About the Grimm MU2 - Part 1 - The Digital part
HTML-код
- Опубликовано: 4 июл 2024
- We visited Grimm Audio in Veldhoven to learn more about the special DAC they created for the Grimm MU2. We interviewed Eelco Grimm to learn how this 1.5-bit DAC actually works.
#Grimm_Audio
#DAC_Technology
#hifi
#audio
#Eelco_Grimm
#Alpha_audio
#interview
#technology 
Наука
A master class from a great teacher. His students are lucky. Can't wait to hear the MU2.
That was a great interview - very informative and inspiring. Greatly refreshing to listen to a designer that speaks plainly, without all the metaphors and anecdotes that typically clog up the hobby.
Thanks
Eelco and Guido, the legends!! Thanks for the vid!
Interesting approach... I use a DIY DSC-2 DAC here, which is single bit and not 1.5 or 5, etc, but uses a similar discrete FIR filter approach. I input DSD 512 to it, oversampled by HQPlayer. As Grimm notes, processing power is at a premium for the SD Modulator with this approach, and that 1.5 bits allows them to have enough processing power to create a near perfect modulator. Of course with HQPlayer running on an I9-9900K, processing power is not limited, and Jussi's (HQPlayer developer) sophisticated modulators are exceptionally good-apparently good enough for single bit use without issues. There are a lot of DACs using variations on this theme now (discrete FIR converters with heavy oversampling, or for use with external oversampling like HQPlayer). Of course there is an advantage to oversampling externally (in my case in another room of the home) as one keeps the electrical noise associated with powerful processing devices away from the audio system. I am sure the MU2 is a wonderful device for those who do not want to bothered with all the complexity of setting up HQPlayer, etc... I suspect the end results are similar, certainly the descriptions of the sonics are very similar to what I experience here: very detailed sound, but fully natural and organic sounding at the same time, unlike the sound of most standard DAC approaches.
It's hard to call a square window a real FIR. Cause every FIR assumes coefficients (depending on the window function). But you can find a Japan project which implements two functions: sine and Nuttall
Interesting, informative and entertaining,, 👍🌟👍
Thank you ever so much for your work and time,, 💖🙏💖
Thanks!
Wow...a fascinating and insightful conversation. Felt like I went back to university! Thank You
Welcome!
Wow the designer by name and action. Super interview.
Always interesting to listen to Eelco and Guido, thanks!
Welcome!
Great interview. I finally got around to listening to it.
1 bit does not in practice have to create any unpleasant noise. Most of the better R2R ladder dacs sound wonderful and the majority have 1 bit decoding at least for sure on the DSD stuff. Cheers.
Great interview!
Great Interview so much is explained Is am a new DAC/Streamer builder
Welcome!
NOS does not relate to an architecture. It relates to a PCM stream processing, and not about a conversion. Probably he meant R2R when said NOS.
Thank you for the interview.
Welcome!
I find refreshing Eelco’s clear and simple analysis focusing on *analog* as the cause of why sources of the same bits may cause them to sound different. Extrapolating this line of thinking, then, when hearing such differences, something is lacking between the digital source and the DAC’s analog output. There is poor analog control/isolation in the DAC’s design and/or responsibility for the clocking the bits has been pushed back to the digital source (SPDIF/TOSLink) making jitter. Therefore, aside from jitter, since it is difficult to characterize how the bit’s source maybe influencing the downstream analog aspects, it is misguided to solve this problem on the digital source side (say by preferring CDs or local files over streaming, using “audio” grade networking, etc). Instead, avoid DAC interfaces like SPDIF/TosLink that permit the source to cause jitter, and employ DACs with well designed analog control/isolation.
Perhaps, if a DAC sounds different when supplied the same bits from different sources (via an unclocked interface, i.e., the DAC is solely responsible for jitter), this should be a negative in a DAC’s review, rather than writing this off to upstream issues or using it as a measure of the DAC’s accuracy.
awesome ty
That pretty much went over my head, but given the U2 is 27k here it’s not likely to make it on my shopping list any time soon. Or ever.
please ask them can they make audio grade switch mode power supply? they make mu1 and mu2 ,can they use that switch mode power supply to create other DC power supply?
We talk about the SMPS in part two. But you can ask them anything via their website.
I dont understand most of this, but thank you anyway❤😅
Haha! Welcome!
An excellent overview of oversampling & noise shaping. I have a couple of questions:
1. I believe one of the disadvantages of over sampling is the jitter requirement of the over sampling clock is proportionally higher (to maintain the same N bit accuracy) compared with NOS? If true, if NOS uses a crystal clock, what clock technology is required for over sampling?
2. By 1 1/2 bits I assume what is meant is a DAC is producing +V, 0, -V ? Where +V & -V are balanced to N bit accuracy?
Thanks!
The 1.5 bits is not +/0/-. Its more like 0, 0.5 and 1. We explain it in our review.
OK, the digital code is 0,0.5, 1 but the DAC is probably +/0/-. with no DC bias therefore no need for output coupling caps?
How about (1.) regarding clock jitter? I've had second thoughts on this. I believe the output rate & resolution sets the jitter requirement. The oversampling jitter only has to meet the requirement that when divided, it meets the output jitter requirement. A hardware divider (binary power of two?) is not perfect, so the degree of imperfection has to be factored when specifying the oversampling jitter.
Different sources, stream vs local it is possible to sound different with the same bits.
With local storage, the entire track can be loaded into memory eliminating network traffic during playback. This can result in different levels of noise vs the same bits streamed while playing.
That’s partly true. Software like Audirvana also loads the stream into memory, so it is not a sufficient explanation?
Thank you for the interview and looking forward to Part 2. A question for Eelco: when will the first MU2s be shipped to the U.S.?
I think you can try to contact them via their website?
Indeed, a good teacher. How does the fpga in the chord Dave/Mscaler compare to the GRIMM MU2 fpga dac?
Totally different technologies and approach. You can't compare it 1 to 1.
From MU2 to power Amp what does Grimm recommend RCA or XLR ?
I don’t think DAC is fully balanced my guess would be RCA but have you tried both and have any recommendations ?
I have chord Dave and use RCA to the power amp . Thank you
De dac is fully balanced and the pre amp as well.
@@TheAlphaAudio nice , sounds like XLR is preferred over RCA . Can’t wait to try MU2 , I’ve had Chord Dave for past 3 years.
I hope MU2 takes my digital experience to another level
I was wondering how the streamer talks to the DAC. Is it over i2s?
I believe so, yes
Just wondering from the previous MU1 video, upgradable path was mentioned. Was the MU1 internal dac possibility mentioned during the meeting?
? There is no dac in the Mu1.
From your previous video, at 14:37, Mr Grimm mentioned that he will add a dac in the future. So just wondering if that just means MU2 or they will offer MU1 upgrades to add dac in the future.
ruclips.net/video/X9Jf2gR-f_8/видео.htmlsi=gSG-9UpSyVNgvpfV
@@jeremylo9816 that would be the MU2 he is referring to.
I'm wondering if Grimm will make a stand alone DAC using the same technology as the one inside the MU2. It'll be a lot less expensive than the MU2, I guess.
I cannot comment on that.
Is there a release date for the MU3?
They don't know if there will be one. This is due to space and heat.
@@TheAlphaAudio Those were my thoughts too, I'm still hoping for the one box solution.
So expensive 😢
Not if you realize that it is: a streamer, a dac, a pre-amp, a headphone amp, a Roon core ánd an upsampler... all at a very high level.
@TheAlphaAudio No serious headphone audiophile would use the headphone out. We spend a great deal on Zahl HM1, Mass Kobo or LTA MZ3. Proper headphone amps. I still think it is far too expensive. Even Esoteric are cheaper!
@@johnratcliffe6438serious audiophiles would listen first before having an opinion…