Love this guys vids, just information. He's not going to convince us he's designing a speaker without ever designing any speakers prior, or coming out with an amplifier etc.
😎Great stuff Hans! 👍I can agree with your comments bc I’ve experienced the effects of some of the variables you describe. Thanks for the additional insight into digital techniques.☕️👨🏻
Just wondering what your thoughts are on the sound quality comparison between using a streamer with a variable volume output versus fixed volume output on the streamer and controlling volume with software like Roon?
Thank you, Mr. Beekhuyzen! I would be lying if I claimed to fully understand the concepts of which you explained, but hopefully, I will by the time I record my debut album.
Hey Hans. Your implicit statement that a 3dB attenuation is always enough is not correct. In some situations the (correctly) reconstructed waveform can show much higher peaks, depending on the material. 6dB or more, although at some point the data needs to be prepared in a special way. If you have an infinitely long file you can create an infinitely large swing... -3dB is good, -6dB is safer. And remember, you're taking this off of the LSB of, say, a 24 bit sample. And that information is already buried in noise. You don't lose anything and gain protection from clipping.
I’ve found that with a Mac it’s more like -3dB automatically. On a windows computer it always pushes it right to the edge of 0 and I don’t really prefer that. I’ve always found Mac and iPads to sound better with my dac and I wonder if that has something to do with it. I used a Motu interface to monitor the levels from each one as it has meters on it.
Any idea Hans what percentage on the Qobuz app 3 or 6 db equates to ? I have played around before where reduced Qobuz volume sounds better but got paranoid about missing details and pushed it back to 100%.
After some investigation this is what I've found, JRiver has a excellent spectrum analyzer, I found a test tone on Qobuz and could see -3db equates to 71% and -6db to less than 50%. So anyhow I figured 50% is a good mathematical number to use and interestingly I seem to remember that's the default desktop player volume. Then watching the spectrum analyzer in JRiver and playing several tracks I could see many peaks getting not so far away to 0db / full volume, so I think I'm settled on 50% as the safe volume. Btw there's a trick I found to fixing the Qobuz volume, find a file in the Qobuz dir named 'player-0.json', right at the start you'll see '"volume":50', I then make that file read only, Qobuz then starts with the volume set to that percentage.
Multiple RU-vid reviews complain about digital overshoot and have no Idea how to mitigate because they have chose not to understand anything about the hobby. They sit in front of the speaker and cast a binary judgement Thank you Hans for you commitment to fostering better listening & enjoyment
The Hobby term was for those that swear we hear nothing different than a clock radio on a night stand, because what we do is no more than that!!! But someone with that view has no idea what the miss( stay thirsty for better sound my friends)
Thank you for the informative video! I would like to try this. I am streaming Quobuz from my Mac Mini to an external DAC connected via USB. Both the Quobuz streaming app AND my Mac Mini each have a volume control. Each one is currently set to 100%. Should I apply the -3b reduction to the USB output volume control in my operating system or to the volume control within the Quobuz app itself? Many thanks for clarifying.
Hi thanks for the video really interesting. Can I please ask some advice? So I have Auralic Altair g2.1 feeding active speakers and using Roon. As far as I can tell no way to try the -3db trick via the Altair. However via Roon in the MUSE adjustments you can put in the -3db I tried to compare with this off an on volume matched - no dramatic difference. Possible slightly smoother sound when -3db. Is there any theoretical merit doing this in my set up or will it make little difference? Thanks
I don't have hands on experience with the Altair. Perhaps Mr. Wang, who is a very clever designer, has takes measures internally to do the same as my -3 dB trick? Anyway, if you don't hear a difference, you're fine.
Well, converting to DSD the signal has to go through reconstruction filters too plus through the noise shaping algorithm. The question is what filters sound the best: those of the DSD conversion or those in your DAC.
Hi, using the -3db I had to increase the gain of the pre audio research ls26 towards the final mccormack dna 500, By raising the gain did I lower the quality of the signal by inserting more noise?Thank you
Well, you didn't insert noise but, true, your signal is now 3 dB closer to the noise floor of your amp. But the room noise and noise in the recording will be 15 to 25 dB higher than the noise of. your amp. And you increase the sound quality where it matters most: at high levels.
I take it then, a well-made re-sampler/re-clocker using a good FPGA chip, should produce better sound than the standard conversion done in the DAC chip. Yes?
You know I have been in the bitperfect tribe until now. But you made me think again. And remember reviews of dac Cayin Ru6 dongle R2R. They say it has 2 operation modes, that is bitperfect and over sampling modes. And also say over sampling has its merits. Does it have to do with what you say of over sampling advantages? That is, to “help” the dac?
Thank you for this, sir. Another great episode. Now the next question that comes to mind naturally is "should I bypass my preamp". As far as I know, that depends on the impedance matching between the DAC and the power amp. And if the DAC can't drive my power amp, then "should I physically bypass the volume control (DIY) in my preamp". I wonder if anyone experimented with that.
So presumably it's the job of the DAC designer to ensure that with a 0dBFS digital input signal, the analog output signal never clips regardless of the reconstruction filter used? Specifically with a 0dBFS sinewave input you should have 2V RMS (RCA) with as little distortion as possible (i.e. no clipping).
@@TheHansBeekhuyzenChannel I assume the issue is when people start using positive EQ on music that's already recorded near to 0dBFS? You run out of headroom?
Thank you very much for all the information and the time you dedicate to this project. The improvement when using an alternate power supply in series (linear) in a DAC, is it easily noticed or do you have to have a very educated ear to notice it?, I own a Musical Fidelity V90, thanks!
Thank you, this is very helpful. I recently changed audio output from my PC from an internal sound card (Creative) to a DAC from IFI and I had problems. I used PowerDVD that can bypass, but not reduce signal, and that was a problem with some records. I'm using Foobar at the moment, that 'should' sound worse by not bypassing, but it sounds significantly better. 🙂 I should probably buy a better software moving forward, as a good enough DAC will require a lottery win. 😂
A 'good enough' DAC these days continue to fall in price. Unless your audio system is very high-end, there are reasonable consumer DAC alternatives that will likely impress you, over anything you computer can produce. One such DAC is the E30 from Topping. If PC playback is your primary source, then give this DAC a try. You can't beat the price at $200USD. After upgrading your DAC, you may feel the need to re-evaluate your speakers for near field listening. :) Good luck.
@@redjr16 I'm not a 100% sure but I believe ifi is in the same category as topping regarding sq (according to measurements I've seen), it's probably the psu and other surrounding parts that will make more of a difference. Using a PC will be a bottleneck compared to something high-end, but I don't have a dedicated listening room (just different panels and changed furniture) so it might be lost anyway in reverb etc (and I still have some hardware left to try and swap). Speakers are decent, but I might upgrade their DSP later to something with I2S or other digital input.
@@wasaaapdroid9477 There's a bunch of audio processes taking place in windows and they can affect the sound, so some software can route the sound outside of them. At the moment though I believe I just have every windows audio (notifications etc) shut off in windows so there isn't any, it depends on what software and hardware that's used.
Hello Thank You very much, in previous videos You mentioned that in Roon, You use Minimal Smooth Filter while upsampling. I use it us well with my Denafrips Pontus R2R. I have pretty strong PC and would like to take advantage of it and HQ Player. Which filters in video are you recommending for best of the best outcomes from upsampling? Thank You very much for your time.
@@TheHansBeekhuyzenChannel Thank You for your replay unfortunetly its out of my price range, is there any chance You could provide information or make separate video about HQ Player and filter/settings that You recommend for R2R dacs and maybe another episode for DeltaSigma Dacs ? Maybe even settings for different generes of music if we are lucky enough ? We will much appreciate Your knowledge us an audience
Maybe I am confused - but the -3dB trick seems to be addressing an amplitude issue (like amp clipping) and in the digital signal, amplitude is covered by the number of bits of each sample. The issue of trying to encode high(er) frequencies and/or to avoid the need for steep reconstruction filters - is where the sampling frequency/up-sampling frequency comes in - and this is not connected to the amplitude, it seems to me?
Also, read up on what Inter Sample Peaks are. They are a natural result of sample reconstruction and can generate peaks that fall outside the dynamic range of the converter and/or the surrounding electronics. :)
Hallo Hans, Fascinerend. Ik gebruik een Limetree Bridge met Merason Frerot DAC. Ik stream via Qobuz. Kan ik de 3dB trick ook uithalen door digitale volume controle in de Limetree app aan te zetten en die dan net iets onder de max? Dank en groet. Bas
@@TheHansBeekhuyzenChannel Practically all dacs made in the past decades do internal upsampling as they are delta sigma. So the average 44.1kHz dac is upsampling. Upsampling involves a form of reconstruction.
Excellent Description Hans, thank you. After your first video introducing this trick, I set the volume control of my MiniDSP Studio SHD to -3dB, which is receiving its stream from a Mac Mini M2 using Roon and then from the MiniDSP, with DIRAC, to an RME ADI-2 FS DAC. To my surprise, I heard an audible and favourable difference. What I interpreted as quieter was a darker background resulting in more detail space and what I describe as speed. Digital, as you regularly demonstrate has much potential and while binary, there are many things that can influence the sound quality, ultimately! Keep doing what you do, please!
@@TheHansBeekhuyzenChannel I will also share I went all in with Dirac through a miniDSP Studio SHD because of your great advice and I will never look back. Thank you 😊
@@olivergnass7566 I also have an Mac Mini M2. Do you have any advice regarding equipment I need to get to get a superior sound. I only play files from my hard drive. I am a complete noob. Also do I need "bit perfect" software; how do I know if it does that job? I currently use an APP called MusicStreamer
Voor de duidelijkheid, het gaat hier niet om -3 vs -6dB. Het gaat erom dat het signaal zachter wordt gemaakt voordat het een upsampler of een DAC bereikt.
I´m trying your insights this way (my DAC is an entry level one - Topping E30 mark II with WIIN mini streamer, for now). So I use the DAC only (not the preamp include - volume set) - I connect everything to my Quad Artera Preamp and set the -3dB for the output. Did I get it right? Thank you very much Sir for all these great videos.
@@TheHansBeekhuyzenChannel Thank you :) I will try to check if that is possible on the Wiin mini! I will share my findings with you. Once again, really appreciated Hans!
The only way was to set -3dB on a custom equalizer trough all the frequency range on the wiim mini. Seems to my ears at least a little less sibilant. Is that correct Hans? Thank you 🙏
@@AlexandreakaXne surely it's just turning volume down a touch? I've got the wiim pro and mini and I just lower volume a touch on the app via phone so that it is presenting a slightly lower volume to the DAC.
wiim pro, chord mojo2 to amp by using roon. where do i set that -3dB please? and does it also work with a streamer with internal dac used? keep up the great work
Love your videos Hans. Always informative and helpful. I used to have an Ipod Touch 2nd Gen and it required a minimum of -5 dB gain reduction on any mp3 to stop clipping and sometimes even a reduction of -8 or more was necessary to get clean output. Didn't matter which bit rate the files were or the encoding software used to create them.
Sir I just love your videos ! What you think of EQUAL LOUDNESS CONTOURS? I've seen some audiophiles who listen at 75dB SPL equalizing their response to ELC curves ! Well I would love to know your opinion on that ☺️🙏
At least in theory that gives you the flattest response. But our auditory system is rather capable of compensating for other levels. I am more concerned about the stereo they use. Often systems don't perform in balance at lower levels.
I don't think this will actually work since all music is recorded at a different loudness. If your set is flat then you still would need to listen to the music at the appropriate loudness (so at the same loudness it has been recorded at) to combat ELC effects. Having loudness fixed at 75dB basically means you would have to adjust the EQ for every recording.
