No longer working on Arch

[portablevampire]

I'm at a bit of a loss, I've tried googling and reading wiki pages to see if there was a package I'm missing, to no avail. Previously spek was working fine for me on Arch, but I just tried to run it again and got the error:
spek: error while loading shared libraries: libavformat.so.61: cannot open shared object file: No such file or directory

On googling that error, someone with a similar issue on a different program seems to have fixed it by recompiling it. So I tried that. Upon running make, I get the following failure:

make[1]: Entering directory '/home/damiano/spek-0.8.5/src'
  CXX      libspek_a-spek-fft.o
spek-fft.cc:5:10: fatal error: libavcodec/avfft.h: No such file or directory
    5 | #include <libavcodec/avfft.h>
      |          ^~~~~~~~~~~~~~~~~~~~
compilation terminated.
make[1]: *** [Makefile:610: libspek_a-spek-fft.o] Error 1
make[1]: Leaving directory '/home/damiano/spek-0.8.5/src'
make: *** [Makefile:473: install-recursive] Error 1

I have ffmpeg installed, version 2:8.0-3, so it's not like i'm missing libavcodec.or libavformat.
I've followed every lead i can think of to solve this myself but i've run dry. I'd rather not have to downgrade to an earlier version of ffmpeg (if that turns out to be the only solution, i'll just have to live without spek for a while until it catches up i fear)

If there's any other way of solving this, I'd appreciate the help.

[Magikh42]

I have the same issue on EndeavourOS

[alexander-lysenko]

Looks like ffmpeg doesn't include libavcodec/avfft.h anymore so this isn't building again.
ffmpeg4.4 seems to include that file avfft.h but under path usr/include/ffmpeg4.4/libavcodec/avfft.h which I assume isn't being considered in the spek makepkg build process.

I found a fix on the discussion on the Spek's AUR package.
If shortly, create a patch file inside the root path of spek's source code

your-patch-file.patch

From ac42a857038f91c3b796cdbf0f7213b33d70e1b4 Mon Sep 17 00:00:00 2001
From: Mike Wang <mikewang000000@gmail.com>
Date: Thu, 6 Feb 2025 03:14:57 +0800
Subject: [PATCH] fix: Replace deprecated FFmpeg avfft.h APIs

---
 src/spek-fft.cc | 28 +++++++++++++++++++++++++++-
 src/spek-fft.h  |  2 +-
 2 files changed, 28 insertions(+), 2 deletions(-)

diff --git a/src/spek-fft.cc b/src/spek-fft.cc
index 3105213..fca30b0 100644
--- a/src/spek-fft.cc
+++ b/src/spek-fft.cc
@@ -2,7 +2,13 @@

 #define __STDC_CONSTANT_MACROS
 extern "C" {
+#include <libavutil/version.h>
+#if LIBAVUTIL_VERSION_INT >= AV_VERSION_INT(58, 18, 100)
+#define USE_LIBAVUTIL_TX_API
+#include <libavutil/tx.h>
+#else
 #include <libavcodec/avfft.h>
+#endif
 }

 #include "spek-fft.h"
@@ -16,7 +22,12 @@ public:
     void execute() override;

 private:
+#ifdef USE_LIBAVUTIL_TX_API
+    struct AVTXContext *cx;
+    av_tx_fn tx_func;
+#else
     struct RDFTContext *cx;
+#endif
 };

 std::unique_ptr<FFTPlan> FFT::create(int nbits)
@@ -24,18 +35,33 @@ std::unique_ptr<FFTPlan> FFT::create(int nbits)
     return std::unique_ptr<FFTPlan>(new FFTPlanImpl(nbits));
 }

-FFTPlanImpl::FFTPlanImpl(int nbits) : FFTPlan(nbits), cx(av_rdft_init(nbits, DFT_R2C))
+FFTPlanImpl::FFTPlanImpl(int nbits) : FFTPlan(nbits)
 {
+#ifdef USE_LIBAVUTIL_TX_API
+    const float scale = 1.f;
+    av_tx_init(&this->cx, &this->tx_func, AV_TX_FLOAT_RDFT, 0, 1 << nbits, &scale, AV_TX_INPLACE);
+#else
+    this->cx = av_rdft_init(nbits, DFT_R2C);
+#endif
 }

 FFTPlanImpl::~FFTPlanImpl()
 {
+#ifdef USE_LIBAVUTIL_TX_API
+    av_tx_uninit(&this->cx);
+#else
     av_rdft_end(this->cx);
+#endif
 }

 void FFTPlanImpl::execute()
 {
+#ifdef USE_LIBAVUTIL_TX_API
+    float *input = this->get_input();
+    this->tx_func(this->cx, input, input, sizeof(float));
+#else
     av_rdft_calc(this->cx, this->get_input());
+#endif

     // Calculate magnitudes.
     int n = this->get_input_size();
diff --git a/src/spek-fft.h b/src/spek-fft.h
index 409263e..9a16c87 100644
--- a/src/spek-fft.h
+++ b/src/spek-fft.h
@@ -26,7 +26,7 @@ public:
     {
         // FFmpeg uses various assembly optimizations which expect
         // input data to be aligned by up to 32 bytes (e.g. AVX)
-        this->input = (float*) av_malloc(sizeof(float) * input_size);
+        this->input = (float*) av_malloc(sizeof(float) * (input_size + 2));
     }

     virtual ~FFTPlan()
-- 
2.51.0

and apply it with the following command:

patch -Np1 -i your-patch-file.patch

This patch file should help to fix libavcodec/avfft.h build error. Then you can build and install it again

P.S. It worked for me