I can respect that. I have a McIntosh C2600 preamp which is tasty and euphonic. I have rolled tubes and ended up sticking with some used tubes made over 50 years ago - hardly SOTA!
My philosophy on a DAC however is measured performance. I want the cleanest most accurate reliable conversion possible with proven track record and independent measurements to prove it - all for a reasonable cost and not crazy money.
So measurenents are important to me with regard to DAC. So my praise for PS Audio DAC is because they measure beyond reproach. My concern about Schiit being that they rarely if ever measure well.
I am currently using Benchmark DAC3 but could be just as happy with Bryston, Mytek, Oppo, PS Audio, Auralic, Weiss, Bricasti, etc. There are many DACs that measure superbly these days.
Benchmark corrects for inter sample overs. Not sure if any others do. Benchmark is moderately priced for its performance IMHO. You could do as well for less with an Oppo 205 or RME-ADI 2. Of course, this is all based on independent performance measurements - and I missed dozens of DACs that perform well but may not have been measured independently. Of course, if you evaluate purely by listening then you really have to include hundreds more choices and of course many flavours!
I donât think that many of the low level errors measured are often or normally audible and just because it measures well does not guarantee there are no faults. However, I like to use the only yardstick we have (measured performance) and, letâs face it, human ears are simply not as sensitive or accurate as modern test equipment.
Measurements mean very little with DACs just like it is with amps and preamps. I have listened to some mediocre sounding DACs that measure well and some DACs that donât measure quite as well that sound much better. For example, I think the Schiit Gungnir Multibit sounds noticeably better than the Benchmark DAC3. The Schiit Yggdrasil sounds as good as my PS Audio DirectStream. Both sound MUCH better than the Benchmark DAC3.
I havenât seen measurements for the PS Audio DirectStream. I never really cared about them because I know they donât really mean anything. Are they published anywhere?
Human ears hear things the measurements we currently use arenât measuring.
Oh yes I agree about hearing. Tubes would not be so popular if stellar measurements were all that pleased the ear. If you evaluate only by listening then tastes, ancillary equipment, speakers and the room will lead everyone toward different choices in a DAC.
In the context of the entire paragraph I wrote, it is clear that I am expressing my opinion on what I hear. You are being disingenuous by using partial quotes. You are the person with the problem here, not I.
Wrong: we are all the problem here, otherwise weâll enjoy some good music somewhere instead of being here! My problem: I keep telling people that they are wrong!
Here is a comparison. The red curve is Benchmark a DAC 1 (no inter sample overs capability). The grey curve is a Benchmark DAC 3 which has extra headroom to handle inter sample overs.
The problem is worse at low sample rates - for example CD and AAC 22.05. It is ubiquitous in all hit mastered rock and pop. Even audiophile stuff like Steely Dan albums will have many thousands of clips. It is possible that this one of the reasons folks love upsampling to DSD and why they hear audible improvements from the processing to convert to DSDâŚ
Jeremy, I do think you praise Benchmark and its intersample over handling a bit too often. It is not ârocket surgery.â In Roon, turn on volume leveling or headroom management â intersample overs in downstream components eliminated.
Good point. Guilty as charged. Roon can fix the lack of headroom for any DAC by reducing the digital signal by 3 dB (digital volume control) but not if you use bit perfect (I think).
I prefer bit perfect but others may not.
That so few DACs address this common pop/rock issue is rather disconcerting to me but I understand that this doesnât surprise most people.
I wonât raise this anymore as It seems many here know about the issue already.
I believe a little bit of on topic will hurt no one. Anybody?
Bill_Janssen
(Wigwam wool socks now on asymmetrical isolation feet!)
43
I really appreciated the link to the article, as I hadnât heard about this aspect of oversampling before. At least, not in the context of audio D/A converters.
Then, only non oversampling (NOS) 16 bit resistor ladder DACs for Jeremy.
Look at it this way. Bit perfectness has to cease somewhere along the digital signal path. Otherwise, headroom management (to avoid intersample overs), oversampling, sigma delta conversion, etc., are not possible.
Yes, I understand, Jeremy. But software music players are performing many of the same tasks now that used to happen exclusively inside DACs. And whether that sigma delta conversion, for example, occurs in software or silicon is largely immaterial. Either way, bit perfectness ceases at some point. So, you may want to reconsider your perspective. Just some food for thought.