clip around that region. The most sibilant set I own - while the hd800/700 have poor resolution there they soften the tones, while the urbanite slaps a hard edged tizz on top of vocals under certain scenarios. Just like shure it sounds zero treble without amping but with proper amping it has a lot of zing. The second widest soundstage but the most natural soundstage. Both hd800 and hd700 roll off upper treble to make stage width apart from midrange dip. Urbanite xl does it through cup reverb. It sounds a bit like a cave but a acoustically well treated cave."

"I think i got one for everything. Srh1540 + desktop setup when i really want that last ounce of detail, hd800 + apogee groove when i want good fidelity but want it to be easy to carry around, urbanite xl + apogee groove on the go, hd700 when I want to smoothen and listen to bad masters."


If you've bothered to make it this far and are thinking to yourself "What the f**k did I just read?", that's manueljenkin in a nutshell.

 

1. It is my opinion. Even tyll felt srh1540 was super bright. It is well known that hd800 and hd700 are amp/dac picky - I just reiterated that they get unstable when your dac or amp is compressed sounding, fine otherwise. Most people who has heard my setup or tried the dac/amps/software changes I mentioned agree with my opinions, but that's another story.

2. What you posted is irrelevant to this thread. I know who you are and what you seek from doing this. I'm not cleaving to that. Mods can better look into it.

3. I already informed that this was not my original name in the introduction post I made. And I was still welcomed and given permission to post. I find hostile people like you sneaking into every place like this now and then. And yes I'm banned from the pleb-fi place known as headphone headquarters (and you're a mod there). It's funny that guys over there try to show you're superior to people at sbaf and somehow end up here again. Lmao.

 

Have you got a winch on your audio player?