Acourate filter creation settings

Just wanted some guidance about using Acourate filter with HQPlayer. I have found older posts to use 352 or 384 sampling rate, turning off HF extension in HQPlayer.

My question is if I should use 32 or 64 bit wav and single ou double length.

I’d say most optimal for typically available content is to create for 352.8k rate. Remember to check the “Brick” option in Acourate which is equivalent of HQPlayer’s “HF Expand” which lets you leave that option disabled in HQPlayer.

64-bit floating point WAV is perfect.

Hello @jussi_laako ,

I have created the filters. Acourate states that the overall gain should be about -3.9db. HQPlayer log says 3.857db. but the convolution window displays about -13db. Is there anything I should worry about?

I would get started with the assumption that Acourate specifies it correctly. But keep eye on the Limited counter to make sure you are not hitting the limiter.

Thanks. I see now I made a serious typo. HQ log shows -3.857db, so almost the same as Acourate. The only discrepancy is indeed the convolution window.

Since all values are negative, and no compensation is applied, I guess I am safe.

Thanks again

Where is this convolution window you mention? Could you show a screenshot?

It is under the matrix menu. When you click on matrix, you get two options: convolution setup and pipeline setup. That convolution setup is the window to which I am referring

Did you manually enter that -13dB figure ?

Nope. It is calculated automatically, as soon as one uploads the convolution file

Interesting, thanks. @jussi_laako is it possible to see the same when using Pipeline Matrix also?

No, if you like to see gain estimates, you can use convolution dialog. You don’t need to save the settings, you can just cancel after checking the gain figure.

1 Like

Thanks, so if HQP is showing the below, I should be able to safely add +20dB of gain?

image

Yes, but note that it is an estimate, not absolute figure. So it may not be always correct.

That number is vastly different to what I see in “plot” using pipeline matrix.

Can you see where this IR gain of -25dB would come from? Even though i understand it is estimate, it is a huge difference, not small.

If I add +20dB gain to the below, I will definitely clip right? Like at 100Hz as one example

Yes, that’s the case. You have different tools in these two windows.

Ok thanks. Based on this example then, I would never use the “IR gain” calc in Convolution window.

The estimate isn’t even close. I guess similar to what @Jose_Vita found

Plot in piple matrix window is much better.

I haven’t seen where HQPlayer log shows this