Jump to content
IGNORED

"Audiophile" sound recorder?


Recommended Posts

The only clock whose accuracy affects recording quality is the clock within the A/D converter. The only role of the computer is to store on a hard drive the digital values produced by the A/D. Timing and computer system optimization at that stage is irrelevant.

 

HQPlayer (on 3.8 GHz 8-core i7 iMac 2020) > NAA (on 2012 Mac Mini i7) > RME ADI-2 v2 > Benchmark AHB-2 > Thiel 3.7

Link to comment

Agree, using precision low-jitter clock is absolutely critical at the A/D side, but why then vise process - playing back - affected so much by HOG and RAM caching?

 

If the playback side is using asynchronous clocking on the converter side, then there's no difference between A/D and D/A in terms of sensitivity for the software/computer.

 

EMI/RFI could be injected from the computer and disturb the process all the same for both.

 

Biggest difference is of course that for A/D, any disturbance will be part of the data forever.

 

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

I agree with Miska's statement: "EMI/RFI could be injected from the computer and disturb the process all the same for both."

 

Unfortunately, minimizing EMI/RFI is difficult and perplexing.

 

HQPlayer (on 3.8 GHz 8-core i7 iMac 2020) > NAA (on 2012 Mac Mini i7) > RME ADI-2 v2 > Benchmark AHB-2 > Thiel 3.7

Link to comment

If the playback side is using asynchronous clocking on the converter side, then there's no difference between A/D and D/A in terms of sensitivity for the software/computer.

 

Exactly. In this point of view there's no difference in A/D and D/A processes. I'm asking about another part, the software itself!

 

It is obvious to me that playback from HDD without exclusivity of using the DAC (ADC?) is worst versus how Audirvana works. That is why I'm asking is there a better recording software…

 

Maybe there's no need to have HOG mode for ADC device as long as no other participants for coreaudio in this chain, instead of playback when many flows are processed or expected to be processed.

 

But maybe there's a chance to find a good recorder that use RAM caching?

Not Pure Vinyl as it too buggy.

 

Link to comment

I'm asking about another part, the software itself!

 

If it's manipulating the data, it's about quality of the algorithms. If it's "bitperfectly" just transferring data from interface to the disk, there's no difference. It doesn't matter what kind of data it is, audio, video, photos or a spreadsheet. It is just series of bytes.

 

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Disagree. See Audirvana posts, again. Bitperfect seem to be more a legend than it is, as soon you go deep inside to the understanding of the processes with coreaudio and HAL.

 

I'd like to see some technical argumentation of what makes the difference. And then we can deal with those technical details.

 

I don't know about posts or hands waving, I only know about writing software and audio/signal processing engines (~20 years of experience on those).

 

If data doesn't change on the way from A to B it doesn't change. This is something that can be mathematically proven.

 

If we talk about CPU load induced EMI/RFI impact on the audio interface (clocking or whatever) we are not talking about software but hardware issue. Let's keep these separate.

 

(Software is just series of logical operations that doesn't even need hardware to be run. The sequence can be "executed" with a piece of paper and a pen too.)

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Then got this: http://www.amr-audio.co.uk

 

The document is still high level hands waving that every audio software developer is aware about. It doesn't have any technical detail in it.

 

(2, 4(1-2), 5) is handled by all serious software. Nothing special. All the trouble with "integer mode" is an OS X problem due to chosen system architecture and not an issue on any other platform.

 

(1, 3, 4(2-3))

That's about hardware issues, mostly RFI, varying on different software use cases (see my previous message). This is something that can and should be fixed by fixing the hardware. Software shouldn't be mixed here. Given modern multicore CPU archictures, the last bullet point in (4) is just funny.

 

At least all the audio production software I know for Windows/Linux use exclusive low-level access to the hardware. But maybe it's somehow different on OS X.

 

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

is there any sound recorder for Mac OS that use exclusive low-level access to the hardware?

 

Someone else may be able to comment in more detail about Mac, but on PC at least Steinberg Cubase and AVID ProTools use ASIO drivers and have also a Mac version. Or the free one I'm very familiar with from Linux side is Ardour.

 

Maybe Apple's Logic Studio would also do?

 

Another option is to use one of the various harddisk/CF recording devices.

 

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

I created RecPcmWin, WASAPI exclusive memory recording application for Windows .

It won't run on Mac OS...

 

I realized big shortcoming on this approach in the course of the development.

 

If you want true memory recording, Memory buffer must be prepared on heap area before recording starts. But in recording, unlike playback, Its buffer size is unknown in advance. Recording time will be decided by pressing stop button.

 

I think recording is more critical task than playback,

If it fails it results permanent unrecoverable data loss.

 

In this approach you must determine recording time before record starts. I think this restriction is unacceptable on some cases.

 

 

Sunday programmer since 1985

Developer of PlayPcmWin

Link to comment

Tried Cubase 5. Surprise: with all its ASIO and "boosted" performance it didn't beat Soundtrack Pro at all. STP gave a little better music after all.

 

If there's difference in the sound, either one is doing something to the data. Or you have one of those hardware issues.

 

OTOH, being accurate doesn't always sound better.

 

What kind of A/D hardware you are using?

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

They use different drivers. STP work with system driver, while Cubase seems use their own ASIO driver.

 

A/D - Echo Audiofire, firewire interface.

 

I can share a link with two short parts of each application record results, 24/96, if you like.

 

Don't know what hardware issues could happen with Macbook Pro unibody, while it is on battery power, with all wireless (bluetooth, wi-fi) switched off.

 

Link to comment

They use different drivers. STP work with system driver, while Cubase seems use their own ASIO driver.

 

And different OS? Since STP is Mac software and ASIO is currently used on Windows only (it was used on Mac OS 8/9 too in the past).

 

AFAIK, Cubase uses CoreAudio on OS X.

 

A/D - Echo Audiofire, firewire interface.

 

That's a good piece of hardware, shouldn't have any major problems.

 

I can share a link with two short parts of each application record results, 24/96, if you like.

 

That would be interesting!

 

Don't know what hardware issues could happen with Macbook Pro unibody, while it is on battery power.

 

As long as the device is not connected with optical cable, there's always possibility of some RF-noise leaking through the cable.

 

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

Echo: very smart firmware without any glitches.

 

Both Cubase and STP were Mac applications, of course! No differences in setup - only the recording software was different.

Cubase, however, says it use ASIO driver with HOG mode, grabbing the card for itself.

 

RF - again, there's no PSU connected, so there's no ground loops etc.

 

http://dl.dropbox.com/u/2703292/software.zip

Files called in a strange way to make comparison "blind" like.

 

 

 

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


×
×
  • Create New...