High quality nonos PC based DAC

[rfbrw]

Originally posted by fedde


The latency of EQ on a PC will likely be lower than on an external EQ. An external EQ will need to buffer data (another time) in order to filter it.

Fedde

Not so.Even outboard digital EQ's using FIR filters have lower a latency than a PC based system.

[fedde]

Not if you use the PC as source! Eg. mp3 is processed in blocks. These blocks can be directly EQ-ed without further delay. An external EQ will have additional delay due to buffering!

Fedde

[rfbrw]

The sane amongst us adopt the horses for courses approach. On a PC one uses a fomat native to the PC. With outboard EQ one would use a format native to outboard EQ. Only a complete eejit would seek to transfer MP3 as blocks to an outboard EQ. If an outboard EQ is processing audio the odds are the data arrived via one of the serial digital audio formats.
Given a level playing field i.e. dedicated hardware decompression engine connected to a source of MP3 data, dedicated hardware will run rings round a PC.

[deandob]

Regarding PC digital equalisation & room correction, check this thread at avsforum:
Digital Room Correction
Detailed step by step guide

Regarding Ogg vs WMA (lossless) formats, not much difference between the two as they are both lossless, however I have faith that the WMA format will be around for a long time.

Regards,
Dean

[Nielsio]

Originally posted by deandob
Regarding PC digital equalisation & room correction, check this thread at avsforum:
Digital Room Correction
Detailed step by step guide

I've been through it (the first two pages of the thread and also the step by step guide), and it seems that the method doesn't take in mind the non-linearity of the soundcard. If the used soundcard is linear, this shouldn't form a problem.

But moreover: (correct me if I'm wrong) this method tries to flatten the sound 100%. This will not sound pleasant at all. You'd want to add a target curve after flattening.
See: http://www.prijsindex.net/tmp/room%2...0and%20eq.html

Because of this (and also to be able to experiment, especially in the bass region, which is a majorly tricky area, and cannot be dealt with measure-wise) I don't fancy the method.

What I want is a 32-band graphical equalizer without delay.

(I've tested Shibatch Super EQ plugin for Winamp 2 before, and although it's practical in usage, it's not HI-FI material: the sounds loses detail (~unlike my Behringer 8024)).

I'd really like to find one for PC, so I can also start to equalize at systems besides my own.

The delay-thing could be worked around (introduce delay for video also), but the loss of detail is an absolute no-go.

Regarding Ogg vs WMA (lossless) formats, not much difference between the two as they are both lossless, however I have faith that the WMA format will be around for a long time.

Regards,
Dean

I have little faith in Microsoft on being a reliable party. Also: I really, really don't like Windows Media Player, and have therefore never liked the WMA-format. Maybe it's because wma-files are created using this software (right?), and I can't take those people serious (can I?).

[Nielsio]

and it seems that the method doesn't take in mind the non-linearity of the soundcard. If the used soundcard is linear, this shouldn't form a problem.

I'm talking about the recording non-linearity.

[swak]

I also find this a very iteresting thread. I have been reading for a while some threads here and on the avs forum and have decided to give it a try.

A fully implemented computer stereo-system (as i'd like it ) could do digital crossover (2 or 3 way) & dsp (both available now with BruteFir on Linux, or only dsp with DRC on Win) to feed an external multi-channel non-os DAC (tda1541 for mid-highs, tda1543 lows). Now, the link between sonundcard and DAC could be spdif, USB or even i2s. It was suggested on this board that an expensive pro-audio card could be spared and get the I2s out of any non-resampling Envy24 based card, such as the Chaintec AV-710 ($~21) or Terratec Aureon (or on the CMI8738). That would be the cheapest and probably simplest implementation.

Anyone done something similar (i2s)?
these two theads are usefull:

http://www.diyaudio.com/forums/showthread/t-22019.html
http://www.diyaudio.com/forums/showthread/t-22556.html

I'll be doing some reaserch before I start ordering what i need.
Regards,

[swak]

Nielsio:

DRC allows any target curve characteristics, it even has a built in euphonic curve - according , to psychoacoutical tests, that is

[Nielsio]

Originally posted by swak
Nielsio:

DRC allows any target curve characteristics, it even has a built in euphonic curve - according , to psychoacoutical tests, that is

Mmm, what comes to mind: the DRC-method uses (recorded) wave-samples to calculate the filter, right? So then it would be possible for me to (somehow) generate a wave as I would wish and use that for it (..thus would be able to use my own measurement methods and experimenting-possibilities)?

I'm sorry to not dig in yet and try the method et al, to get the answers on my own; but seen (some of) you have experience already, that could save time for me (in case it wouldn't give me what I need).

Something I'd also like to share:
My silent computer project:
http://www.prijsindex.net/cgi-bin/pr...ilent-computer

[MWP]

Im another thats been thinking about this for quite a while.

This is what im thinking of doing:

WinXP box (Envy42HT) -> Optical spdif -> Linux box (Envy24HT) -> BruteFIR (xovers & room EQ) -> Envy24HT -> I2S transmitter (TI MuxIT parts) -> Multiple chan DAC (with ASRCs) -> Gainclone amps

The Envy24HT in the Linux box will be heavily modifed for Optical input and the TI-MuxIT outputs.

Im not quite sure on the details of the DAC ill build yet.
Ill probably have fixed gain (no volume control) between the DAC and the Amps and use the Linux box (BruteFIR) to do volume control in 24bit.

Sound like a good plan?

The only thing im unsure about is if the copper connection between Linux box and DAC will cause noise issues... it would be nice to isolate the two somehow.