I have been reading about hardware upsampling devices and the whole idea sounds (pun intended) interesting, but they are just a computer with a fancy software in pretty enclosure, couldn’t you do the same thing in software if you had pc that’s powerful enough?
Also for people who aren’t into electric guitars, there are so called profiling amps, what they do is create the profile of your tube guitar amps, so basically you get the guitar amp, warm it up and connect it to cabinet and a guitar like you usually would, you dial in a tone that you want and when you’re happy with how it sounds you disconnect it from guitar and a cabinet, hook it up to profiling amp and you’ll get a profile, once you load that profile you’ll get the exactly the same tone, and it sounds great. wouldn’t it be amazing if you could do the same with hifi tube amps, maybe you already can but I don’t know about it. It would sure come in handy, I play a lot of retro games and adding tube buffer between my dac and solid state amp made them sound a lot more pleasant, but I prefer solid state sound otherwise.
You can actually do that in (dedicated) hardware, which a lot of them do for power savings reasons. Since getting custom ASICs or buying FPGAs is expensive, the devices usually are expensive too.
That certainly exists. IIRC Lexicon has multi-effects that can be fed with a “decay sample” (has a different name) and then simulate the same room. This also exists for DAWs.
I don’t know any real time program that does this.
This is what HQPlayer does.
The quality is very much dependent on the algorithms in use for upsampling and filtering, chord for example claim a massive number of taps in their filters. And all the “clever” filters will be proprietary and closely guarded.
The other issue with software filtering, is most DAC’s won’t let you bypass their own processing and filtering, so YMMV.
I haven’t used HQPlayer, but I know a lot of people who like it.
That’s not true in most cases, most dacs will bypass their internal filter if you feed them their maximum sampling rate, for instance RME Adi-2 dac fs will bypass it’s filter if you feed it 768 kHz 32 bit
Modern delta sigma DACs automatically upsample the file exept if the file is sampled already with the maximum supported sample rate, if you want you can upsample the file with some players and the windows audio system before send it to the dac. Which one is best is difficult to say and probably there aren’t differencies. About the profiling amp someting similar in the digital world are the VST and some players like foobar supports them but usually they are effects, instruments and guitar amp simulator so not really usefull in the hifi, the only think that can be usefull is an equalizer.
DACs at a chip level NEVER take in files directly.
They have no concept of filesystems or files. I2S takes in “Word Select” (= Output 1 or Output 2), Data and a Clock signal. That is it.
In case of SPI connected DACs, there is incoming Data of a kind the chip understands.
Resampling/Upsampling is also Digital Domain only, a Digital (to) Analog Converter is a Domain Converter. Some chips include that, most don’t. They just adhere to the provided clock (or the clock rate set internally in case of “smarter” chips).
Edit: Oversampling is something functionally different from Resampling. A common mistake since they sound similar.
Windows does horribly at upsampling (unless they have fixed it in a recent update), it’s the same as integer NOS filter.
I was talking about dac unit not dac chip, and usually people use them connected via usb dacs that have an i2s input are not really common and really few people use them directly via i2s. There is a usb microcontreoller in between the dac chip and the pc and the microcontroller takes a signal and convert it to a file stored in a memory and after it send it as signal to the dac chip so talk about file isn’t wrong
If you found that somewhere, I would now classify that source in the same trustworthiness tier as Bloomberg in regards to tech news.
Because running USB in a mode that supports file transfer will correct for transmission errors since file transfers are not timing sensitive.
Reading material on USB Audio Class:
That’s because you probably consider “file” just informations stored in an non volatile memory managed by a filesystem but actually in computer science is common to refer to a file or stream for any kind of data avaiable over time.
If you think of OSI:
L1 (Physical) just has electrical pulses to get from A to B
L2 (Data) makes sure those Bits make sense
L3 (Network) ensures the “what goes where”
L4 (Transport) handles the “this to there”
L5 (Session) “File was requested to be there”
You could argue that L6 (Presentation) is the first true file-aware layer, is a bit pedantic IMO.
Or if you instead wanted go Software/Hardware, the ISA has no concept of “file”. Machine Code could maybe be argued (depends though). Assembly and above can open files.
I can’t see why I should use an abstract model especially used the in communication for say that a programmer calls file or stream a data stored on memory
I can’t see why refer to the ISA if I’m taking about a file since it’s an abstract model as well and it’s also pretty wrong consider it at software/hardware level. At most you can say that an ISA is implemented in an hardware
Here you’re right a file/stream are the words people use when they are dialing with data stored on memory
Could we stay on topic please.
Update; HQplayer has all upsampling options you could ever want.
Thanks for this interesting information!